Last week’s caveat was that 2026 sits in the awkward Year +2 after a halving as we wrote in our previous The Insider (5 Oct). In our sample, those Octobers returned -13.1% in 2014, -4.0% in 2018 and +5.5% in 2022, for a median of -4.0%. The seasonal did not disappear; the cycle analogue simply mattered more.

Our base case has not flipped bearish. Sentiment has deteriorated faster than price, leverage has come down, and BTC is still defending the psychological $80,000 area. That is the sort of setup from which a bear trap can form. But bulls now need to reclaim $82,000 first, then $87,000. Do that and a final squeeze towards $90,000 becomes credible. Lose $80,000 cleanly and $77,000 is the next meaningful support test.

Crypto

The $82,000 break changes the map, not necessarily the destination.

The first positive is positioning. Altcoin open interest has fallen through October and our cross-market leverage measure is down roughly 10% on the month. That is very different from early October, when the uncomfortable part of the setup was the amount of leverage sitting behind spot. Recent market data had already shown leverage migrating away from BTC and towards the alt complex, making the latter the more fragile part of the system.

Deleveraging does not create demand, but it removes fuel from the liquidation machine. With sentiment now poor and BTC still above $80,000, the risk/reward is becoming more asymmetric. A clean loss of $77,000 would invalidate that view. Until then, the more interesting possibility is a flush of late longs before one last move higher rather than the start of a straight-line bear leg.

ETF flows support that “messy, not broken” interpretation. Bitcoin spot ETFs were net positive on the week in our data, while Ethereum ETFs continued to see outflows. Yet October remains a net-outflow month for BTC, and 22 October printed roughly $635m of Bitcoin ETF outflows, the largest single-day withdrawal since July. Institutional demand is still present, but it is not providing a persistent marginal bid.

The ETF impulse does not need to become spectacular for a squeeze. It only needs to stop getting worse while leverage is lighter: fewer forced sellers, more negative sentiment, and a market sitting just above obvious support.

Uptober, meanwhile, has lost the marketing campaign. Our own work showed the second half of October as historically strongest, but Year +2 after a halving as the weakest cycle slot. This year is so far behaving much more like the cycle analogue than the calendar average.

Ethereum looks weaker. Price is around the mid-$2,300s and the supplied chart puts the next meaningful support cluster near $2,200–$2,230. The narrative premium has cooled while ETF demand remains softer than Bitcoin’s. BitMine, chaired by Tom Lee, has been pursuing a target of owning 5% of Ether supply and has moved close to that objective. A marginal buyer approaching a stated accumulation target is a different setup from a fresh buyer beginning one.

So crypto enters the final week of October bruised but cleaner. BTC below $82,000 is a technical negative. Lower system leverage is a positioning positive. ETF flows are mixed rather than catastrophic. Sentiment now looks considerably worse than price.

That is why we still lean bear trap, with $77,000 as the line that would make us change our mind.

Macro

The Fed meets on 27–28 October, and the important distinction is between a pause and a pivot. The September FOMC raised the target range by 25bp to 3.75%–4.00%, while describing activity as solid and inflation as still elevated. The next decision lands on Wednesday, and the market overwhelmingly expects no change: a Reuters poll found 94 of 100 economists expecting a hold, while futures by Sunday implied only around a 9% probability of another October hike. December remains the first serious tightening window.

The data give the Fed room to wait, but little reason to declare victory. Flash US composite PMI eased to 57.6 in October, still a strong expansion, while new orders hit their highest since March 2022 and output-price inflation rose to its strongest since August 2022. Initial jobless claims were only 194,000, near multi-decade lows.

The weak spot is housing. September single-family starts fell 4.6%, permits dropped 4.0% to their lowest since March 2023, and builder sentiment is near levels last seen in 2012. That leaves an awkward mix: rate-sensitive sectors are hurting, while demand and energy remain strong enough to keep inflation risk alive.

The real macro swing factor next week is geopolitics. Brent moved above $116 on Friday as Middle East risk intensified, its highest level since 2022. Saudi-Houthi escalation has revived concerns around regional infrastructure and Red Sea shipping, while Russia-Ukraine fighting remains active rather than moving towards a durable settlement.

This is also the clearest upside catalyst for risk. A temporary ceasefire in Ukraine, a de-escalation in the Middle East, or both would remove part of oil’s geopolitical premium quickly. Lower oil would ease inflation expectations, relieve pressure on long yields and make a December Fed hike look less urgent. In a market already deleveraged through October, that could be enough for a sharp risk squeeze.

But that is a scenario, not the current reality. The current reality is still $100-plus oil, a Fed that has not finished tightening, and several geopolitical fronts capable of generating a weekend headline. A quieter tape into the midterms is tactically plausible. It is not a structural resolution.

Equities and the AI Trade

Equities had a deceptively calm weekly close after a much uglier tape underneath.

For the week, the S&P 500 fell about 0.5% and the Nasdaq 0.7%, while the Dow gained 0.4% and the Russell 2000 rose 2.0%. The important move came inside the AI complex: on Thursday the Nasdaq dropped 2.6% and the Global Semiconductor Index lost 6.7%, with several major chip names falling more than 8%. The Index is down 20% in October alone.

The market’s AI question is changing. Six months ago it was whether demand was real. It is. Today the question is who funds the next wave of infrastructure, and what return they earn on it.

Meta made that tension visible. It reported quarterly revenue of roughly $64.6bn, above expectations, but guided to $135bn–$145bn of capital expenditure and posted negative free cash flow of about $2.1bn. The shares fell more than 7%. Tesla moved in the opposite direction, cutting its 2026 capex forecast to around $20bn from more than $25bn, yet the stock still fell as investors questioned how its AI and robotics ambitions translate into earnings.

That tells you the market is no longer rewarding “AI spend” as a category. It wants evidence of returns.

The financing stack is also getting heavier. Nvidia-backed Nscale has been discussing another cloud deal with Anthropic potentially worth around $45bn, while SpaceX has reportedly secured $40bn of debt financing for GPUs due in 2027. AI capex is migrating from an income-statement story into balance sheets, debt markets and long-duration commitments.

That does not make us bearish on AI. Roughly 82% of the S&P 500 companies that had reported by Friday had beaten estimates, and Nvidia plus Micron alone are expected to account for roughly a third of third-quarter earnings growth. But concentration is now a feature, not a footnote. When crowded AI positions unwind, the damage can be violent: hedge-fund technology strategies suffered some of their sharpest losses of the year during Thursday’s sell-off.

The trade is not broken. The financing model is under audit.

Bottom Line

October has delivered the caveat, not the seasonal. Our base case remains a bear trap before one more attempt higher. Reclaim $82,000 and then $87,000 and $90,000 is back in play. Lose $77,000 cleanly and the thesis is wrong.

The Fed should pause on Wednesday, but a pause is not a pivot; December tightening remains live. The cleanest bullish catalyst would be a geopolitical de-escalation that knocks oil lower and takes pressure off inflation and yields. The cleanest bearish catalyst is the opposite: another energy shock into a still-hawkish Fed.

Equities are sending the same message as crypto. The headline indexes are holding up, but the most crowded part of the market is being repriced around funding costs and return on capital.

Before the question was $82,000 or $87,000 first.

This week it is $77,000 or $87,000.

Share

Subscribe now

That leaves us in the waiting room. Last week we moved the framework from “does $82,000 break?” to “does $82,000 or $87,000 break first?” after the former resistance level gave way. The bullish case has better breadth and a friendlier Fed path; the bearish case has slower ETF flows, heavy leverage and a market that is still tactically stretched. That is an extension of the framework laid out in last week’s The Line is Broken.

The interesting part is timing. “Uptober” is statistically real in our data, but it is not really an October 1 effect. Almost all of the historical edge arrives after about 12 October. We are entering the month’s strongest seasonal window, just not quite yet.

Crypto

The headline return hides a healthier relative picture underneath. Across the top 200 assets, 54% have outperformed Bitcoin over the last 90 days. The strongest cohort is ranks 50–100, where 58% have beaten BTC; ranks 20–50 are at 56.7% and 100–200 at 53%. Even the top ten are split 50/50.

That matters because relative strength has been spreading further down the market-cap curve even while headline prices have gone sideways. It is a constructive breadth signal and, if BTC eventually clears $87,000, suggests there is still crypto beta available outside the index leader.

The flow picture is less impressive. US spot Bitcoin ETFs took in only about $82.9m from 28 September to 2 October, down from $2.39bn the previous week. The prior burst was extraordinary: 21 September alone brought about $999m, the largest daily inflow of 2026, and September finished with roughly $2.65bn of net inflows. This week was not an outflow week, but the marginal impulse clearly faded. Ethereum ETFs also swung to roughly $118m of weekly outflows.

The distinction matters. ETF demand remains structurally supportive, but it is no longer accelerating enough to do the market’s work for it. Bitcoin now needs either a renewed flow impulse or an endogenous breakout. The first two October sessions did at least turn positive again, with BTC ETFs taking in $102.7m on Thursday and $31.7m on Friday.

Leverage remains the uncomfortable part. OTHERS open interest is around $25bn versus roughly $27bn for BTC before equity perpetuals. Add equity perps and OTHERS rises to around $31bn. That does not make the market bearish, but it does make it fragile: a relatively small spot move can force a much larger derivatives unwind.

So our map is unchanged. Above $87,000, the range resolves higher and breadth argues the move can broaden. Below $82,000, the breakout structure fails and leverage becomes the problem rather than the fuel. In between, there is not much edge in pretending every $1,000 move is a new regime. Bitcoin was still trading in roughly the mid-$80,000s over the weekend after repeatedly testing but failing to establish itself above the upper end of the range.

Uptober

Is Uptober real? As real as a 13-observation seasonal can reasonably be.

From 2013 to 2025, October produced a median BTC return of +14.4% and a mean of +19.4%, finishing green in 10 of 13 years. Against all other months, the edge survives three tests in our study: Welch t-test p=0.023, Mann-Whitney p=0.014 and permutation p=0.042.

But the useful result is not “October is good”. It is when October becomes good.

From 1–12 October, the median return is only +1.0%, with a 62% hit rate and no statistically meaningful edge (p=0.13). From 13–31 October, the median jumps to +10.2%, the hit rate rises to 85% in 11 of 13 years and the p-value falls to 0.007. The final week alone has a +3.0% median return and a 77% hit rate.

In other words, the first half of Uptober has mostly been marketing. The second half has been the trade. A simple “buy the 12 October close, sell the 31 October close” rule compounded roughly +383% across the 13 historical observations, with the worst year around -10%.

The result is not carried by the obvious blow-off years. Remove 2013, 2017 and 2021 and the 13–31 October median is still about +10%, with an 80% hit rate. Restrict the sample to the post-2018 institutional era and the median remains +9.9%, with an 88% hit rate.

Now the caveat. Year +2 after a halving where 2026 sits has historically been October’s weakest slot: 2014 returned -13.1%, 2018 -4.0% and 2022 +5.5%, for a median of -4.0%. And 2025 already broke the usual year +1 pattern with a -3.9% October.

So the honest framing is not “Uptober guarantees upside”. The calendar becomes materially more favourable around 12 October, while the cycle analogue is unusually poor. BTC was about +2.2% month-to-date by Friday’s close. The next two weeks will tell us which effect wins.

Macro

Macro got friendlier at the policy end of the curve and stayed hostile at the long end.

August PCE, released Wednesday, showed headline inflation at 3.4% year-on-year and core at 3.0%. Monthly core PCE rose 0.2%. Then Friday’s labour report delivered only 29,000 new payrolls in September, with unemployment at 4.2% and the prior two months revised down by a combined 60,000 jobs. Average hourly earnings rose just 0.1% month-on-month and 3.0% year-on-year.

That was enough to flip October Fed pricing. A week earlier, futures implied roughly a two-thirds probability of another 25bp hike. By Friday that had collapsed to about 23%; as of Sunday, CME pricing implied roughly a 77.9% probability of no change and only 22.1% for a 25bp hike. The Fed’s next meeting is 27–28 October.

The oddity is that Treasuries did not celebrate. The 10-year yield touched roughly 5.34% during the week and ended Friday at 5.28%, while the 30-year finished around 5.63%. Policy expectations softened, but the long end remained under pressure amid persistent energy, inflation and sovereign-debt concerns.

That split matters for risk assets. A Fed pause removes one immediate tightening catalyst, but a 5%+ long rate still compresses valuation headroom. The backdrop is less “easy money is back” and more “the Fed may stop making difficult financial conditions worse.”

Equities and the AI Trade

Seasonal equity weakness never really arrived. The S&P 500 slipped only 0.3% over the week and finished Friday within roughly 1% of its record high; the Nasdaq gained 0.5%, while the Dow fell 1.3% and the Russell 2000 lost 0.2%. Friday alone saw the S&P gain 0.7% and the Nasdaq 1.2% as the weak jobs report reduced rate-hike fears.

The bear case has instead migrated to breadth. It is genuinely poor: only about 24.6% of S&P 500 members were above their 50-day moving average on Friday, and only 42.3% were above the 200-day. Across the broader US equity universe the readings were 25.7% and 38.6%. That deserves respect, but not automatic bearishness. Breadth has repeatedly narrowed during this bull market and later repaired through catch-up rather than index collapse.

The other wall-of-worry brick is yields. But 5% is not 7%. In the historical bucket study, months when the 10-year sat between 5% and 6% and yields were rising still produced an average S&P return of roughly +0.8%. At 6–7%, that falls toward zero. Above 7% and rising, average monthly returns turn negative. The current 5.28% is a valuation headwind, not historically an automatic equity-killer.

AI is why the index continues to absorb it. Micron’s latest revenue outlook beat expectations as long-term customer commitments jumped to $32bn from $22bn and remaining performance obligations rose to $150bn. Nvidia authorised a $150bn buyback. AMD agreed to acquire World Labs for $8.2bn to deepen its push into physical AI. US equity funds still attracted $20.6bn in the week to 30 September, with AI optimism cited as a major support.

The caveat is becoming financial rather than technological. Anthropic’s IPO filing showed at least $518bn of planned AI-infrastructure commitments over a decade, about 80% of them non-cancellable or payable regardless of usage. That is the next phase of the trade: the AI boom is moving from an earnings story into balance sheets, credit and long-duration capital commitments.

We remain selective, not bearish. AI demand is real. So is the cost of funding it at a 5%+ risk-free rate.

Bottom Line

Crypto has spent a week refusing to choose. That is not a reason to invent conviction.

BTC remains in the $82,000–$87,000 waiting room. Breadth is improving, with a majority of the top 200 now outperforming Bitcoin over 90 days, and the Fed is increasingly expected to pause in October. Against that, ETF inflows have slowed from $2.39bn to roughly $83m in a week and altcoin leverage remains large enough to turn any failed breakout into a fast unwind.

Seasonality becomes the most interesting new input. The historical October edge is strong, but almost all of it begins after 12 October, and 2026 sits in the one halving-cycle slot where that seasonal has historically disappointed.

Equities are telling a similar story. Breadth is bad, yields are high and concentration is obvious, yet the indexes remain close to records because the AI capex and earnings impulse is still powerful.

For now, price gets the final vote. $82,000 or $87,000 first.

That forces a change in our framework. Last week, $82,000 was the line in the sand: a convincing break would invalidate our call for one more reset towards the low-$70,000s before a stronger Q4. That line has now gone. The next level we care about is $87,000.

The uncomfortable part is that both things can be true. The market is stronger than we expected, and it is also running very hot. ETF and stablecoin flows are constructive and regulation is improving. But breadth is overbought, open interest is extreme and Monday’s breakout was amplified by a major short squeeze. We are no longer waiting for the low-$70,000s, but this is not a clean place to chase.

Crypto

The most important signal this week was breadth. Bitcoin rose 4.8%, but SOL gained about 12% and TOTALE100 rose 10.43%. That is what we wanted to see if the market was going to invalidate the summer range: capital moving further out the risk curve rather than hiding in BTC alone.

The spot bid was real. US spot Bitcoin ETFs took in about $2.39bn between Monday and Friday, their strongest week of 2026 and best since October 2025. Monday alone brought roughly $999m, the largest daily inflow of the year and the largest since 6 October 2025. Ether ETFs added about $690m and Solana ETFs another $188m. The caveat is the shape of the flows: BTC ETF inflows fell every day from $999m on Monday to roughly $134m on Friday. Demand stayed positive, but the initial impulse cooled.

Our stablecoin data tell a similar, constructive story. The seven-day average net inflow has risen to roughly $237m, versus a one-year average near $37m. Historically, readings above roughly $200m have tended to coincide with stronger liquidity and momentum.

But there is a second side. Across our top-100 universe, excluding stablecoins and wrapped or staked derivatives, 25 tokens now have daily RSI above 70 and four: QNT, NEAR, PYTH and ENA, are above 80. Roughly half the universe is more than 20% above its 50-day moving average; around half has Stochastic %K above 80 and CCI above 100. On a composite basis, 53 of the top 100 are flashing at least two overheated signals. Another 30 sit in the RSI 65–70 band, including BTC around 66 and SOL close to 70.

Overbought is not the same as bearish. Strong trends can stay overbought for weeks. The concern is leverage. OTHERS open-interest series is now roughly $26.67bn versus about $27bn for BTC, before equity perps. Add equity perpetuals and OTHERS rises to about $32.66bn. The last two comparable setups in the data: December 2024 and October 2025, were followed by sharp corrections.

Monday showed how much derivatives positioning is influencing the tape. Around $648m of crypto shorts were liquidated over 24 hours as BTC broke the range, yet aggregate crypto open interest still rose 7.6% to roughly $156bn. Old shorts were removed and fresh risk was added behind them.

Our read is therefore more nuanced than “breakout = bullish”. The $82,000 break invalidates our previous tactical base case. But broad technical froth and alt leverage mean the next correction can be fast. A sustained move through $87,000, continued ETF and stablecoin inflows, and no further acceleration in leverage would be healthier confirmation

The relative-value case also remains interesting. TOTALES/Nasdaq chart still shows crypto relatively cheap versus US technology after spending much of 2026 lagging the AI complex. If that ratio clears long-term resistance while BTC holds its breakout, the rotation into crypto beta probably has further to run.

Regulation

The most important regulatory development this week was a technical SEC clarification that could materially change how the market thinks about token value accrual.

On 25 September, the SEC’s Division of Corporation Finance said that, where a crypto system is functional, announcing a buyback of a non-security crypto asset would not by itself amount to a promise of “essential managerial efforts”. The FAQ explicitly recognises treasury management, supply reduction, protocol-funded burns and rebalancing. Where a system is not yet functional, however, a buyback marketed as creating yield or return could still contribute to an investment-contract analysis.

That gives mature protocols a cleaner route to connect protocol revenue with token economics. It does NOT mean “buybacks make a token legal”: the FAQs are staff guidance, not a binding Commission rule, and the underlying asset still has to qualify as a non-security crypto asset. The SEC’s broader March interpretation also makes clear that the treatment depends on the characteristics of the asset and the circumstances in which it is offered or sold.

Our inference is that the cleanest winners are already-functional, fee-generating protocols where value accrual is an extension of an existing product rather than the promise that sells the token. Aave is a good example: its DAO already uses protocol revenue to buy AAVE in the open market and has adjusted the programme according to revenue and treasury sustainability. Hyperliquid and pump.fun show why the theme matters economically: together they have accounted for close to 90% of tracked crypto buyback spending this year. That does not mean the SEC has specifically blessed either token; it means the regulatory framework is becoming more compatible with the type of fee-funded token economics they represent.

The losers are the reverse: pre-functional projects whose main investment case is financial engineering. A buyback cannot substitute for usage, recurring fees or a functioning network.

This fits the SEC’s broader direction. Its March interpretation created a five-part taxonomy: digital commodities, digital collectibles, digital tools, stablecoins and digital securities, while last week’s five-year “Innovation Exemption” opened a controlled route for tokenised NMS equities to trade through permissioned onchain AMM venues. Functional networks and real onchain financial infrastructure are gaining more room to operate.

Macro

Macro did not become easier this week. Crypto simply traded through it.

With a quiet week last week and no major data releases, leaving rates, energy and geopolitics to dominate the tape. By Friday the US 10-year Treasury yield had traded as high as 5.23%, its highest since 2007, while the 30-year reached levels not seen since 2004. That is a meaningful tightening in financial conditions for long-duration assets.

Oil remained volatile. Brent settled Friday around $104.32, up less than 1% for the week, while WTI ended around $92.41 and was down roughly 8%. Markets were caught between hopes of a phased US-Iran path out of the conflict and renewed concern around Houthi attacks on Saudi infrastructure.

Seasonality is worth being precise about. September is the weak part of the midterm-election pattern, not necessarily October. Citadel Securities’ long-run work shows the second half of September averaging about -0.9%; in midterm years September has averaged about -1.5%, with the path weakening into quarter-end before recovering through October and accelerating around Election Day. Its more recent midterm study shows the average path gaining 5.6% from the end of September into year-end.

That argues for caution into month-end, but it also means the seasonal window starts to improve soon. This distinction matters for our crypto view. We were expecting weak risk-asset seasonality to spill over into digital assets. Instead, crypto has broken higher. If BTC can hold above $82,000 and push through $87,000 while 10-year yields sit above 5%, energy risk remains elevated and the weakest seasonal window is still in play, we should treat that as evidence that crypto-specific liquidity and regulatory catalysts are overpowering the macro headwind.

Equities and the AI Trade

Equities reinforced the same message. The S&P 500 gained about 1.2% over the week and the Nasdaq Composite roughly 2.1%, while the Russell 2000 fell about 0.8%. A 5%+ 10-year yield is not being ignored; rather, the market is concentrating into areas where investors still see exceptional growth. AI optimism being a key supporter for equities even as yields and oil remained elevated.

AI was again the centre of gravity. AMD crossed a $1tn market capitalisation on Monday after a 9.6% rise, while the semiconductor index jumped 4.3% that day. Microsoft ended the week with a fresh expansion of Copilot, including code generation and an always-on enterprise agent. Akamai also announced an $11.6bn, seven-year cloud-services deal with Anthropic, showing that the AI buildout is broadening beyond GPUs into CPUs, networking and distributed cloud infrastructure.

At the same time, financing is getting more aggressive. SoftBank launched more than $11bn of bonds to help fund its next OpenAI investment, replacing bridge financing with longer-dated debt. The dollar tranches were marketed in high-yield territory. That is not evidence that AI demand is breaking. It is evidence that the capital intensity of the buildout is moving from an equity-market narrative into balance sheets and credit markets.

Our equity view therefore remains selective rather than bearish. We prefer companies with visible AI monetisation, contracted infrastructure demand, power or networking advantages and real enterprise distribution. With long yields above 5%, there is less room for businesses whose valuation depends only on distant optionality.

Bottom Line

The line in the sand has broken. We said a convincing move above $82,000 would invalidate our low-$70,000s-before-Q4 base case, and it has. We should not defend an old forecast when the market has given us new information.

But invalidating a bearish tactical view is not the same as giving the market a free pass. ETF demand is exceptionally strong, stablecoin liquidity has improved, regulation is moving in the right direction and relative value versus the Nasdaq still has room. At the same time, technical breadth is stretched and altcoin leverage is at levels that have preceded painful resets before. The fact that open interest rose even through Monday’s short-liquidation event only strengthens that concern.

The new line is $87,000. Hold $82,000, clear $87,000 and keep the relative-strength and flow data intact, and the bull case strengthens materially. Lose the breakout while leverage continues to build, and this can still become a fast false-break unwind.

For now, the market has earned the benefit of the doubt. It has not earned complacency.

Share

Subscribe now

That resilience matters. But we are not through the key level yet.

For Bitcoin, $82,000 is now the line in the sand. Until we see a convincing break above it, our base case remains one more correction towards the low-$70,000s before a stronger Q4. If BTC clears that level while crypto continues to outperform into an unfriendly macro and seasonal backdrop, we would consider that view invalidated.

Crypto

Crypto finally did what we wanted to see: it absorbed a Fed hike, a failed US market-structure vote and another push higher in bond yields, and still finished materially stronger. At the point of writing, total crypto market capitalisation was up roughly 6.2% on the week, BTC 5.8%, ETH 6.3% and SOL 11.1%. The important part is not just direction. Breadth improved, ETH/BTC pushed higher and higher-beta assets outperformed Bitcoin. That deserves respect. It does not yet deserve confirmation.

Bitcoin is still inside the broad $74,000–$82,000 range that has defined the summer. Last week we said the market needed to reclaim roughly $80,400–$80,800 on a weekly basis before we would call the structure repaired; it cleared that test. This week BTC recovered $80,000, but $82,000 is now the cleaner line in the sand. A sustained weekly break above it would invalidate our current base case: a difficult back half of September, spillover from weaker risk assets, and one more correction towards the low-$70,000s before a stronger Q4.

The strongest argument for the bulls is relative performance. The TOTALES/Nasdaq ratio, which stalled at long-term resistance last week, has now pushed through that area in our chart. Crypto spent much of 2026 lagging the AI/equity complex; a clean relative breakout suggests capital may be rotating back towards digital-asset beta.

ETH/BTC is telling a similar story, up about 27.7% from its June low and pressing higher through the long-term zone we have been watching.

If BTC breaks the range while dominance does not rise with it, that would be a different signal from a narrow Bitcoin-led rally. It would suggest the market is expanding rather than merely rotating. The tape is also discriminating: some large-cap alts remain below prior highs while stronger narratives lead. We prefer that to an indiscriminate melt-up.

There is a catch. Leverage is rising. Altcoin perpetual open interest recently exceeded Bitcoin open interest, with roughly $29bn (+$11bn in since 1 Aug) in alt OI versus about $28bn in BTC in early September, while consolidated BTC perpetual open interest also rose into this week’s move. A more financed market can accelerate faster on the way up, but it can liquidate more violently on the way down.

Policy was the other major driver. On Thursday the SEC created a temporary, conditional five-year exemptive regime for tokenised NMS stocks. Eligible venues can use permissioned automated-market-maker pools without registering as national securities exchanges, while qualifying liquidity providers receive conditional dealer-registration relief. This is narrower than a blanket green light: tokenised shares must preserve the rights of the underlying stock and the regime carries limits, disclosures and an expiry. Still, it is meaningful. The SEC is permitting a controlled form of onchain market structure for real US equities, and DeFi/tokenisation-linked assets rallied sharply on the news.

That came after the Senate failed to advance the CLARITY Act earlier in the week, initially hitting crypto hard. Bitcoin then recovered above $80,000 by Friday. We care less about the legislative scorecard than the price response: a market that can absorb a negative Washington headline, a Fed hike and higher yields in the same week is giving us information.

Our framework is simple. Below $82,000, we still respect September seasonality and the possibility that leverage gets cleaned out before Q4. Above $82,000 on a weekly basis, especially if ETH/BTC and TOTALES/Nasdaq hold their breakouts, we would stop waiting for the low-$70,000s. When a market refuses to go down in a hostile macro tape, the view has to adapt.

Macro

The Fed delivered the first rate increase since 2023, lifting the target range 25bp to 3.75%–4.00% in a unanimous vote. More important was the path. The September projections put the median policy rate at 4.1% at the end of both 2026 and 2027. With the current midpoint at 3.875%, that is broadly consistent with another 25bp hike this year and no net easing next year. The Fed also sees 2026 PCE inflation at 3.7% and core PCE at 3.4%. This was not a one-and-done signal.

Nor is this only a US story. The Bank of Japan raised its policy rate to 1.25%, the highest in roughly three decades; the ECB had already lifted its deposit rate to 2.50% the prior week. The Bank of England held at 3.75%, but three policymakers preferred another increase. Global policy is moving against the liquidity impulse that normally helps long-duration equities and crypto.

Meanwhile the US 10-year yield has pushed back towards 5% and crude has traded above $100. Yet the Nasdaq finished the week higher and crypto rallied. Relative strength in a bad macro environment matters.

Seasonality is why we are not declaring victory. As we wrote in The Insider - Macro Lit the Match

So the cleaner takeaway is not that September is automatically bearish, but that the seasonal risk tends to build as the month progresses

Geopolitics adds another layer. A NATO fighter shot down a suspected Russian drone in Lithuanian airspace this week, while Poland’s prime minister warned that Moscow could seek to test NATO resolve through strikes made to look accidental. On Sunday, a Russian drone attack in the Kyiv region killed four people, and President Zelensky is due to meet President Trump in New York during UN General Assembly week. Europe increasingly has to price the tail risk that an incident around Ukraine becomes a direct NATO problem.

That is the test for risk assets over the next two weeks. If crypto can keep climbing while rates, oil, seasonality and European security risk all argue for caution, that would be stronger evidence than a breakout occurring in an easy liquidity regime.

Equities and the AI Trade

Equities were more mixed than crypto. The S&P 500 finished roughly flat on the week, the Nasdaq gained about 0.7%, while the Dow fell around 1.7%. The split is notable because the 10-year yield was testing 5% at the same time. Technology is still absorbing a higher discount rate better than the rest of the market, although breadth is far from clean.

The AI story also evolved. Last week the focus was on frontier labs openly discussing whether the pace of model development should slow. This week some of that concern moved from rhetoric into governance spending. Anthropic and Accenture each committed at least $1bn over five years to independent model evaluation and red-teaming, while OpenAI said it would publish regular reports on unexpected or concerning model behaviour and has been working with other leading labs on safety.

Our read is that this is not yet evidence of an AI capex bust. Safety, evaluation and enterprise controls are becoming another layer of the AI stack. Large users are also demanding tighter guarantees around where models run and how proprietary data is protected, which can shift value towards secure cloud environments, evaluation tooling and enterprise infrastructure even if frontier-model growth becomes more measured.

Market action supports that distinction. The Nasdaq rose and AMD gained about 8.5% on the week despite long yields near 5% and oil around $100. Investors are not pricing the end of the AI buildout; they are becoming more selective about who captures the economics.

That is also our equity stance. We are not structurally bearish on AI, but higher rates leave less room for narrative-only valuation. We would rather own businesses with visible demand, backlog, power access, data-centre utilisation or security relevance than buy every asset with AI attached to the story. The next leg should be about earnings capture, not merely exposure.

Bottom Line

This was a strong and broad week for crypto. Alts outperformed, ETH/BTC improved, crypto broke higher relative to the Nasdaq and policy delivered a genuine tokenisation catalyst. The market also absorbed a failed CLARITY vote, the first Fed hike since 2023 and a 10-year yield near 5%.

But the decision level is still $82,000.

A weekly BTC break above it, with dominance contained and the relative-strength charts holding, would invalidate our base case for a low-$70,000s reset before Q4. At that point we would rather follow the tape than defend a forecast.

Until then, the back half of September remains the difficult part of the calendar. Leverage is rising, central banks are tightening and geopolitical risks are harder to dismiss. We remain constructive medium term, but tactically patient.

The market has earned more respect this week. It has not yet earned a free pass.

Share

Subscribe now

Underneath the headline move, the tape was more interesting. US spot Bitcoin ETFs flipped from three straight weeks of inflows to outflow, however ETH went the other way. Last week we said Bitcoin needed to reclaim the $80,400–80,800 area on a weekly basis before we would call the structure repaired. It did not. Crypto relative to the Nasdaq also reached the resistance area we were watching and was rejected. Friday gave the market another attempt after a relatively benign CPI print, but the breakout did not stick.

None of this changes our medium-term view. It changes how we want to express it. The back half of September still looks difficult, and early October may remain messy. We would rather let the market test whether buyers genuinely defend support than chase a failed breakout.

Crypto

The ETF divergence is the cleanest signal from the week. Bitcoin ETF had approximately $460M of outflow between Tuesday and Friday, ending the three-week inflow streak that had brought in more than $3.8bn. ETH on the other hand finished the same four sessions with about $197m of net inflows, helped by a $216m Friday print.

That does not automatically mean institutions are rotating from BTC into ETH. Four days is too small a sample. But ETH’s relative resilience is consistent with price, it was almost unchanged while BTC and SOL fell materially.

In a weak tape, what refuses to go down is worth watching.

Technically, Bitcoin is still in the same place conceptually as last week, just with less momentum. The $80,400 area around the 50-week moving average and the roughly $80,800 breakout level remain the line between recovery and confirmation. Last week’s note explicitly framed a weekly close through that zone as the confirmation we wanted; we never got it.

The first leg of the rally rewarded the obvious high-beta beneficiaries, particularly names tied to Robinhood Chain and tokenisation. Activity accelerated sharply into early September and ARB more than doubled from late-August levels. The trade became crowded quickly.

The most useful chart for us this week is not BTC in dollars. It is crypto relative to equities.

The TOTALES/Nasdaq ratio bounced hard from long-term rising support through the summer, helping turn us more constructive on crypto beta. It has now hit old horizontal resistance and stalled. Relative value improved enough to attract capital, but not enough to establish new leadership. Crypto still looks cheap relative to the AI/equity complex after its 2026 underperformance. The opportunity is becoming more interesting precisely as the short-term setup gets less comfortable.

There are early signs of life. ETH is holding better than BTC, ETF demand has shifted towards ETH rather than leaving crypto entirely, and after Robinhood-linked beneficiaries ran too far, parts of the market are repricing towards levels where fundamentals matter.

ETH’s chart captures the point. It has reclaimed the roughly $2,200 area but remains below the larger $3,000–3,060 resistance band. We would use the same framework across crypto: avoid paying for completed narrative moves and look for assets where usage, cash-flow capture or structural demand is improving faster than price.

September can still punish indiscriminate beta while creating a very good shopping list.

Macro

The macro setup is less friendly than Friday’s equity rally suggests.

August CPI rose 0.4% month-on-month and 3.4% year-on-year, broadly matching the headline expectation. Core CPI rose 0.3% on the month and 2.4% year-on-year. It was not an inflation shock, but it was hardly an all-clear: inflation remains above the Fed’s 2% objective and monthly core was slightly hotter than consensus.

Markets responded by raising the probability of a 25bp Fed hike. By Friday, pricing had moved to roughly 86% for a hike at Wednesday’s decision, from about 72% before CPI. The FOMC meets September 15–16, with the statement due at 2pm ET Wednesday.

The uncomfortable part is that this is not just a Fed story. The Bank of Japan is also widely expected to tighten, with a September hike carrying a high market probability as policymakers warn that rates may need to rise further.

Energy is also back in the equation. August producer prices rose 5.4% year-on-year, with energy a major contributor, while oil has moved sharply higher amid Middle East supply risks. That is an awkward combination: financial conditions can tighten even without a growth collapse if inflation keeps central banks hawkish and long-end yields elevated.

The key on Wednesday is not simply a 25bp hike, but the path afterwards. A one-and-done move is digestible. A broader tightening cycle while oil is firm is not.

This is why we still think the back end of September, and potentially early October, will be tricky. Into the US midterms, inflation is sensitive, central banks are tightening and positioning in technology is hardly washed out. Expect volatility.

Equities and the AI Trade

Equities showed some fatigue. The S&P 500 fell roughly 0.8% last week, the Nasdaq about 0.7% and the Dow 1.6%, while the 10-year Treasury yield approached 5%. Friday’s rebound did not erase the weekly damage.

The more interesting development was the sudden convergence among frontier AI leaders around slowing the pace of model development.

Anthropic CEO Dario Amodei has argued for “pacing the frontier”, including independent evaluators with deep access to leading labs and greater coordination on safety. Sam Altman subsequently said OpenAI was open to slowing development alongside other leading labs. The comments followed concerns around increasingly autonomous model behaviour. OpenAI itself disclosed that during July cyber evaluations its models bypassed controls designed to isolate them from the internet, exploited infrastructure and accessed third-party systems including Hugging Face.

There are two readings.

1. Capabilities have advanced faster than expected and recent evaluations changed the risk assessment of people closest to the frontier. When aggressive competitors arrive at a similar conclusion, investors should not dismiss it as public relations. “The builders are asking for time” is information.

2. Frontier AI is becoming extraordinarily expensive just as open and lower-cost models attack inference economics. DeepSeek’s V4.1-Flash is a useful example. The company says the 552bn-parameter mixture-of-experts model activates only 8bn parameters on input and 16bn on output, while its KV cache uses one-quarter of the HBM and one-eighth of the SSD storage of the prior generation. It is also lowering API prices and routing V4-Pro traffic to the cheaper Flash model from September 14.

That is a reminder that the AI trade is not simply “more intelligence equals more silicon forever”. Better architecture can reduce hardware required per unit of useful output. But efficiency is not automatically bearish for semiconductors: lower inference cost can create more usage, and one architecture does not prove aggregate GPU or HBM orders are peaking.

More importantly, CAPEX has not rolled over. Oracle disclosed more than $30bn of new AI cloud contracts and a $664bn revenue backlog, while Google announced roughly $15bn of AI infrastructure investment in Finland over the next two years alongside a long-term nuclear-power agreement.

Our view sits between the extremes. The safety concern looks real, but a coordinated slowdown also has attractive economics for incumbent frontier labs. It buys time to monetise the infrastructure already deployed, moderates the need for an endless step-up in training costs and slows the pace at which cheaper open models commoditise yesterday’s frontier. That is our inference, not the companies’ stated motivation.

For equities, this distinction matters. We are not calling the end of the AI trade; we are saying the next leg should be less forgiving. If hyperscaler capex, cloud backlog and data-centre utilisation remain strong, the medium-term infrastructure thesis survives even if frontier progress becomes more measured. If those numbers begin to slow at the same time the labs reduce scaling, the market will have to revisit the earnings assumptions supporting the AI complex.

Until then, we remain selectively long rather than structurally bearish. But with rates higher, valuations elevated and a debate beginning around the frontier, this is no longer an environment to own every AI-adjacent asset.

Bottom Line

Crypto failed the test we set last week. BTC did not achieve the weekly break above $80,400–80,800, and crypto relative to Nasdaq was rejected at resistance. Bitcoin ETF flows reversed sharply, although ETH’s continued inflows and relative strength are encouraging.

We expect support to be tested before another durable leg higher. That may be exactly what creates the opportunity: the easiest narrative beta has been repriced, dispersion is increasing and parts of crypto are showing fundamental life. We would use weakness to look for deep value rather than chase what worked first.

Macro is the near-term constraint. Wednesday’s Fed decision, a likely BoJ hike, higher energy prices and weak back-half September seasonality argue for volatility. AI adds another question: the frontier may slow because safety risk is worse than investors appreciate, because frontier economics are becoming less attractive, or both.

The medium-term outlook has not changed.

For the next few weeks, selectivity matters more than conviction.

Share

Subscribe now

Since late August the strongest part of the market has been the names sitting further down the risk spectrum. The top 50-100 names has outperformed both top 50, and top 100. That is after a year where most capital has either hidden in BTC or left crypto all together.

Simultaneously, ETF demand is coming back. BTC had $987.6M of inflows this week, making it the third consecutive week. White the index traded flatish, participation has improved along with risk appetite.

However, from a technical point of view, BTC is sitting exactly where it needs to prove direction. Walking into a September setup that we dont particularly love. The backend of the month has been historically dificult for risk, as we wrote about in last weeks The Insider. Rates remains capable of ruining the party and positioning across equities are hardly washed out.

Market is improving for now, however we remain cautious short term as we havent broken out just yet.

Crypto

The most encouraging about the price action over the last couple of week has not been BTC price action alone, but the breath of the crypto market. Bitcoin has recovered, ETF flows have stayed positive, volatility has settled and now capital appears to be moving further out along the risk curve.

The top 50-100 bucket is the stand out.

A similar signal can be seen in the relative performance of the bucket to BTC over a 90 day look back period. 36% of the 50-100 market cap names have outperformed.

This can be interpret in different ways. The more cynical one is that investors are reaching for beta after the majors have moved. This more constructive take is that sidelined capital are beginning to participate again, which is our core view. You tend to see this progression when market starts to repair itself. Capital firstly moves to the safest expression, if the move holds, investors look further down the curve.

That dosent mean that altseason has arrived, as currently only 35% of the top 200 has outperformed BTC over the last 90 days.

But it does mean the market is becoming less dependent on BTC to do all the work. ETF flow is showing the same point.

Bitcoin ETF inflow over the last 3 weeks has surpassed $3.8B cumulatively, with over $980M this week. What matter is consistency. Three weeks of sustained demand starts to look more structural, also explains why BTC is sitting near resistance without immediately rolling over.

But resistance is still resistance. The area is sitting around 80,400 (50W MA) and 80,800 breakout level.

The 50-week moving average sits around $80,400, while we would like to see a proper weekly break above roughly $80,800 before calling the trend confirmed again. Until that happens, BTC is still recovering into supply.

The distinction matters. A weekly close above that zone would materially improve the structure and probably give the rest of crypto permission to extend. Failure here would keep us inside the same broader range and make another move back toward support entirely plausible.

We are also watching crypto relative to equities

As we’ve highlighted before, long-term, crypto still looks cheap vs NASDAQ after a big relative underperformance in 2026. The ratio bounced on its long term support. The ratio has bounced hard from long-term trend support, but it is now running into an old resistance area. That makes the next move quite important. If crypto can break through relative resistance at the same time BTC clears the 50-week MA, the argument for a genuine capital rotation becomes much stronger.

RWA Tokenisation

Tokenisation has been one of those themes that everyone within blockchain has agreed on for years. Eventually everything goes on-chain. Eventually settlement becomes instant. Eventually securities become programmable.

The problem was that for a long time very little actually happened, however that is beginning to change.

Token Terminal now shows roughly $44.9bn of real-world assets on-chain. US Treasuries remain the biggest category at around $15bn, followed by yield strategies at $9B. More interestingly lately, tokenised equities are starting to become more relevant. Its onchain market cap has reached $3.1B, with ETFs being around $644m. Still small numbers compared to traditional markets, but its mainly the volume that caught out attention.

Over the past 90 days, tokenised stocks has generated around $15.9B of DEX volume. Uniswap being the leader with roughly 6.6B followed by PancakeSwap $5.5 and Raydium $2B. The top three accounts for close to 90%.

On the issuer side the concentration is similar.

Binance bStocks generated $7.2B of issuance vs Robinhood $4.2B and xStocks $2.5B. The top three represent >86% of the volume.

The split between chains is also something worth watching.

BNB Chain is currently in the lead with around 32% of the tokenised stocks market cap, followed by Ethereum at 25% and Solana at 24%. Distribution matters as much as decentralisation: issuer partnerships, cheap execution, stablecoin liquidity, wallet access and the ability to reach users where they already are.

That also means the investment opportunity may not be where people first look.

Macro

The macro setup became slightly less friendly this week, Friday’s US payroll report came in stronger than expected, with nonfarm payrolls rising 162,000 in August and unemployment holding at 4.1%.

The immediate read-through was higher rate expectations. That matters because the market now has CPI and PPI directly ahead of the September Fed meeting.

For crypto, the equation is fairly simple. ETF demand is providing a cushion. Breadth is getting better. But if inflation surprises higher again and yields reprice, BTC will be trying to clear major weekly resistance while financial conditions are moving against it.

September seasonality does not help much either. The important nuance, as we wrote last week, is that September weakness has historically been much more concentrated in the second half of the month rather than appearing immediately from day one.

Equities and the AI Trade

Equities themselves were almost unchanged on the week. The S&P gained roughly 0.1%, the NASDAQ around 0.4%, while the Dow slipped slightly.

Underneath that calm surface, fund flows were a little weaker. US equity funds saw roughly $11bn of outflows while money-market funds attracted close to $49bn. There is plenty of cash sitting on the sideline.

We still think the “SaaSpocalypse” narrative got ahead of itself, the idea that enterprises were suddenly going to rip out entire software stacks and replace them with autonomous agents always felt more like a positioning unwind than a realistic six-month business plan. That does not mean AI is irrelevant to software.

It means adoption takes time, integration is messy, enterprises move slowly and incumbents have more distribution than the market occasionally gives them credit for.

The better distinction now is between companies actually monetising AI and companies merely associated with it. With yields still elevated, investors can afford to be picky. Scarce infrastructure, real revenue growth and pricing power can continue to command premium valuations. Pure narrative becomes harder to justify.

What has changed is that AI is no longer the only interesting part of the market.

Crypto has badly underperformed the AI complex this year. With ETF flows staying strong and breadth finally improving, some of that relative-value gap is beginning to attract attention. We would not call that the end of the AI trade. We would call it the beginning of a less one-dimensional market.

Bottom Line

Crypto is improving under the surface. ETF demand remains strong, smaller names are beginning to outperform, and capital is finally reaching for beta again. But Bitcoin is sitting at the level where the market has to prove it. A weekly break above the $80,400–80,800 area would change the picture materially. Until then, we are still testing resistance.

Tokenisation, meanwhile, is quietly becoming one of the more interesting structural themes in the market. The $44.9bn of RWAs is impressive, but the $15.9bn of tokenised-stock DEX turnover is the number we care more about.

September probably gets harder from here, particularly in the back half. That does not make us bearish. It makes us less interested in chasing and more interested in setting levels which makes sense medium term from a value point of view.

The medium-term setup still looks good.

For the next couple of weeks, entry matters more than excitement.

Share

Subscribe now

But the deeper story was not just flows. It was macro.

The market latched onto Treasury’s move into the long end and immediately translated it into the language of the debasement trade. Gold ripped, Bitcoin ripped, and the whole “hard assets vs. paper promises” complex went multi-standard-deviation in a hurry. That does not mean the market is wrong to react, but it does mean the reaction may be over-reading the announcement.

In our view, the starting conditions mattered more than the headline itself. Gold had been basing for months, Bitcoin positioning had been washed out, shorts were meaningful, crypto had been left for dead, and sentiment was still skeptical. Bulls had been waiting for a spark. Bessent’s move gave them one.

While the medium-term picture has improved, we think the short-term trade has become less asymmetric.

Crypto

From a market-structure perspective, Bitcoin is now doing exactly what strong assets do after a regime change in perception: it is pressing into resistance rather than fading from it. The current picture still looks like a range market that is trying to transition into expansion. Price has pushed back into the upper end of the range, near the range highs around $80k, after reclaiming the middle of the structure.

That is constructive. But it is also where markets often stall, chop, or throw a local fakeout before deciding whether they truly have enough fuel to break out.

So the short-term setup is this: if BTC can absorb supply up here without giving back the move, then the market can start talking about continuation. If it loses momentum, the more likely path is a cooling move back toward the mid-range. That would be healthy, not bearish.

Ethereum looks even more tactical. The move has been impressive, especially because ETH spent so long looking structurally broken relative to the narrative around it. It has now reclaimed the lower support band and is pushing back into a zone that matters. But here too, the market is now overbought on short-term momentum measures. That makes ETH very tradable, but not necessarily very buyable at this exact moment unless one has a much longer time horizon.

The same logic applies across the rest of crypto. Our baseline remains that the broader debasement trade is alive, but the current extension looks vulnerable to a pullback. The market is acting as if the Treasury long-end buyback announcement was a watershed event for the entire monetary order. That may be overstating it. It is more likely that this was an important headline that landed into a market already primed to move. The debasement trade probably wanted to rally anyway. Now that it has rallied, the bar for fresh upside without a reset is higher.

Macro

Treasury announced increased sizes for its nominal long-end liquidity support buybacks beginning September 9, a move the market quickly read as supportive for duration and friendly to the anti-fiat trade. Meanwhile, at Jackson Hole, Chairman Warsh emphasised that the Fed must be confident underlying inflation is moving to target and said he is “committed to a discipline, not to a decision.” In other words, he did not signal a dramatic policy pivot; he signaled conditionality and continued inflation vigilance.

The market heard two things at once: Treasury is willing to help the long end on the debt-management side, while the Fed is not yet promising a wholesale easing regime on the monetary side. That combination was enough to energize risk and hard assets, but not enough to fundamentally rewrite the macro playbook.【

This distinction matters. Treasury buying long-end bonds is not QE. It does not magically create a fresh tidal wave of reserves. What it does do is help support liquidity and shape expectations around long-duration issuance and market functioning. Markets interpret that through the lens of term premium, financing conditions, and the broader politics of debt management. So yes, it can be bullish for gold and bitcoin at the margin. But no, it is not automatically the beginning of some new debasement supercycle. That is why I think the reaction function was partly legitimate and partly narrative overshoot.

The market is still behaving as if there is only one more hike in the next 12 months, which tells you how skeptical it remains of a truly aggressive hiking path from here. That makes the labor data next week absolutely central. Jackson Hole gave us tone. The data will decide the path.

Next week’s calendar is effectively a referendum on the labor market. Here is the clean framework:

1. Monday: markets reopen and digest the weekend macro narrative, including the implications of the Venezuela oil deal and the continued debate around Treasury’s long-end operations.

2. Tuesday: ISM Manufacturing and JOLTS are both high-impact for the rates narrative.

3. Wednesday: ADP employment change acts as a preview, though it is rarely the last word.

4. Thursday: ISM Services matters because it speaks directly to nominal growth and labor resilience.

5. Friday: Nonfarm payrolls and unemployment will be the biggest data of the week and arguably the final major labor print before the market fully settles into its FOMC expectations.

If the job market deteriorates, the logic is straightforward: the Fed cannot credibly threaten a more aggressive hiking path into weakening labor conditions, so hike odds fall and risk likely breathes easier. If the job market re-accelerates, the exact opposite happens. A stronger labor print would reopen the door to a more hawkish September conversation and likely hit both equities and crypto in the near term.

That is why macro is the main story here. Crypto moved, but macro gave it permission.

Equities and the AI trade

On the equities side, our core view has not changed much. The Situational Awareness low remains in place, and the SaaSpocalypse has been delayed or perhaps cancelled altogether. The market has, quite sensibly, come around to the idea that large enterprises are not about to replace entire software stacks with a pile of autonomous agents overnight. That narrative got way ahead of reality. It was a fear trade dressed up as inevitability. Markets do that all the time: they take a kernel of truth, wrap it in greed or panic, and run way past what is actually likely in the near term.

We still think the bigger macro backdrop is the same one we have been describing for a while, nothing Warsh said at Jackson Hole meaningfully changes that. The biggest risk to equities is not some sudden disappearance of liquidity. It is positioning and seasonality. The market is pretty fully bulled up into a historically weak seasonal window.

That said, the important nuance is timing. September is not uniformly bearish from the opening bell. The more recent 2020–2025 sample is particularly weak, with an average full-month return of -2.6%, driven largely by the back half of the month, where the average decline was roughly -2.0% versus -0.7% in the first half.

Expanding the study to 2016–2025 softens the magnitude but not the pattern: the average September return improves to -1.1%, with the first half roughly flat at -0.2% and the second half still doing most of the damage at around -1.0%.

So the cleaner takeaway is not that September is automatically bearish, but that the seasonal risk tends to build as the month progresses. Bulls can still have room early, but the historical backdrop becomes materially less friendly in the back half.

So for equities, the working view is: the storm around AI/SaaS has calmed, the macro backdrop is still supportive enough to keep the broader bull alive, but the tape is crowded and the calendar is getting less friendly. That is not a call for disaster. It is a call for better timing.

Bottom line

Bringing it all together, our stance is fairly simple. The medium-term case for the debasement trade is intact. The shorter-term case is much less attractive after this week’s squeeze. Crypto, gold, and the broader anti-fiat complex have already had the move. When markets go vertical on a news catalyst that arrives into favorable positioning, the first job is not to get more excited. The first job is to ask how much of the good news is already in the price.

Right now, probably a lot.

So the tactical playbook is patience. If labor weakens and risk wobbles only modestly, that likely sets up the next buying opportunity. If labor comes in hot and markets finally respect the possibility of a more hawkish September path, that probably gives us the correction that crypto now looks ripe for. Either way, we would rather be a little careful here than chase after a week in which almost every crowded debasement expression already went bananas.

The bigger trend is still higher. But after this move, the better trade is probably waiting for the market to remind everyone that even bullish tapes need to breathe.

Share

Subscribe now

Our previous piece was titled Bottoming, Not Breaking Out. The core message was that forced selling was fading, ETH was beginning to lead, and the TOTALES/Nasdaq ratio was sitting on long-term support, but BTC still needed to reclaim $65.8k and, more importantly, the 200-day moving average around $73.9k before we could call the trend repaired. That framework worked. BTC cleared both levels, ETH has been the stronger major, and the TOTALES/Nasdaq ratio has bounced sharply from its rising structural trend line.

The question has therefore changed. We are no longer asking whether crypto is bottoming. We are asking how much of the new move has already been pulled forward by leverage and forced buying.

Crypto

Last week was the cleanest confirmation of the July thesis, but also the point where the short-term risk/reward became less attractive.

BTC exploded through the $69k–$75k area and briefly traded close to $79.5k before settling around $77k into the weekend. Structurally, this is exactly the price response we wanted to see after the July accumulation phase: the lower support zone held, the November 2021 ATH around $69k was reclaimed and price finally broke through the more important $73k–$74k resistance.

August 19 then produced the largest short-liquidation event in crypto history. The significance is not simply the dollar amount. It tells us a meaningful part of the move through resistance was mechanical: once BTC cleared the old highs, shorts were forced to buy back into an already thin market.

That accelerates price, but it also means the first move through resistance can overshoot its sustainable short-term equilibrium.

The $74k–$75k zone we were watching has now been reached. That was the upside objective; it is not where we want to become more aggressive. Daily BTC RSI is now around 80. ETH is close to the same level, while the TOTALES/Nasdaq ratio is similarly stretched following its bounce from long-term support. Price can remain overbought for longer than expected during a strong trend, but the combination of strong resistance, record liquidations and elevated momentum readings makes chasing here unattractive. Our view is a short-term correction or consolidation.

The first level is the breakout around $73.8k. If that gives way, $69k–$70k becomes the more attractive retracement area. It would retest the November 2021 high, revisit the order block created during last week’s breakout and allow positioning to reset before another attempt higher.

Importantly, a correction here would not invalidate the broader move. It would probably improve it.

ETH continues to look better on a relative basis. It has reclaimed the $2,200–$2,230 area and is now trading around $2,450. The next meaningful resistance sits around $3,040, followed by the larger historical supply zone around $3,800–$4,100. That sequencing matters.

In July we wrote that better-quality beta would likely improve before broader altcoin participation. That is what has happened: BTC stabilised first, ETH then became the cleaner high-beta expression, while most of the alt market lagged.

The next phase may be broader altcoin participation, but there is one positioning signal we do not want to ignore.

OTHERS open-interest dominance has now moved above BTC open-interest dominance. In the mainly 2024–2025 sample, these flips have tended to mark short-term local tops for alts rather than full-cycle tops. The typical move afterwards has been an 18–30% correction over the following one to two weeks. March 2024 produced roughly a 22% average decline, while December 2024 and September/October 2025 showed similar leverage resets.

We would treat this as a leverage warning, not a reason to abandon the broader thesis. Twenty-to-thirty percent drawdowns are perfectly normal in strong crypto trends. There is also an important caveat. “Others” OI is less pure than it used to be. Tokenised equities, equity perpetuals, gold and other non-alt products, particularly on venues such as Hyperliquid, increasingly sit inside the bucket. The historical signal is therefore somewhat muddier today.

The setup we see is relatively simple: BTC cools toward $69k–$70k, ETH retains its reclaimed support, leverage gets washed out and broader altcoin participation follows.

That would be a much healthier environment for the next leg than simply extending this week’s squeeze vertically.

Macro

The most interesting macro development this week was not the Fed. It was the Treasury.

Stress had been building at the long end of the U.S. curve, with 30-year yields returning to levels seen before the latest round of intervention. Treasury subsequently increased its buying of long-duration bonds, while Scott Bessent has made clear that authorities have room to intervene further if market functioning deteriorates.

The distinction matters. Treasury buybacks are not QE. They do not change monetary policy in the traditional sense. What they do is provide a buyer into a part of the curve where liquidity and term premium have become increasingly problematic.

The market’s interpretation appears to be that policymakers have developed a pain threshold for the long end.

This is marginally supportive for liquidity-sensitive assets because it reduces the probability of an uncontrolled long-end tightening shock. But it does not solve the underlying fiscal or inflation problem. If anything, persistent intervention without an improvement in fiscal expectations could eventually cause investors to demand a larger term premium.

For now, Treasury has bought optionality rather than solved the problem. The Fed side is less aggressive than it looked earlier in the summer. Markets are effectively pricing roughly one additional hike over the coming 12 months, rather than the beginning of a sustained hiking cycle. That matters more than whether the next meeting itself produces a hike. Risk assets care about the expected path.

Oil remains the complication.

Traffic through the Strait of Hormuz has recovered from the most stressed point, but energy prices remain supported. That divergence is important. More vessels moving through the strait reduces immediate tail risk, but the market is still attaching a geopolitical premium to supply, insurance and the possibility that conditions deteriorate again.

In other words, higher traffic has reduced panic without fully removing the oil premium. That leaves the Fed in an awkward but manageable position: inflation risk has not disappeared, but neither oil nor the bond market currently looks disorderly enough to force a major policy response.

For crypto, that is a considerably better macro backdrop than the one we were dealing with before the summer.

Equities and the AI trade

The other major development since July has been the reset in the AI trade.

In our last Insider, we highlighted the relative risk/reward developing between crypto and AI equities. The TOTALES/Nasdaq ratio was sitting directly on long-term structural support, while semiconductor positioning was crowded and investors were beginning to question how the economics of AI capex would eventually be distributed.

That ratio has now responded.

At almost exactly the same time, the crowded side of the AI trade experienced one of its largest forced-positioning events. Situational Awareness suffered a massive liquidation of its AI-heavy portfolio. The important development since then is that the market has absorbed much of that supply. Citadel subsequently distributed more than 80% of the portfolio it acquired, reducing one of the largest technical overhangs sitting over AI equities.

The AI market has rebounded since. That does not mean the AI trade is finished. Quite the opposite. The fundamental AI capex cycle remains enormous. But the market is becoming more selective about where the returns accrue, while the liquidation removed some of the most extreme leverage in the system. This brings us back to the relative-value argument from July.

The cleanest way to quantify it today is through flows.

We are not yet seeing a wholesale exit from AI into crypto. That would be too strong a conclusion. But the relative marginal flow has shifted: semiconductor beta has experienced redemptions at the same time institutional BTC demand has accelerated.

That is exactly what an early-stage rotation should look like.

The next major test arrives this week with Nvidia reporting Wednesday after the close. NVDA remains the centre of the AI complex. A strong print and strong forward guide would reinforce the argument that July was mainly a leverage and positioning event rather than a deterioration in underlying AI demand. A weaker guide, particularly around capex returns, hyperscaler demand or margins, would reopen the debate very quickly.

For crypto, the optimal outcome is probably not an AI collapse. It is an AI market that remains healthy but stops absorbing virtually every marginal dollar of speculative growth capital. That creates room for another high-beta theme to compete. Crypto is finally beginning to do that again.

Bottom line

The medium-term setup is constructive; the short-term setup is stretched.

July was about identifying a bottom. August has given us the confirmation: BTC reclaimed the levels we needed, ETH led, the crypto-versus-Nasdaq ratio bounced from structural support, and ETF demand came back with real size.

But record short liquidations, RSI readings near 80 and the alt-OI dominance flip tell us that this is not the moment to confuse confirmation with a good entry.

Our view is a BTC retracement toward $69k–$70k, followed by renewed strength and a broadening from ETH into a wider set of alts. If BTC holds $73.8k without giving much back, the market may simply consolidate higher. If $69k fails, the $65k–$67k zone becomes the key test. If the reset holds and BTC subsequently establishes acceptance above $80k, the next higher-timeframe area we are watching is the $100k–$108k supply zone.

The bigger picture is improving. The Treasury has shown it will respond to stress in the long end, the Fed is no longer being priced for an aggressive hiking cycle, the Situational Awareness overhang has largely been distributed, and crypto is beginning to compete for institutional risk capital again.

We would buy the reset rather than chase the squeeze.

Share

Subscribe now

Macro remains volatile. Softer U.S. labour data have weakened the case for an immediate tightening shock, but shifting rate expectations, persistent inflation risks and oil volatility mean the Federal Reserve has not turned decisively dovish. The most supportive outcome for risk assets would be an extended pause followed by a gradual softening in policy expectations.

Equities remain an important transmission channel for crypto. Investors are beginning to question the durability and distribution of returns from AI capital expenditure, with attention potentially shifting from semiconductor suppliers towards hyperscalers and software companies. A broadening or cooling of AI leadership, rather than a collapse, could allow crypto to attract more relative-value capital.

Crypto

BTC’s tactical message did not materially change this week: selling pressure eased, but a durable uptrend has not yet been confirmed. BTC spent much of the week in the low-to-mid $60,000s, briefly trading near $61,335 after Strategy disclosed fresh BTC sales before recovering towards $64,000 by Friday.

Strategy disclosed that it sold 3,588 BTC for approximately $216m between 29 June and 5 July. The announcement initially revived concerns that one of the market’s most important treasury buyers had become a seller. However, the company indicated that the sales were driven by balance-sheet considerations rather than the start of a broader corporate liquidation. The market therefore treated the disclosure more as the removal of an overhang than as a new systemic threat, although confidence in the crypto-treasury model remains weaker.

The ETF story matters more than the price bounce. The week of 6–10 July produced approximately $197.4m of net inflows after eight consecutive outflows. This is constructive, but the scale remains modest: it is enough to suggest that the bleeding has paused, not enough to conclude that institutional sponsorship has been fully repaired.

A bottom can form while flows are merely becoming less negative, but a new bull leg would normally require consistent and repeatable absorption rather than a single positive week after a prolonged washout.

Positioning and market structure therefore remain more important than excitement. BTC is still a bottoming market rather than a reclaimed trend. The 200-week moving average continues to provide support, while approximately $65.8k is the first meaningful resistance level. The more important trend checkpoint is the 200-day moving average near $73.9k.

Until BTC reclaims that higher trend line, the better interpretation is that forced selling and panic may be behind us, but the market has not automatically returned to a bull regime. Capitulation appears to have occurred, early accumulation signals are emerging, and the market is now determining how long that phase must continue before markup can begin.

ETH has quietly done the more impressive work. It defended the 1,500 area and outperformed BTC during the week. The weekly performance snapshot shows ETH up 2.0%, BTC up 0.91% and SOL down 4.8%.

That relative strength matters because better-quality beta will often begin improving before broader altcoin participation follows. SOL, by contrast, remained weaker. The resulting hierarchy is familiar for an early bottoming phase: BTC is trying to stabilise, ETH is attempting to lead, and higher-beta alt exposure remains too early to chase aggressively.

This is the type of price action expected in a low-volatility accumulation zone rather than a clean breakout regime. Support is better defined than it was two weeks ago, but a stronger price response is still needed to confirm that accumulation is transitioning into markup.

Macro

Macro is still the first derivative for crypto. The June U.S. payrolls report showed only 57,000 jobs added, well below expectations, with April and May revised down by a combined 74,000. The unemployment rate dipped to 4.2%, but only because labour-force participation fell to 61.5%, the lowest since March 2021, while wage growth ran at 3.5% year-on-year. In practice, that report reduced the urgency for the Fed to tighten immediately.

That softer labour pulse has to be weighed against a still-uncomfortable inflation backdrop. The probabilities are implied from 30-day Fed Funds futures, and the June FOMC minutes showed a real split inside the Committee: many participants still saw a case for inflation easing on its own, but many others thought rates might need to move higher if price pressures persisted. Fed’s monetary policy report made clear that AI-related investment, tariffs and conflict-linked energy costs were all being discussed as upside inflation risks. In other words, the Fed is not relaxed. It is conflicted. Market has moved from pricing in 50/50 unchanged/hike in September to 51% for a hike and 18% chance of a 50bp hike.

Oil is the spoiler, but this week it was volatile rather than one-way. Crude began the week under pressure as OPEC+ approved more August supply and traders focused on recovering flows through Hormuz. Mid-week, Brent spiked towards $78–$80 after Trump said the ceasefire with Iran was over and fresh hostilities hit sentiment.

By Sunday, however, Trump was again saying the strait remained open to commercial traffic, while Reuters described Gulf bourses as subdued rather than panicked. The practical implication is that energy remains a source of volatility, but not yet a sustained inflation shock large enough to force the Fed’s hand immediately.

For crypto, that matters. A Fed that faces softer jobs data, a still-tight but not decisively re-accelerating inflation picture, and oil that is no longer relentlessly moving in one direction has more room to lean less hawkish at the margin. That is not the same as imminent cuts. Our view is that if markets correct again and growth data continue to soften, the path of least resistance for the Fed could drift from “threaten hikes” towards “wait, then soften”. That is not a bullish catalyst by itself, but it is consistent with a grinding low-volatility bear-market endgame rather than a new leg lower forced by policy.

Equities and the AI trade

The equity market remains the other key transmission channel. At the start of the week that investors was pivoting from chipmakers to hyperscalers, and separately that hedge funds had sold tech hardware and semiconductor exposure for a fourth straight week.

Looking at the IGV [Software ETF] versus SMH [Semiconductor ETF] overlay: software and semis are still moving with an unusually tight inverse relationship, as the market is increasingly debating whether AI capex remains a pure positive for chip suppliers or whether marginal benefit now accrues more to the firms that can monetise the software layer.

That debate matters for crypto because the TOTALES/Nasdaq ratio only needs AI to stop being the sole destination for risk capital.

The semiconductor index rebounded 2.2% on Monday, even as the broader conversation turned to second-quarter earnings and whether the AI trade can keep justifying itself. The next real test is earnings season, which begins in earnest in late July. Meta remains one of the most important names in that debate because it raised 2026 capex guidance to $125bn–$145bn in April; if management later signals that the slope of spending is flattening, software could outperform while semis de-rate at the margin. A part of our rotation thesis we think can play out in the near future.

Bottom line

The base case in crypto remains the same. The market is showing signs of bottoming, but not enough to declare the bear market over. The most likely path over the next one to two months is low-volatility chop, a gradual accumulation process, and the possibility of one final scare below 60k before a cleaner trend can form. The probability of a move into the low 50s looks low; a sustained break below 50k looks lower still. What would change that view is not simply more bad macro headlines, but a failure of support to hold after ETF flows have already started to stabilise.

The bull case is straightforward. BTC breaks and closes through 65.8k, reclaims the high-60s with authority, and then starts working back towards the 200-day moving average around 73.9k. If that level is reclaimed, the market can start talking about trend repair rather than merely survival. In that scenario, ETH likely continues to lead on a relative basis, and the TOTALES/Nasdaq ratio probably confirms the idea that crypto is beginning to win back risk capital from a more uncertain AI complex.

The bear case is also straightforward. ETF demand fades again after one positive week, oil volatility re-accelerates, the Fed turns more hawkish into August, and BTC loses the 200-week moving average decisively. That would reopen the case for a final liquidation flush. For now, though, the market is not trading that path as the central case. It is trading something more nuanced: fear has eased, accumulation may have started, but price still needs to do the last piece of work.

Share

Subscribe now

Crypto

What changed this week was price. What has not yet clearly changed is sponsorship. The ETF complex still tells a story of institutional hesitation. June was the worst month on record for U.S. spot bitcoin ETF flows, with about $4.06bn in outflows, enough to push the category negative for 2026 as a whole before a $221.7m inflow on Thursday, 2 July, broke a 10-day outflow streak. Even after that bounce, the message remains the same: one green day is not a repaired buyer base. The market still needs repeatable absorption, not just a reflex rally.

That is why the tone should stay cautious even though the tape looks better. Price responded to a softer macro pulse, lighter positioning, and a short squeeze. But durable lows in crypto tend to be built on one of two foundations: either macro conditions are moving meaningfully easier, or flows return and prove that institutions have decided the washout is over. At the moment, we have the beginnings of the first and only a hint of the second. A market can rally hard under those conditions, but it is still vulnerable to disappointment.

The Strategy overhang also remains central. The company’s enterprise value fell below the value of its bitcoin holdings for the first time, with its mNAV ratio slipping to 0.99. That is important because the market is no longer valuing Strategy as a simple perpetual premium vehicle on top of bitcoin. It is questioning the durability of the structure itself. Strategy also announced a share repurchase programme, and authorised up to $1.25bn in bitcoin sales to build a liquidity cushion. Whether or not a larger liquidation ever happens, the market now has to price the possibility that liquidity management takes precedence over relentless BTC accumulation.

That overhang is why this still feels more like a low-building process than a completed low. The cleaner bullish read is that a real reset is happening underneath the surface. CoinDesk reported that wallets holding large balances accumulated more than 270,000 BTC, roughly $16.7bn, over the past two weeks even as ETF demand remained deeply negative. That kind of split, non-native institutional flows selling while larger, longer-duration holders absorb supply, is exactly the sort of divergence that often appears close to important lows. Looking at the Supply in profit (%) for BTC, we hit the first time loss-making supply has overtaken profitable in this cycle.

That does not prove the bottom is in, but it does say stress is becoming real enough to matter.

Galaxy scorecard fits that interpretation. Four of the thirteen bear-bottom indicators have now reached their bear-market zones, with two still sitting there. That is not a full capitulation signature, but it is enough to say the market is no longer obviously expensive on cycle metrics. The right read, in our view, is not “all clear”; it is that downside asymmetry is improving while uncertainty remains unresolved.

Relative value is also getting more interesting. One of the clearest themes through June was that crypto was not just fighting its own internal liquidation story; it was competing for capital against the AI complex. At the start of last week, U.S. investors were already questioning the sustainability of AI spending, while semiconductor and other AI-linked stocks fell on Wall Street mid-week as investors rotated into financials and healthcare.

That is exactly why the TOTALES (total market cap ex-stables) / Nasdaq ratio matters here. The ratio is testing a structural support zone rather than breaking cleanly through it. If the AI/momentum leadership continues to wobble at the margin while crypto stops leaking flows, the asymmetry in that ratio improves quickly. It is not yet a trigger, but it is no longer just a theoretical conversation. Crypto does not need AI to collapse; it only needs the market to stop treating AI as the only destination for risk capital.

Macro

Macro remains on top of the stack. The single most important development this week was the June U.S. jobs report. The Bureau of Labor Statistics reported payroll growth of just 57,000, well below consensus expectations around 110,000. April and May were revised lower by a combined 74,000. The unemployment rate was 4.2%, but that was flattered by a sharp drop in labour-force participation to 61.5%, the lowest in more than five years. Average hourly earnings rose 0.3% month on month and 3.5% year on year. That is not a collapse, but it is a meaningful cooling in labour momentum and not the sort of report that screams for an urgent rate hike.

The complication is that inflation still looked sticky coming into the week. The Fed’s preferred price gauge showed headline PCE inflation at 4.1% in May and core PCE at 3.4%, both still well above target. That sticky inflation print was the reason the market had spent late June leaning hawkish again.

But the jobs report changed the tone materially. After payrolls, markets cut the perceived chance of a July hike to below 20% and reduced September hike odds to a coinflip while still keeping one hike somewhere later in 2026 alive in the curve. That is a crucial distinction. The market is no longer confidently pricing a hawkish Fed; it is pricing a Fed that might still want to threaten one.

The other key macro release came through energy. Oil rolled over hard as the week progressed. Brent settling sub $72 on 1 July, the lowest closes since March, as U.S.-Iran talks eased supply concerns and the reopening in Hormuz progressed faster than feared. OPEC+ approved another August quota increase as prices had fallen back towards pre-war levels around $72. If June’s hawkish repricing was partly an energy story, then July has begun by pulling one of those pillars away.

That leaves the Fed in a trickier position than rate markets implied even a week ago. A central bank facing 57,000 payroll growth, softer wages than feared, falling oil and still-tight but not obviously re-accelerating inflation has much less reason to rush. Reuters’ market commentary after payrolls captured the point neatly: policymakers now have less impetus to raise rates, and at least some investors are beginning to argue that there may be no hike at all this year. That feels directionally right. The market may still be too aggressive in pricing hikes that never get delivered.

Bottom line

The message this week is simpler than the volatility: crypto needed a bounce, and it got one. But the bounce arrived before the market has fully repaired the two things that matter most: flows and overhangs. ETF demand remains fragile, Strategy remains unresolved, and the macro backdrop has improved enough to help but not enough to declare victory.

Still, the setup is better than it was. On-chain accumulation has improved. More BTC is now held at a loss than at a profit, which historically has mattered. AI leadership has started to wobble just as crypto relative value is approaching major support. Oil is down, payrolls are softer, and rate-hike conviction has faded. That is enough for a relief rally, and perhaps enough for a local low to keep forming. It is not yet enough, in our view, for a durable all-clear.

July is still a seasonally strong month for the Nasdaq, but momentum itself often struggles even when the broader index does fine. If that seasonal pattern holds, then the path of least resistance may be one where leadership broadens, AI cools at the margin rather than collapses, and crypto gets room to mean-revert from an oversold, flow-starved position. That is the constructive case. The cautious case is that one green week is still only one green week. Both can be true at once.

Share

Subscribe now

The result is that the market feels closer to a low than it did a fortnight ago, but not cleaner. Bitcoin is hovering around the $60,000 area after extending its 2026 drawdown, while broader risk appetite remains highly selective. BTC finished the week -7.1%, ETH -9.7% and SOL -3.8%. This is not the indiscriminate risk-off of a panic, nor is it the broad-based sponsorship you want to see at the beginning of a fresh crypto up-leg. It is a market still trying to absorb forced selling, tighter financial conditions and a funding squeeze in one of its most important bellwethers.

Crypto

Flows remain the missing confirmation. Since inception BTC ETF has now had 7 weeks of consecutive outflows, with last week alone being the second biggest in its history of -$1.78B.

Within the top 5 the last 7 weeks has manage to take the top 3 stops and cummulatively this 7 week window has led to over $7.7B of outflows. That is not just a sentiment indicator. It tells you institutional sponsorship has not returned in size, and in a market like this that means every other overhang matters more. When the ETF complex is absorbing redemptions instead of fresh inflows, price becomes more vulnerable to treasury de-risking, cross-asset reallocations and funding stress elsewhere in the ecosystem.

That is why we still do not think the sell-off is over. The market may be close enough to a local low to produce sharp squeezes, but a durable low usually comes with either a clear macro easing impulse or a visible return of sponsorship.

We have neither. The tape has become oversold enough for rebounds, the market may be trying to form an early accumulation structure near the high-$50,000s, but it has not yet produced the kind of sign of strength that usually confirms the range is changing character. For now, it still looks more like an attempted base than a completed one.

That also fits the broader intermarket picture. Crypto is no longer trading as a self-contained narrative asset. It is trading as a flow-sensitive, macro-sensitive risk asset that competes directly with the AI complex for capital. Semiconductor ETFs have absorbed more than $21 billion of year-to-date inflows while bitcoin ETFs have been bleeding capital. That helps explain why crypto can look cheap on its own chart and still fail to rally: it is not enough for something to be oversold if capital is still rotating aggressively elsewhere.

The more constructive read is relative value. TOTALES (Total crypto market ex stablecoins) vs Nasdaq is now close enough to long-term support that the asymmetry is improving. Not yet a trigger, not yet a conviction trade, but a level worth respecting. If that ratio begins to stabilise while ETF selling slows, crypto-versus-AI rotation becomes a serious conversation rather than a theoretical one.

Macro

Macro still sits on top of the stack. The Federal Reserve held rates at 3.50%–3.75% in June, but the message markets heard was unambiguously hawkish.

• May payrolls increased by 172,000

• Unemployment held at 4.3%

• Retail sales rose 0.9%

• First-quarter GDP was revised up to 2.1%

• Weekly jobless claims fell to 215,000.

At the same time, the Fed’s preferred inflation gauge rose 4.1% year on year in May, with core PCE at 3.4%. That is not a recessionary mix, and it is not weak enough to force the Fed into comfort. It is a combination that leaves policy restrictive and keeps rate-hike risk alive.

The market’s reaction has been meaningful. Currently pricing in a 30% chance of a rate hike in July, but higher probability of hikes in Q3/Q4.

A Reuters poll of economists, however, still found the modal base case was no further hikes in 2026. That gap is important. It means financial conditions have tightened more through communication and repricing than through realised policy. In effect, the market is doing some of the tightening for the Fed.

We do not believe the current degree of hawkishness is necessarily durable. The markets have rewarded Warsh’s inflation posture, with long-term break-evens falling sharply since mid-May. But if inflation starts to fall, and especially if the energy impulse continues to fade, we suspect the tone to soften. The simple point is that the market is still acting as if second-round inflation effects are the main danger. Our bias is that once first-round energy effects fade, the appetite to keep talking like a hawk will fade with them.

There is a useful historical analogy here. Reuters’ review of Powell’s tenure reminded markets that Powell’s October 2018 “long way from neutral” remark tightened conditions sharply, before he pivoted weeks later to saying rates were “just below” neutral. New Fed chairs do sometimes feel the need to establish anti-inflation credibility early. When conditions tighten fast enough, that posture can change quickly. That does not mean Warsh is about to turn dovish. It does mean the market may be extrapolating too much from the first act of a new chairmanship.

There is another reason the long end still matters even if the front end stabilises: AI financing demand. SpaceX sold $25 billion of bonds after generating an order book of roughly $85 billion, while the BIS warned that debt-funded AI investment and fragility in core bond markets are becoming more tightly linked. That matters for crypto because even a softer oil tape does not automatically deliver easier long-end conditions if a huge capital-spending cycle is competing for debt-market capacity at the same time. In other words, macro relief can arrive more slowly than bulls expect.

Strategy and STRC

The deeper crypto-specific overhang is still Strategy. The company remains the largest corporate holder of bitcoin and has continued to buy selectively, with holdings reaching 847,363 BTC by late June. But the market is no longer focused only on the size of the bitcoin stack. It is focused on the cost and durability of the capital structure funding that stack.

That is why STRC matters so much, which closed at $74 on Friday, well below its $100 target trading level, sending the effective yield into the low- to mid-teens.

Another step-up in the dividend rate is likely, from 11.5% to at least 11.75%, and potentially higher if pricing remains weak, on about $10.5 billion outstanding. The current annual preferred dividends is at around $1.7 billion. That is not a side plot. That is the cost of capital story now sitting on top of the bitcoin story.

What Strategy has done in response is revealing. The company has continued to issue common equity, but the latest fundraising was used far more to rebuild cash reserves than to add meaningfully to BTC. Around $300 million of recent issuance went into the reserve, taking cash to roughly $1.4 billion. That tells you management is prioritising liquidity and optionality. Rationally, that makes sense. But for the market it changes the read-through. Strategy is not currently functioning like a relentless marginal buyer. It is functioning like a levered bitcoin treasury trying to protect the plumbing of its own funding model.

This is why we think bitcoin will struggle to print a durable low until there is greater clarity around Strategy and its BTC. A forced sale is not our base case today. The balance sheet is not there yet. But the market is trying to price the path by which one could become conceivable: the common loses more of its premium, the preferred remains below par, issuing more common becomes increasingly dilutive, bitcoin purchases shrink further, and any renewed pressure on reserves or redemptions raises the spectre of BTC monetisation. Whether that ends in an actual sale or merely in the fear of one, it is the overhang that still needs to clear.

That is also why a genuine capitulation event linked to Strategy could be the type of washout that finally produces a cleaner low. Not because it would be healthy, but because it would resolve uncertainty. Markets often bottom not when news becomes good, but when the largest unresolved risk becomes concrete enough to discount fully.

Bottom line

The message this week is simpler than the tape. The market is trying to build a low, but it has not yet earned the right to trend. ETF flows remain poor, macro remains restrictive, and the Strategy overhang remains unresolved. That is enough to keep us cautious even though positioning, sentiment and price all look much more washed out than they did earlier in the month.

The constructive case is not hard to see. Warsh may not stay this hawkish if inflation rolls over. Oil has already shown it can fall quickly on any genuine easing in Middle East tensions. Relative crypto-versus-tech valuations are becoming more interesting. And BTC itself is close enough to major support that the downside from here is more asymmetrically two-sided than it was from much higher levels.

But that is not the same as a clean bottom. For that call to improve materially, we would want to see outflows stop, the Fed’s effective hawkishness fade, the long end calm, and evidence that the STRC problem is being solved rather than merely financed around. Until then, the right posture is still the same as last week: respect the bounce when it comes, stay alert to relative-value opportunities, but do not confuse a technical reversal with cleared downside.

Share

Subscribe now

The point, up front

Two technologies are reshaping the economy at once: blockchains, which give the world a native way to move and own value without intermediaries, and artificial intelligence, which gives software the ability to reason and act. Archimed exists to invest where these two forces meet and, more specifically, where they create durable value rather than merely excitement.

Our founding conviction has not changed since 2023. Crypto is building a parallel financial system that is global, open, programmable and permissionless, the TCP/IP for money and we are still early on its adoption curve. What has changed is that AI now sits alongside it as the other defining technology of the decade, and the two are converging in concrete, investable ways. The crossover is not just on-chain plumbing for the future: there is a crypto-adjacent AI stack above all the energy and data-center capacity that crypto miners have spent a decade building that is ready to scale AI activity today. AI needs exactly what crypto provides: scarce power and infrastructure now, and a neutral, always-on settlement and verification layer as autonomous software learns to transact.

But being right that crypto and AI are huge is not the same as making money from them. The two are easily confused, and that confusion is where most investors lose. Our discipline is simple and runs through everything below: we buy value accrual, not narrative. We want assets where real economic activity flows back to the token through fees, and we avoid the parts of the stack engineered to get cheaper and more abundant every year. In a gold rush, the reliable fortunes are made selling picks and shovels and owning the scarce ground not panning for gold. That is how we are positioning Archimed.

The vision that still holds: From centralized control to decentralized value

The early internet promised universal connectivity and digital equality. Over time, centralized platforms came to dominate it, concentrating power, extracting value, and deciding what innovation was allowed. Blockchains push the other way. By distributing control across a global user base, they return the internet toward its open roots, and they do something the early web never could: they let value itself move natively online, without a trusted intermediary in the middle.

The core innovation is embedded trust. Intermediaries; banks, brokers, clearing houses, platforms, have long acted as the gatekeepers of trust, and in doing so concentrated knowledge, cost and control. A blockchain encodes trust directly into transparent, verifiable, distributed technology. Records become auditable, settlement becomes final, and rules are enforced by code rather than by a counterparty you have to trust. Removing fallible intermediaries lays the foundation for commerce and coordination built on embedded trust rather than brokered trust, a shift whose implications unfold over decades, not quarters.

This is also what makes a genuinely global, real-time economy possible: digital dollars that settle worldwide in seconds, markets open every second of every day, and capital that can reach anyone with a connection. Much of the 2023 vision borderless value transfer, individual ownership of one’s financial and digital life, open and composable applications, has only become more credible. Stablecoins now move trillions, on-chain markets clear billions a day, and real-world assets are beginning to migrate on-chain. The Internet of Value is no longer a forecast; it is under construction.

Usage is not value capture: The lesson history keeps teaching

Here is the trap that catches most technology investors. A technology can transform the world and still ruin the people who invested in the obvious way. The internet build-out of the late 1990s is the cleanest example: the technology was every bit as transformative as promised, vast amounts of fibre were laid for the coming wave of data, the wave arrived , and most of those companies still went bankrupt, because they had built far more capacity than was needed and it sat unused.

It gets sharper. Over the following two decades mobile data traffic grew more than a thousand-fold. If you had known that in advance and bought the network operators, you would have made almost nothing, they spent constantly to carry the traffic while competing on price. The fortunes were made by the platforms that owned the scarce, defensible layer on top.

That single lesson sits at the centre of how we invest.

Crypto has its own version of this, and it is the most important structural fact in the asset class today: the cost of blockspace is trending toward zero. A cycle ago, transaction capacity was scarce and expensive. Today, after years of scaling and dozens of competing chains and rollups, blockspace is abundant and cheap and getting cheaper. Generic blockspace, generic yield, and undifferentiated chains are commoditizing exactly the way AI models are commoditizing as their cost falls roughly ten-fold a year. Building a thesis around ‘which chain is fastest’ is like building an AI thesis around ‘who has the best model’: you are anchoring to the layer engineered to commoditize. Value does not disappear, it migrates upward, to whatever stays scarce and to whatever can actually charge a fee.

Where value accrues: Fees flowing to the token

If blockspace and raw capacity are heading to zero, the durable returns belong to assets that capture real economic activity and route it back to token holders. This is the heart of our work, and we apply a consistent test: does the protocol earn genuine fees from real usage, and does the token have a credible claim on those fees through burns, buybacks, revenue share, or staking rewards funded by actual economic activity rather than inflation?

We look for several scarce, fee-generating positions in particular:

• Settlement and credible neutrality. The deepest moat in crypto is the trust that a network will not change its rules and cannot be captured. Only a handful of chains have earned it through years of survival, and on those, demand for blockspace and the asset’s role as collateral and store of value drive real value accrual even as raw capacity gets cheap.

• Protocols with real cash flows. Exchanges, lending markets, derivatives venues and asset-issuance protocols that earn fees from genuine volume and increasingly direct those fees to the token. A protocol that earns and shares real revenue is worth more than a hundred that merely promise future utility.

• Deep, sticky liquidity. Liquidity begets liquidity. The venues where real volume already lives enjoy a flywheel newcomers cannot easily break, and they monetize it through fees.

• Stablecoin issuance and distribution. The right to issue a trusted dollar on-chain, and the rails to distribute it, is becoming one of the most valuable franchises in finance, scarce, increasingly regulated, and highly cash-generative.

• Trusted data and verification. As more value and more AI activity move on-chain, feeding reliable real-world data into contracts and verifying that an output or asset is what it claims to be becomes scarce and monetizable.

The principle uniting these is simple: we want to own the parts of the stack where economic activity is rising and a fee can be charged on it, and where the token has a real claim on that fee. Growth in usage only makes us money if value accrues to the asset we hold.

The growth still ahead

None of this discipline implies caution about the size of the prize. The long-term addressable market remains vast, the hundreds of trillions of dollars of assets, transactions and economic activity that could benefit from moving onto open, programmable rails. Crypto continues to trace a classic adoption S-curve, with the steepest part of the climb still ahead as the technology crosses from early adopters toward the mainstream, much as the internet and the smartphone did before it.

The drivers are now concrete rather than speculative: stablecoins becoming default settlement for global payments; tokenization of treasuries, funds and private assets bringing real-world value on-chain; clearer regulation pulling institutional capital in; and - newly - AI agents that will need to transact natively in digital value. We expect adoption to be powerful but uneven, arriving faster in finance than elsewhere, and we plan around that. The direction of travel is not in doubt; the timing and the winners are where the work is.

The crossover: The crypto-adjacent AI stack

The most important development since our first thesis is that crypto and AI have stopped being separate stories. But the convergence we find most investable today is not the speculative one. It is a crypto-adjacent AI stack that can scale AI activity right now, anchored by the single scarcest ingredient in AI: power. The lesson from the AI build-out is that the binding bottleneck has moved from chips to electricity, the developed world cannot build power and data-center capacity fast enough to feed planned AI demand. Whoever already controls that scarce capacity captures the value, and crypto operators turn out to be sitting on it.

The power play: Miners pivoting to AI. For over a decade, crypto miners did the unglamorous work of securing cheap power, land, grid interconnects, substations, cooling and round-the-clock data-center operations. That is precisely the asset AI now needs most and cannot quickly reproduce. Many of these operators are now converting energy-rich sites from proof-of-work mining to AI and high-performance computing. This is not a someday story: they have power and ready-to-energize shells today, and are signing that capacity to AI demand now, with real contracts, real revenue and hard assets behind it, not tokens. This is an area we understand particularly well, and we view it as the most concrete, near-term expression of the crossover: owning the energy and infrastructure bottleneck rather than the AI model that runs on top of it.

Compute and infrastructure ready to scale. Beyond miners, a broader crypto-adjacent infrastructure layer, data-center operators, GPU-backed compute providers, and the public companies bridging crypto balance sheets and energy assets into AI compute is positioned to monetize AI demand today. These businesses give us exposure to the AI build-out through cash-generative, asset-backed operators rather than richly priced headline names.

Around this near-term core sit the on-chain rails the crossover will increasingly run on, which we sort honestly by maturity:

• Already working settlement for machine commerce. Stablecoins and on-chain payments are the natural settlement layer for machine-to-machine and AI-agent transactions, because they are programmable, global and available every second of every day in a way bank rails are not. An economy of autonomous agents needs money that software can hold and move without a human in the loop, and that demand routes real fees to the rails.

• Real but earlier: data, provenance and verification. On-chain provenance and verification for AI’s outputs and training data, and tokenized markets for data, are a genuine and monetizable need. The businesses are real but younger than the power-and-compute layer.

• Frontier, not yet ready to scale: decentralized inference and compute. Decentralized compute and inference networks are compelling and worth positioning ahead of, but they cannot yet absorb AI activity at the scale the centralized, power-rich operators can today. We treat them as a frontier, not the near-term trade — the distinction the market most often gets wrong.

• Mostly narrative: AI agent tokens. The broad basket of ‘AI agent tokens’ and autonomous on-chain funds runs well ahead of durable value. Useful agent infrastructure is realistic this decade; most of today’s AI tokens are not the way to own it.

Our stance follows the same discipline as everywhere else in this thesis: own the scarce, ready-now layers first: power, infrastructure and settlement, position ahead of the frontier, and avoid the narrative basket. The crossover only deepens our core conviction: crypto is not a competitor to the AI age, it is one of its foundations, and in the case of power and infrastructure, one of the few parts of that age ready to scale today.

Our Strategy

Archimed is a long-biased fund focused primarily on liquid crypto assets, with a disciplined allocation, up to 15%, to selected illiquid and early-stage positions where the asymmetry justifies the risk. We take a top-down view of market cycles and liquidity to decide how much to bet and when, and a bottom-up, fundamental view to decide what to own. The two combine into a portfolio built to capture the upside of the on-chain economy through the full cycle.

We lean toward the scarce, fee-generating, often less-exciting parts of the stack: settlement layers with credible neutrality and real demand for blockspace, protocols that earn and return genuine cash flow, stablecoin and tokenization rails, deep-liquidity venues. An area we understand particularly well and believe is under-owned is the crypto-adjacent AI stack, above all the power and data-center capacity that miners and infrastructure operators can scale into AI demand today. We lean away from the most crowded and fragile parts: undifferentiated chains competing on speed and subsidies, narrative tokens with no value capture, opaque and leverage-heavy corners exposed when the cycle turns, and the speculative ‘AI token’ basket where the story runs ahead of the substance.

Our research process is unchanged in spirit and sharper in execution. We build token valuation models grounded in real fees and cash flows; we track on-chain analytics for genuine usage rather than incentivized activity; and we study teams, governance, token distribution and catalysts closely.

Alongside the liquid book we run crypto-native yield strategies staking, lending, liquidity provision, to keep capital productive and generate returns less correlated to price. For early-stage, we keep a sniper’s discipline: a small number of high-conviction positions in the scarce layers above, sized for the risk, often alongside top-tier partners. A few early winners can meaningfully lift the whole fund.

Why be different at all? Because we are a focused fund, not a giant. We cannot win by trading faster or spending more than the largest players, and simply owning what they already own means buying yesterday’s winners at today’s prices. Our edge is a clear, defensible view of where value in crypto and AI will actually accrue and the discipline to buy the scarce, fee-generating foundations before the crowd.

The honest risks and what would change our minds

We would rather state our risks plainly than have you discover them later. The central risk in crypto remains leverage and opaque counterparties: a disappointment, a fraud, or a liquidity shock tends to ripple through the financial plumbing of the industry before it shows up in fundamentals, and it hits the over-leveraged and the opaque hardest. This is exactly why we favour transparent, cash-generative protocols. Crypto also remains volatile, sensitive to global liquidity, and exposed to the chance that regulation turns hostile in a key jurisdiction.

We are also honest that we could be wrong about the big picture. We would revisit our thinking if any of the following happened:

• Adoption of real-world assets and institutional crypto accelerated dramatically faster than we expect, justifying a harder lean into the build-out and earlier-stage bets;

• A single chain or platform achieved such durable dominance that credible neutrality stopped being scarce and became a winner-take-all monopoly;

• The crypto-AI convergence produced genuine, revenue-generating businesses far sooner than we assume;

• Or, conversely, regulation turned sharply hostile in a major market, or stablecoins suffered a confidence shock — which would undercut the rails we believe are most durable.

We will watch these honestly and tell you if our view changes. We would rather update in the open than defend a position we no longer believe.

The future is still decentralized

We remain in the opening chapters of a transformation that will reshape finance and beyond. Open blockchain protocols are becoming the backbone of a more transparent, accessible and innovation-driven economy; tokenized value lets individuals and institutions transact without depending on centralized intermediaries; and a new layer of intelligent, autonomous software will increasingly run on these same open rails. Crypto and AI are not separate revolutions, they are two halves of the same shift toward systems that are open, verifiable, and owned by their participants rather than their gatekeepers.

The one idea we ask you to weigh is the one this document exists to make plain: blockspace and raw intelligence are becoming cheap and abundant, and the lasting profits will belong to whoever owns the scarce things the on-chain economy depends on and the assets where real fees flow back to the token. Everything we do follows from that belief.

Drawing inspiration from our namesake Archimedes, give us a lever and a place to stand, we intend to apply our capital where it has the most durable effect. The window of generational opportunity is open. We are here for the long haul, and the future of finance is decentralized.

Team Archimed

That matters because the market has repriced the Fed higher. After the June meeting, futures shifted towards a meaningful chance of a hike by October and a still larger probability by December, while Warsh also launched a review of Fed communications that investors fear could reduce forward guidance and push term premia higher. In other words, even if headline oil prints cool for a few sessions, financial conditions are not cleanly easing. They are becoming more conditional, more volatile and more dependent on the long end.

Crypto

BTC traded like a market trying to bounce before it has earned the right to trend. Early in the week the tape improved as news of a U.S.-Iran framework deal and a drop in oil lifted risk appetite, pushing bitcoin back into the mid-$66k area and briefly above $67k intraday. But that move faded as traders turned their attention back to the Fed, weak institutional sponsorship and the reality that a hawkish policy surprise is still a stronger driver of crypto multiples than one good geopolitical headline. By end of the week we closed around 64k.

The BTC structure remains a rebound inside a damaged range, not a repaired trend.

Likewise, on the broader tape, the TOTALES, Total Market cap ex-stables, looks like a market leaning on structural support rather than launching into fresh impulse. The setup is better than it was at the panic lows, but it still resembles a technical reversal more than a final low. The market has become less fragile, but it has not become safe.

Flows are the missing confirmation. Bitcoin ETFs continued to bleed this week and have now posted six consecutive weeks of net outflows, the longest stretch since inception with total outflow of $5.9B. The pace is slowing, which matters, as capital rotations has shifted towards specualtive capital in AI and megacap equity stories. We have enough for a tradable bounce: support held, momentum is oversold, and forced selling looks less aggressive. What we do not yet have is the kind of sustained flow recovery that normally sits underneath a fresh uptrend. Until ETF outflows stop and turn back into consistent inflows, the bounce is real, but still unconfirmed.

Strategy helped at the margin, but only at the margin. After unnerving the market with its earlier bitcoin sale, the company disclosed that it had bought 1,550 BTC for about $101.3 million at an average price of $65,332, bringing holdings to 845,256 BTC, while also increasing cash reserves to $1 billion. That removed the immediate fear that the company had become a persistent marginal seller into weakness. But the size of the purchase was small by Strategy’s own recent standards, and it did not change the bigger story: Strategy is now managing capital structure stress as much as it is managing bitcoin accumulation.

Macro

Macro was supposed to get easier last week. The script was straightforward: a softer oil tape after the U.S.-Iran framework, weaker inflation impulse, lower long yields, a calmer Fed and some breathing room for crypto. Instead, we got the first half of that script for about forty-eight hours, and then the Fed took over. U.S. retail sales for May rose 0.9%, the labour market remained stable with 172,000 payrolls added and unemployment at 4.3%, and inflation stayed too hot for comfort, with May CPI at 4.2% year on year and core CPI at 2.9%. That is not recessionary enough for easy cuts and not disinflationary enough for complacency.

The Fed’s response reflected that mix. The committee held rates steady at 3.50%-3.75%, but its projections and market interpretation were hawkish: policymakers leaned towards at least one hike this year, short-dated Treasury yields jumped, and the market moved to price a much higher chance of a year-end increase. Warsh’s emphasis on rethinking communication, including scepticism around the dot plot and forward guidance, only added to the sense that the easing bias is gone. The market is no longer debating when cuts arrive. It is debating how high the Fed’s tolerance for sticky inflation has become.

That reset matters for crypto because it changes which yield matters most. The front end still anchors policy expectations, but the long end sets mortgage rates, discount rates and refinancing conditions. There was a brief reminder of that early in the week when cooling Iran-war fears pushed the 10-year Treasury yield lower and the average 30-year mortgage rate down to 6.47%. Yet the relief was fragile. As U.S.-Iran talks frayed again into the weekend and Iran threatened renewed closure of Hormuz, oil downside lost credibility and one of the easiest channels for a disinflationary macro repricing started to close.

Our macro conclusion is unchanged in spirit but updated in emphasis. The downside is not cleared, even if we get a temporary macro repricing lower in yields. Oil can still reprice higher if diplomacy fails. The Fed has clearly shifted hawkish. And there is a new structural force now operating on the long end: AI funding demand. That is where this week’s deeper story sits.

Strategy and STRC

The cleaner single-stock story this week is STRC. Strategy’s Stretch preferred was designed to trade near par by adjusting its dividend, but the mechanism is now working against the issuer. The outstanding size is about $10.5 billion, the dividend has already been raised from 9% at launch to 11.5%, and even with that support the security has recently traded around 87-89, implying a market yield closer to 13%. Multiple reports this week argued that the formula points to another increase, at least to 11.75% and potentially towards 12% or higher if prices do not recover.

That leaves Saylor with a narrow menu. Continuing to issue STRC at a discount is unattractive. Selling common equity can still raise cash and buy more BTC, and last week’s issuance did exactly that, but it does not fix STRC directly; it mainly shifts strain onto the common by diluting holders while topping up the dividend reserve. Strategy’s recent actions, using common issuance to buy bitcoin and add cash, and increasing the reserve to $1 billion, tell you the company is prioritising liquidity and optionality, not resolving the preferred overhang.

Our view is that there are really only three credible paths from here. One is to keep lifting the dividend and hope the market eventually pays par again, which protects credibility but raises the carrying cost of the entire capital structure. The second is to buy back or tender for STRC below par, which is economically attractive if Strategy can use a still-elevated common equity multiple to fund it. The third is to keep selling common stock while the mNAV premium exists, building cash until the company can more aggressively defend or retire the preferred. Of those, the buyback/tender route looks the cleanest economically, the dividend hike route looks the most credible optically, and pure common dilution looks like the default bridge, not the endgame.

That last point is the important one: diluting common to buy more BTC may support the bitcoin narrative, but it does not solve the STRC problem on its own. That is an inference from the structure rather than a stated company plan.

Bottom line

The weekly message is simple. Last week’s “technical reversal, not final low” still stands. Crypto stabilised, but it did not break free. Macro got a brief oil-led reprieve, but the Fed replaced it with a hawkish repricing. The deeper medium-term issue is that AI capex is no longer just an equity valuation story; it is becoming a capital-markets story that can keep the long end firmer than bulls want.

So the right posture remains the same as last week, just with a sharper macro lens. Respect the bounce. Be open to relative crypto-versus-tech opportunities if that ratio keeps leaning on long-term support. But do not confuse a tactical reversal with cleared downside. We still do not think the downside is cleanly behind us, even if yields retrace a bit and oil stays softer for a few sessions. For that call to improve materially, we would want a calmer long end, less hawkish Fed pricing, cleaner ETF participation, and evidence that the STRC problem is being solved rather than financed around.

Share

Subscribe now

The clearest improvement from last week is Strategy. After unnerving the market with its first BTC sale in years, the company disclosed that it bought 1,550 bitcoin for about $101.3 million at an average price of $65,332, lifting holdings to 845,256 BTC. The same filing also said cash reserves were increased by $100 million to $1 billion.

In other words, one obvious overhang has shifted from “maybe marginal seller” back toward “still buying when the balance sheet can support it”. That is constructive, but it is not a sufficient condition for calling the low.

Crypto

The bullish case for last week was straightforward. BTC flushed below $60,000 on 5 June, leverage got carried out, and the market was ripe for a rebound.

BTC ETF flows remain weak. Last week saw another -$319M of net outflows. That is a much slower pace than the prior weeks, which is constructive at the margin, but it is still an outflow week. More importantly, it marks the fifth consecutive week of negative BTC ETF flows, the longest negative streak we have seen so far. The cumulative number matters too. Over the last five weeks, BTC ETFs have seen roughly $5.076B of outflows. That makes this the second-largest five-week outflow window on record, behind only the March 2025 stress period, when outflows reached roughly $5.4B.

That is why we treat the bounce as technical first. The story is that sell pressure slowed enough for an oversold market to bounce, with sentiment still hurt by Strategy’s earlier sale, weak ETF demand and the continuing appeal of equity markets. On the price side, your BTC chart is saying something similar: the market has stabilised around the low-$60,000s, but it has not yet reclaimed enough territory to say the trend has repaired. This is exactly what a bottoming process tends to look like in the middle, not at the end.

The liquidation backdrop also supports that reading. The sell-off was large enough to clear weaker hands, with reports of roughly $1.77 billion in crypto liquidations during the worst of the move. That helps explain why the rebound was sharp: once the weak longs are gone, the market does not need much good news to lift. But sharp relief does not mean the market has earned the right to trend yet. It means the market is less fragile than it was one week ago.

The same logic applies to the broader market.

TOTALES ex-stables bounced from an important support band around the $1.8T area. That level matters because it sits near the prior structural breakout region. If it holds, the market can keep building a bottoming structure. If it fails, the next clean level is much lower.

So the update on crypto is simple. The market was oversold enough to bounce. Strategy buying again removes one near-term fear. But the flow picture still looks tactical rather than structural, and that keeps us in the camp of buyable reversal, not confirmed bottom.

AI versus crypto

SpaceX continues to have all eyes this week. The IPO raised $75 billion by selling 555.6 million shares at $135, valuing the company at about $1.78 trillion. The stock gained roughly 19% on their first day, pushing the valuation to around $2.1 trillion. However, digging a bit deeper, this is what in crypto is called a low float, high FDV: the raise at pricing represents only about 4.3% of the implied value, and versus the first-day valuation it is closer to 3.6%. That helps explain why the stock could trade higher across the day without telling you much about long-run clearing demand.

More importantly, SpaceX is not a one-off. Anthropic recently raised $65 billion at a $965 billion valuation and has confidentially filed for an IPO. OpenAI has confidentially filed and could target a valuation of up to $1 trillion, with its latest prive around raised $122B at a $852B valuation. Put those together with SpaceX and the capital call is massive: about $262 billion of reported or recently committed capital and roughly $3.6 trillion of headline value even before assuming a full $1 trillion OpenAI IPO mark. Morgan Stanley, meanwhile, says global AI-related debt issuance could top $570 billion in 2026. This is not background noise. This is a competing sink for global risk capital.

That is why the “AI versus crypto” framework still works. The SpaceX deal and the AI pipeline to liquidity being drawn away from crypto, and the crypto tape fits that explanation. But the relative setup is getting more interesting. Looking at TOTALES/Nasdaq chart, crypto versus tech is back on long-term relative support and looks oversold.

Nasdaq is sitting between the 20-day and 50-day moving averages. Break either level and trend should follow. If AI hype unwinds slowly, that relative setup could become a positive tailwind for crypto.

If we instead get a broader risk-off move, the correlation of BTC/Nasdaq chart, now back around 0.5 after negative readings late in May, is a reminder that crypto may go lower first before it outperforms later.

Macro

Macro is still the brake. The May U.S. jobs report showed nonfarm payrolls rising by 172,000, with unemployment unchanged at 4.3%. Then came firmer inflation prints. CPI rose 4.2% year on year in May, with core CPI at 2.9%. Producer prices were even hotter: headline PPI rose 6.5% year on year, and final demand less foods, energy and trade services rose 5.1%. That is not the recipe for an easy liquidity tailwind into risk assets. It is the kind of mix that keeps the market nervous about rates staying restrictive for longer.

The next FOMC meeting this week on 16–17 June, and the Federal Reserve calendar marks it as one of the meetings accompanied by fresh Summary of Economic Projections. Market is currently pricing unchanged rate path fir 2026, with a slight chance for potential hikes.

If Chair Warsh leans hawkish into sticky inflation, it is hard to see crypto getting the macro clearance needed for a durable low just yet.

The swing variable is energy. Reuters reported today that Iran says a draft memorandum with the U.S. would include immediate reopening of the Strait of Hormuz and temporary oil-sanctions relief. If that de-escalation sticks, one major inflation impulse could cool. If it does not, energy remains an upside risk to inflation and yields. For crypto, that means the macro tape is not outright hostile in one direction only; it is conditional. A softer energy backdrop could help. A renewed oil squeeze would probably hurt everything speculative, including crypto, before any relative rotation story can play out.

Bottom signal checklist

The house checklist still says bottoming process, not confirmed low. These are some of the things we look at:

Twenty-five delta skew. This is really the price of downside insurance versus upside. You want fear to have shown up, but you also want to see it stop worsening. Right now the message feels more like defensive but easing than fully reset

Perp funding negative. This is one of the cleaner washout tells because it shows frothy long positioning has finally been punished. We have seen enough stress in the move for that box to have flashed. That is constructive.

Quarterly basis. You want basis to compress hard and ideally flirt with inversion. We are clearly closer to exhaustion than to froth, but not yet in the sort of durable backwardation that usually shouts final low.

Stablecoin APR on Aave. We want cash-on-chain yields to be boring, not euphoric. The market is moving towards that low-APR zone, which is quietly positive. It says leverage demand is cooling

Term structure. Near-dated futures over far-dated futures is a genuine stress signal. We are much closer to flat than to healthy bull-market contango, but not yet at the type of inversion that would make me say the job is done.

Open interest divergence. If price is falling and OI is rising, new bearish positioning is still entering the market instead of the move being fully exhausted through liquidation. That keeps the structure looking more like redistribution than a final capitulation.

Total liquidations. The flush was long-led, which is what you want to see into a local low. This is one of the stronger constructive tells on the board. It supports the bounce call. It does not, by itself, prove the ultimate bottom is in.

Net-net, the checklist is moving in the right direction, but only some boxes are properly ticked. Enough for a tradeable reversal. Not enough for the all-clear.

Bottom line

We expected a bounce and we got one. But this still looks more like a technical reversal than a final bottom. Strategy buying again removes one fear. SpaceX’s blockbuster debut and the Anthropic/OpenAI pipeline keep AI as a competing sink for capital. Macro is still awkward rather than supportive. That keeps the right framing as crypto is getting attractive on a relative basis before it is fully safe on an absolute basis.

What would make the call cleaner? A calmer skew, a clearer funding washout, more obvious front-end stress in futures, and a Fed outcome that does not tighten financial conditions by rhetoric alone. Until then, the better posture is probably to respect the bounce, keep leaning into relative crypto-versus-Nasdaq attractiveness when the setup is there, and keep describing the market honestly: bottoming process, not confirmed bottom.

Share

Subscribe now

Total crypto market cap fell 12.8%. BTC was down 13.7%, ETH down 19.3%, and SOL down 20.8%. Total market cap now sits around $2.11T, down 52% from the October ATH and 30% from the 2021 ATH. Ex-stables, the drawdown is even worse: -56% from October and -40% from the 2021 high. The reason this matters is simple: TOTALES is now sitting directly on support.

If this level breaks, the next clean support is around $1.4T. That would take us back to November 2023 levels and effectively reset the entire post-2023 crypto beta trade.

The theme this week is not “all clear”. It is bottoming process. The market is oversold enough for violent reflex bounces. BTC and SOL are pinned to the floor of their one-year value areas, RSI is washed, and reversal trend is flashing. This is the type of zone where bounces usually start. But bounces are not bottoms by default.

The macro tape is still hostile, flows are still poor, and the institutional bid that carried much of this cycle has clearly weakened. We are no longer in the middle of the range. We are at the floor. That does not mean the low is in.

It means the market is finally trading where risk/reward starts to change.

Crypto

The most important thing in crypto this week was not just the price action. It was the flow picture. One distinction matters. U.S. spot bitcoin ETF outflows worsened from roughly $1.0bn in the week of 11–15 May to $1.26bn, $1.42bn and then $1.72bn in the week of 1–5 June. That is not noise. That is persistent de-risking.

The structural bid is no longer absorbing supply in the same way. For most of this cycle, the ETF bid was the cleanest expression of demand. When that bid turns into repeated weekly outflows, the market starts to feel heavy. Not because there are no buyers, but because the marginal buyer has stepped back.

Strategy is the other pressure point. The 30 May / 1 June disclosure showed that between 26 and 31 May, Strategy sold 32 BTC for roughly $2.5M. It also sold around 801,994 MSTR shares for $128.3M in net proceeds, issued no preferred stock during the period, and ended May with 843,706 BTC plus a $900M USD reserve.

So the answer to the most important question is nuanced. Capital markets were still open to the common-equity ATM. But the BTC treasury itself was a net source of cash at the margin, not a buyer. That matters because the entire market has spent years treating Strategy as a one-way structural bid. Sell equity. Issue preferred. Buy BTC. Lift NAV. Repeat. This week showed something different.

The sale was tiny in size, but large in signal. If Strategy sold more, the move makes sense. The market is pricing a shift from permanent buyer to potential marginal seller. If they managed to tap capital markets and buy more, that is neutral. If they were buying while BTC still sold off this hard, then flows are strongly against BTC.

STRC is where the stress is showing up.

Strategy’s STRC now trades below par, with the latest price around $93.40, an effective yield of 12.31%, and a variable annual dividend rate of 11.50%. This is a security designed to trade around the $100 stated amount. When something engineered to hug par is sitting materially below par, the market is telling you the funding flywheel is under strain.

That does not mean crisis. It means the marginal source of easy crypto-linked financing is less powerful than it was.

Technically, the setup is much cleaner than the tape feels.

BTC, ETH, SOL are both pinned to the floor of their one-year value areas. BTC is around $60k. ETH around xxx SOL is around $61. RSI is in the mid-teens fresh trend reversal setups. This is exactly the kind of place where reflex bounces start.

But oversold is not the same as repaired. All are still below major moving averages. Volume thins quickly below spot. Any near-term strength should be treated as tactical until the market proves otherwise.

For BTC, the map is simple.

Zone 1: $59k–$61k: This is the first accumulation area and the current floor.
Zone 2: $56k–$57.5k: This is where the air pocket starts.
Zone 3: $53k–$55k: This is the capitulation flush / deep-value zone.

If BTC loses $60k and then $59k, the lack of volume support below means the move can accelerate quickly.

For ETH, the support zones are:

Zone 1: $1,500–$1,650: Current support. If ETH is going to stabilize without a full flush, it probably has to do it here.
Zone 2: $1,250–$1,425: The next obvious support pocket if $1,500 fails.
Zone 3: $900–$1,050: The deep-value washout area.

Macro

Macro is not yet giving crypto a clean tailwind. The latest official U.S. inflation data show core PCE at 3.3% year on year in April, up from 3.2% in March, while headline PCE rose to 3.8% from 3.5%. The Dallas Fed’s trimmed-mean PCE measure is still lower at 2.3%, but that gap is precisely why the market remains edgy: if underlying inflation measures keep firming, the disinflation story gets harder to defend. At the same time, Kevin Warsh has just taken office as Fed chair, and the next FOMC meeting is scheduled for 16–17 June.

The growth side of the macro tape also pushed in a hawkish direction. U.S. nonfarm payrolls rose by 172,000 in May, well above expectations, while unemployment held at 4.3%. That is one reason equities sold off hard into the weekend and why Warsh’s first meeting now looks more like a risk event than a relief event. Put simply: inflation is not cooling fast enough, growth is not weakening fast enough, and markets have had to reprice the chance that rates stay restrictive for longer.

That repricing hit growth equities directly. S&P 500 posted a weekly decline after nine straight weeks of gains, while Friday’s selloff was driven by stronger jobs data and a hit to semiconductors after Broadcom disappointed. Broadcom left its fiscal 2027 AI revenue forecast unchanged and missed second-quarter revenue expectations, sparking a broader chip selloff. In other words, the macro backdrop is tight enough that “good but not perfect” is suddenly not good enough for crowded AI winners.

The AI trade

The bear case on AI financials is no longer fringe with key focal points around cost, tokens and monetisation rather than growth. The key point is simple: tokens have become the revenue engine for model providers, but they are also exposing how expensive AI usage is for customers. A growing backlash from enterprise buyers who are burning through token budgets without a clear line of sight to business value, while Google has already begun pitching Gemini as a cheaper option for firms trying to manage spend. ROI on AI investment remain hazy, daily token costs can exceed human labour for some tasks, and boards are likely to impose much tighter financial discipline.

That fits the broader chip-cost story as well. Surging memory prices are now feeding “chipflation”, pushing manufacturers toward harder choices on pricing, margins and component use. Put differently, the cure for very high AI costs is likely to be efficiency, not unlimited spending. That does not kill the AI story. It changes the market’s willingness to pay peak multiples for every name leveraged to it.

SpaceX is the immediate liquidity event hanging over that trade. The IPO has been priced at $135 a share, is targeting a $75bn raise at a $1.75tn valuation, and is expected to be listed 12 June. The deal is 2x oversubscribed. Damodaran’s latest post is a useful counterweight: he argues that the selling machine around this IPO will be intense, that the equity story will be marketed partly as an AI vehicle, and that investors should pressure-test the plausibility of the numbers rather than just buy the narrative.

SpaceX is not the only draw on capital. Anthropic has now confidentially filed for a U.S. IPO, while OpenAI is also preparing to file. The combined demand for capital from SpaceX, OpenAI and Anthropic is likely to be large enough to disrupt flows across public markets.

There is one possible offset. Reports has started to circulate that senior U.S. officials have held preliminary discussions about the government potentially taking stakes in AI companies, and Trump said his team would look into the idea. AI policy headlines might help sentiment at the margin for a few days, but they are not the same thing as a durable fix for stretched positioning and rising cost anxiety. As we’ve stated many times in the past, if/when equities get a stronger correction, crypto wont be able to withhold the further downside pressure.

Why this still looks like a bottoming process

The case for a crypto bounce is becoming stronger, even if the case for a durable low is not there yet. BTC having one of its worst weekly drop since late 2022, while ETF outflows have accelerated and capital continues to been rotating into AI. Bitcoin’s relationship with the S&P 500 has flipped deeply negative as AI equities continue to attract money while crypto lags. That is exactly the kind of dislocation that can produce sharp, tradable mean reversion once the forced sellers are done.

So the working view should be: massively oversold, probably due a bounce, but not obviously the final low. That means patience and staggered spot buying make more sense than swinging for the fences. The confirmation signals to watch are straightforward: skew settling down, heavy volume into support followed by reversal, ETF redemptions easing, and no further unpleasant surprises from Strategy’s treasury or STRC. Until those things improve, treat the setup as a bottoming process, not a confirmed bottom.

What to watch this week

The immediate line in the sand is still BTC $59k–$61k. Hold that, and the market can stage a sharp relief move simply because positioning and sentiment are so washed out. Lose it, and the air pocket into $56k–$57.5k opens fast. The same logic applies to SOL $60–$63, with $54–$56.5 the next destination if support fails, and to ETH $1,500–$1,650, where failure would likely drag the market towards $1,250–$1,425.

The checklist for the week is clear. Watch daily U.S. spot ETF flows. Watch Strategy’s purchases page and any fresh SEC filings for evidence of renewed BTC buying or further treasury sales. Watch whether STRC can stabilise near par or whether it keeps drifting lower. Watch SpaceX pricing and listing developments, and watch whether Anthropic/OpenAI headlines keep pulling speculative cash into AI instead of back into crypto. All of that sits in front of Warsh’s first FOMC on 16–17 June, which is now shaping up as the next major macro hazard.

The bottom line is simple. Crypto is getting close to deep value in a number of important names, but the market has not yet earned the right to call the all-clear. Start buying the better zones slowly. Respect the air pockets. Expect a bounce before a clean trend turn. And if the tape does not improve on flows, be ready for one more flush before the real reset is complete.

Share

Subscribe now

That is the important tension this week. Crypto is maturing into a more institutional market, but it is doing so inside a world where capital has a higher hurdle rate, the U.S. Treasury needs more buyers, and every asset class is competing for balance sheet.

Last week was about the US-Iran process as a rates trade. This week is about something slightly different: who absorbs all the paper? Treasury paper, AI capex paper, stablecoin collateral, crypto leverage, and risk assets all sit inside the same balance-sheet system. When balance sheet is abundant, everything can work at once. When balance sheet becomes more selective, dispersion replaces beta.

Crypto

BTC remains the asset the market trusts most, but it is not trading like a clean breakout asset yet. That distinction matters. It can be the strongest asset in crypto and still not be ready to lead a broad risk cycle.

The reason is simple: BTC is now sitting between two stories. The structural story is strong. ETF ownership, treasury adoption, reduced exchange float, and institutional familiarity all continue to improve. But the tactical story is still messy. When real yields stay elevated and the long end refuses to behave, BTC does not trade like a pure monetary debasement hedge. It trades like high-quality macro collateral with no yield.

That is still much better than being a low-quality altcoin, but it is not the same as being immune.

ETH remains the problem child. It is not that Ethereum has no activity. It is that the market is not paying for that activity right now. ETH still feels like an asset caught between narratives: too mature to trade like a reflexive growth chain, too fragmented to trade like clean monetary premium, and too dependent on lower macro volatility to lead. If BTC is crypto’s collateral asset, ETH is still trading like crypto’s operating system equity. In this environment, collateral is being rewarded more than operating leverage.

SOL remains the opposite. It is easier, cleaner, faster beta. When the market wants risk, SOL still gets reached for first. When the market de-risks, that same simplicity becomes a liability. So the playbook remains tactical rather than structural: SOL can outperform hard in a relief rally, but it is not the asset you want to hide in if yields turn higher again.

The more important update is market structure. Coinbase and Kalshi are bringing regulated perpetual crypto futures to U.S. investors, moving one of crypto’s most important offshore products into domestic regulated venues.

Perpetual futures onchain has increased >200% in both 2024 and 2025 and in 2025 alone about $8.2 trillion of volume alone.

This means crypto leverage is slowly being pulled out of offshore opacity and into regulated collateral systems. The onchain perp market reached as high as 15% of total CEX perps.

The irony is that this happens at the exact moment when leverage is becoming more expensive everywhere else. The bullish read is structural: better venues, deeper liquidity, more institutional comfort. The cautious read is cyclical: regulated leverage still needs balance sheet, and balance sheet is no longer free.

Stablecoins sit in the same bucket. They are no longer just crypto casino chips. They are becoming a shadow demand channel for short-duration U.S. government collateral. Q1 2026 data indicated that Tether holds $141 billion of U.S. Treasury bills in Q1 2025, equal to roughly 1.6% of outstanding Treasury bills, making it one of the largest non-sovereign buyers in that market, and top 17th largest holder of U.S. debt globally.

Stablecoins are not just liquidity inside crypto. They are a bridge between offshore dollar demand, Treasury collateral, and on-chain settlement. If stablecoin supply expands, crypto gets more transactional liquidity and the Treasury gets another buyer. If stablecoin growth stalls, crypto loses one of its cleanest plumbing tailwinds.

Stablecoins have continued to grow even as 3-month bill yields have fallen, which suggests demand is not simply a yield-arbitrage trade. The GENIUS Act gives the sector a clearer U.S. regulatory framework, and Bessent has argued that stablecoins could grow roughly tenfold to around $3tn by 2030, creating a new source of demand for short-duration Treasuries. That makes stablecoins look less like offshore crypto plumbing and more like a structural dollar-liquidity rail: crypto gets a deeper liquidity base, while the Treasury gets another marginal buyer for bills.

Macro

Macro is also becoming less about the Fed meeting and more about the buyer of last resort.

The Fed still matters, but the long end is doing most of the talking. In May, global bond markets were hit by the Iran war, hotter inflation, and growing concern about public debt. The 30-year U.S. Treasury yield briefly reached around 5.2% on May 20, its highest level since 2007, before peace-talk hopes and softer growth data helped yields calm down.

Market has repriced the potential of a fed hike again this week, currently looking for an unchanged trajectory for a longer period of time.

That is the macro regime in one paragraph. The market is not simply pricing inflation. It is pricing inflation plus supply plus fiscal risk plus uncertainty over who absorbs duration.

The fiscal numbers are not catastrophic day to day, but they are heavy enough to matter. The U.S. posted a $215 billion April surplus, which sounds fine until you notice it was down 17% from a year earlier, while outlays rose and gross interest on the public debt hit a record monthly $112 billion. For the first seven months of fiscal 2026, gross interest costs reached $734 billion, up $50 billion from the year before.

That is why the bond market is no longer treating deficits as background noise. Interest expense is becoming a flow. It is not just a spreadsheet problem for 2034. It is a monthly cash-flow item that has to be financed today.

Treasury borrowing also remains a live issue. The Treasury lifted its April-June borrowing estimate to $189 billion, $79 billion above its February projection, and projected $671 billion of borrowing in the July-September quarter. That does not mean the market cannot absorb it. It means absorption is now a price-sensitive process.

This is the part that connects directly to risk assets. When the government needs more buyers, when AI companies need more financing, when stablecoin issuers need more bills, and when investors can earn real yield in cash or Treasuries, every speculative asset has to compete harder for capital.

Inflation makes that competition worse. April PCE rose 3.8% year on year, the fastest pace since May 2023, while core PCE rose 3.3%. The monthly core number was less alarming at 0.2%, but the headline pressure still came through energy, with the Iran war and Hormuz disruption feeding directly into prices. Consumer spending rose, but the saving rate fell to 2.6%, and real disposable income weakened.

That is not a clean recession setup. It is also not a clean soft-landing setup. It is a nominal resilience setup. Households are still spending, AI capex is still running, government outlays are still large, but the financing cost of that resilience is going up.

This is why the market can rally on peace headlines and still feel fragile. A credible US-Iran deal would lower oil risk premium and help the inflation path. But it does not solve Treasury supply. It does not solve interest expense. It does not solve the fact that AI capex is competing for capital at the same time the government is issuing more debt.

The Collateral Trade

The theme this week is that the market is becoming a collateral trade.

That sounds abstract, but it is very practical. The assets that matter most right now are the ones that either create collateral, absorb collateral, or compete with collateral.

Treasuries are the base collateral. Stablecoins are becoming a new wrapper around short-duration Treasuries. BTC is trying to become pristine non-sovereign collateral. AI infrastructure is issuing claims on future productivity. Equities are claims on earnings. Altcoins are claims on future network value. All of them are competing for the same investor balance sheet.

In the zero-rate world, the market could buy all of them. In this world, it has to choose.

That is why dispersion should increase. The market will not simply buy “crypto” because regulation improves. It will buy the parts of crypto that improve the collateral stack. The market will not simply buy “AI” because capex is large. It will buy the parts of AI that own bottlenecks, power, grid access, memory, cooling, or financing channels. The market will not simply buy “risk” because oil falls. It will buy assets whose liquidity improves when oil falls.

BTC fits that world better than most crypto. Stablecoins fit it better than most DeFi tokens. Regulated derivatives venues fit it better than offshore leverage tokens. SOL can still outperform tactically, but structurally the market is moving toward collateral, settlement, and regulated leverage.

That is also why ETH needs to prove what it is. Is it money? Is it equity in the settlement layer? Is it yield-bearing internet collateral? The market is not giving full credit for all three at once anymore. It wants one clean answer.

[Chart 6: BTC market cap vs stablecoin supply vs Treasury bill issuance. Place at start of theme section.]

Looking Ahead

Next week, the question is not just whether BTC holds a level or whether the Fed sounds hawkish. The better question is whether balance sheet becomes easier or harder.

For crypto, watch three things medium term.
I. BTC performance while real yields stay elevated.
II. Stablecoin supply resumes growth rather than just staying flat.
III. Onshore perp story creates genuine liquidity or simply more regulated leverage in a market that is not ready to use it.

For macro, watch the long end more than the Fed speakers. If 30-year yields calm down and oil keeps falling, risk assets can breathe. If long-end yields rise even as oil falls, that is a much more important warning, because it would mean the issue is no longer just inflation. It would mean the market is pushing back on fiscal supply.

The bottom line is that this is not last week’s market anymore. The US-Iran deal still matters, but it is no longer the whole story. The deeper issue is absorption. Who buys the Treasuries? Who finances the AI buildout? Who holds the stablecoin collateral? Who warehouses crypto leverage? Who pays for duration?

In a market like this, the winners are not always the assets with the best story. They are the assets that can become part of the collateral system.

Share

Subscribe now

Crypto

The map is still broadly the same as last week’s note: BTC remains the anchor, the upper-$70k area is still the line that matters, and the real overhead supply still sits much higher, first in the $90k-$99k area and then again above $100k. What changed this week is not the structural chart, but the character of the tape. BTC absorbed macro pressure better than expected, and that matters because last week’s framework was built around BTC defending support while macro stayed difficult rather than cleanly improving.

The relative performance is even more informative than the outright move. ETH underperforming in a week like this still looks like a duration problem: it behaves like long-duration tech beta without BTC’s scarcity bid and without SOL’s reflexive momentum. SOL finishing best does not mean a new alt season has begun. It means that once the macro backdrop stops getting worse, the market still prefers the higher-beta chain with stronger tactical flows. That is a tradable distinction, but it is tactical rather than structural.

Macro

The Fed itself said on 29 April that economic activity had been expanding at a solid pace, that unemployment had been little changed, and that inflation remained elevated, in part because of higher global energy prices. Since then, the hard data has kept the hawkish discussion alive: April payrolls rose by 115,000, unemployment stayed at 4.3%, headline CPI accelerated to 3.8% year on year, and producer prices rose 6.0% year on year, the hottest annual PPI print since late 2022.

That said, the soft-landing story has not broken. The Atlanta Fed’s GDPNow estimate for Q2 was running at 4.3% on 21 May, which is far too strong for a recession narrative. The hidden wrinkle is that some of the apparent resilience looks precautionary rather than organic: S&P Global’s flash May manufacturing PMI rose to 55.3, but firms are building inventories against shortages and cost shocks tied to the Hormuz disruption, while Chris Williamson warned that Q2 growth might still struggle to exceed 1% once you strip out that stock-building effect. Our read is that the US economy is still landing softly, but part of that softness is being financed by bringing future demand forward into current inventories.

That distinction matters for the Fed. The market is now well set up for hawkish outcomes not because growth is booming, but because growth is good enough and inflation is hot enough. Kevin Warsh was sworn in as Fed chair on 22 May and was immediately elected chair of the FOMC, while Philadelphia Fed President Anna Paulson said earlier in the week that it was “healthy” for markets to consider an extended hold or even hikes from here. June is still overwhelmingly centred on the current range, but by year-end the distribution has shifted materially upward.

This is why we still think one or two hikes are not automatically the end of the world. The cleaner read is Fed net liquidity (Fed balance sheet - Treasury General Account - reverse repo ) and that still does not look like a systemic liquidity squeeze. After sliding into early 2026, net liquidity has rebounded back inside the post-2022 range and is now sitting around the lower-to-middle part of that band rather than breaking down. That matters. A Fed hike would tighten the price of money, but the quantity of liquidity available to the system is not collapsing.

The place where tightening is already biting is the long end. The 10-year Treasury yield had moved above 4.5%, its highest level since early 2025, as bond markets pressured Washington over inflation and the Iran shock. That is real tightening whether or not the Fed hikes.

It hits housing, compresses equity multiples, and raises the hurdle rate for every speculative duration asset, including crypto.

Positioning makes the setup more fragile. Bank of America’s May survey showed the biggest ever jump in equity allocations, and Barron’s summary of that same survey said semiconductors had become the most crowded trade after rising oil had held that title in April. That is the key behavioural point for this week: the market is already long the two stories it trusts most, energy scarcity and AI scarcity. Whenever both become consensus at the same time, the next move usually comes from whichever consensus gets challenged first.

The US-Iran Deal Is Really a Rates Trade

The market desperately wants a deal, and there is enough smoke that you cannot dismiss it. Trump said on 23 May that a memorandum of understanding was “largely negotiated”, and report that a proposed framework being discussed through mediators could include a 60-day ceasefire extension, reopening the Strait of Hormuz, sanctions waivers, access to Iranian ports, and potentially some asset relief for Tehran.

But the deal is not done, and that is the most important nuance. Iran has not agreed to hand over its highly enriched uranium stockpile, that Trump said on 24 May there was “no rush” and that the US blockade would remain until there is a certified agreement, and that Rubio has warned any Hormuz tolling system would make diplomacy unworkable. In other words, there may be a framework for de-escalation, but the implementation details still sit exactly where markets usually get trapped.

Why does this matter so much? Because the transmission is not really diplomatic, it is macro-financial. The Hormuz disruption had threatened roughly one-fifth of global oil and gas flows, while 10-year yield repricing sharply higher on inflation fears tied to the same conflict. So if the market starts to believe the Strait reopens in a credible, durable way, the first effect is not peace dividend sentiment. It is lower oil risk premia, softer inflation expectations, less Fed-hike premium in the front end, and breathing room for duration-sensitive risk. That is why a real deal would be bullish for BTC even before the physical oil market fully normalises. This is an inference, but it follows directly from the way oil and yields have traded the conflict.

The catch is timing. Physical reopening, insurance normalisation, and actual shipping recovery will lag headline diplomacy. If traders price a full disinflation reset too quickly, they risk front-running a process that is still politically and mechanically messy. And if talks stall, the market has no room for complacency because CPI and PPI are already hot enough that another energy leg higher would re-ignite the second-round inflation debate immediately.

AI Capex Is Not Slowing

The other reason we do not think modest additional Fed tightening necessarily ends the cycle is simple: AI capex is still going full speed.

• Google and Blackstone are launching a new AI cloud venture with an initial $5 billion equity commitment and 500 megawatts of planned data-centre capacity by 2027

• Meta sold $25 billion of bonds after lifting 2026 capex to $125 billion-$145 billion

• Amazon still sees roughly $200 billion of capex in 2026

• Major tech AI infrastructure spending is expected to exceed $700 billion this year.

ASML’s chief executive is still talking about a supply-limited chip market, while Nvidia’s results again underscored how extreme the demand side remains.

That means the hidden alpha is probably no longer “just buy semis.” If semiconductors are already the most crowded trade, then the cleaner scarcity may now be one layer underneath: power, grid access, cooling, memory, and financing channels. AI-related convertible issuance reached about $34 billion in the first four months of 2026, and that NextEra’s planned acquisition of Dominion is explicitly about owning more of the electricity footprint demanded by data centres in Virginia and the broader PJM region. If oil cools because US-Iran tensions genuinely ease, while AI capex keeps eating power and financing capacity, we would expect leadership to broaden from front-end GPU proxies into infrastructure toll-collectors. That is where the next leg of the AI trade may become less crowded and more interesting. This is again an inference, but it is supported by where the new financing and industrial capacity are actually going.

Looking Ahead

Next week is about confirmation, not storytelling. Macro will care most about whether the next inflation prints validate the recent CPI/PPI heat, whether the long end can stop pushing higher, and whether US-Iran headlines move from “framework” to “mechanism”. Core PCE as the next critical inflation test on Thursday, to help settle market debating hikes rather than cuts.

For crypto, the filter remains simple. If BTC keeps holding the upper-$70k area and yields stop making fresh highs, the local base remains valid. In that world, BTC should keep acting as the defensive leader, SOL remains the best tactical beta if risk re-expands, and ETH likely continues to lag until macro volatility falls. If the Iran deal becomes real, the cleanest squeeze is lower oil, lower yields, and better crypto beta. If talks fail, respect the downside quickly, because the market is already positioned for a hawkish Fed and the crowded trades are crowded for a reason.

The bottom line is straightforward. The US economy still looks more soft landing than hard stop. Liquidity is still supportive enough that one or two hikes would hurt but not necessarily break everything. AI capex still refuses to slow. The real swing variable is whether the US-Iran process can compress the oil premium before second-round inflation psychology becomes embedded. If it can, this week will look like a washout inside a base. If it cannot, macro keeps driving the bus.

Share

Subscribe now

BTC struggled to build momentum, alts lagged, yields moved higher, and Fed pricing shifted further hawkish. In that context, Bitcoin’s ability to defend the same support area we have been tracking for weeks is meaningful. The setup is still fragile, and this is not a clean risk-on environment, but unless macro deteriorates further, the case for a local BTC bottom continues to improve.

BTC is still the anchor. The key level remains $78k. Price pushed into the low $80ks earlier in the month, failed to extend, and has now pulled back toward the upper end of the major support zone. That is where bulls need to step in. As long as BTC holds around $78k, the market can continue to build a base. A reclaim of $80k would open the door to the mid-$80ks, where momentum can return quickly. Lose $78k decisively, and $65k–$68k becomes the obvious downside magnet again.

The larger structure is also unchanged. BTC still has heavy overhead supply above. $90k–$99k remains the first major resistance area, and $100k–$108k the next battleground. For now, the conversation is about whether BTC can keep holding support while macro pushes against risk assets.

That is why this week was constructive, even if the price action did not feel good. BTC was down around 3% from the start of the week, roughly in line with total crypto market cap. ETH was weaker, down close to 6%, and SOL was weaker still, down more than 7%. This is not the tape of an aggressive risk-on rotation. It is defensive leadership. BTC remains the safest expression of the crypto trade, while alts are still waiting for permission.

But the alt setup is becoming more interesting. Altcoins have already taken a lot of pain. TOTAL3/BTC is sitting right on a long-term support shelf, and historically these levels tend to matter. Either alts break down violently from here, or they set up a strong reflexive bounce. Given BTC is holding support and has not made a fresh leg lower despite the macro pressure, the risk/reward for alts is improving tactically.

That does not mean we are calling for a new alt season. It means that if BTC has bottomed locally, alts should perform better on the upside. The first step is BTC stabilising above $78k. The second is reclaiming $80k. The third is macro not getting worse. If those boxes are checked, the alt complex can bounce hard from here.

The important word is “tactically.” This is not a structural all-clear. The market still has to deal with rising yields, sticky inflation, and a Fed being forced away from cuts. But crypto positioning already reflects a lot of fear. BTC has absorbed a difficult macro tape without breaking. Alts have lagged hard. Relative charts are sitting at support. That combination is usually where tradable rallies begin.

The other positive this week came from regulation. The Senate Banking Committee passed the Clarity Act in a 15–9 vote, sending the bill to the full Senate. Crypto has been waiting for clearer U.S. market structure for years, and the Clarity Act is one of the more serious attempts to draw lines between securities-style and commodities-style oversight. The market does not need every question solved tomorrow, but it does need direction. This vote gives it direction.

Still, we should not overstate it. A committee vote is not law. The bill still needs to clear the full Senate, and the politics remain messy. The Block highlighted comments from GSR’s legal chief, who put the odds of passage below 50%. Polymarket is more optimistic, with odds around 61% that the Clarity Act is signed into law in 2026. That gap is useful: lawyers are cautious, prediction markets are more bullish, and the truth is probably somewhere in the middle.

For crypto, the direction of travel is positive. Better market structure, clearer rules, and a more mature U.S. framework are all supportive for institutional capital over time. But regulation is not the main driver of the tape right now. It is a supportive background catalyst, not the thing deciding whether BTC breaks $80k or retests $65k.

Macro is still driving the bus.

Macro

The macro story became more uncomfortable this week.

U.S. data is running hot again, and American exceptionalism is back on screens. Not because the economy is perfect, but because the data keeps the Fed uncomfortable. CPI is hot. PPI is hot. Import prices are rising. Retail sales are holding up. The labour market has cooled from the post-pandemic boom, but it has not broken. That is the problem.

At the start of the year, the market was still dreaming about cuts. That dream is gone. The question is no longer “how many cuts does the Fed deliver?” The question is now “can the Fed avoid hiking if inflation keeps moving higher?”

That is a huge change.

The war and the Strait of Hormuz remain the centre of the inflation problem. The double blockade has dragged on long enough that the initial shock is starting to look less like a headline and more like a macro regime. Commodity, shipping, and supply chain experts continue to warn that the shutdown will eventually create broader supply shocks. Those warnings can feel stale because markets have short attention spans. But the longer the disruption lasts, the harder it becomes to dismiss.

Energy shocks are never just energy shocks. First they hit oil. Then gasoline. Then freight. Then import prices. Then producer prices. Then businesses begin passing higher costs through to consumers. At that point, the Fed cannot simply call it a one-time price-level adjustment and move on.

This is why the “transitory war shock” argument is too easy. Yes, part of the inflation impulse is caused by the war. Yes, if Hormuz reopens, some of that pressure can fade. But central banks do not just care about the first shock. They care about the psychology that follows. Once companies realise they can raise prices and consumers keep paying, inflation becomes harder to kill.

The Fed pricing chart captures the shift perfectly.

At the start of the year, the market was still pricing a cutting cycle. The January snapshot showed policy drifting lower through 2026. By mid-February, that path had already moved higher because the labour market refused to crack. By April 1, after the war premium started to matter, the easing cycle had largely disappeared. The market had moved from “cuts are coming” to “maybe the Fed just stays on hold.”

Now the red line is above all of them.

That is the big macro story this week. The market has gone from pricing cuts to pricing hikes. Kevin Warsh taking the helm of the Fed makes the setup even more interesting. The crazy reality is that his first move could be a hike, not a cut. That is not necessarily the base case yet, but it is no longer a tail risk that can be ignored. If inflation keeps rising and the labour market stays in equilibrium, Warsh and Waller could both sound hawkish by mid-summer. A December hike is now a realistic market conversation.

The bond market is already taking notice. The 10-year and 30-year yields are pushing higher again, with the long end doing real tightening work. Higher long-end yields pressure equity multiples, tighten financial conditions, and make speculative assets harder to own.

The AI capex boom is the other side of the macro story. It continues to support equities, keep some regional labour markets tight, and maintain the American exceptionalism narrative. The hyperscalers are still spending aggressively. Data centres, power, chips, cooling, construction, and infrastructure are all absorbing resources. That keeps parts of the economy hotter than they otherwise would be.

So the market is stuck between two forces. On one side, AI capex and resilient U.S. growth keep risk appetite alive. On the other side, inflation, energy, shipping, and higher yields are pulling the Fed more hawkish. Stocks have been outperforming, but the move could be running out of steam. If yields keep rising, eventually the discount-rate math starts to matter again.

For crypto, the macro filter remains simple. If yields stabilise and Fed pricing stops moving hawkish, BTC can hold support and alts can bounce. If oil pressure eases and Hormuz moves toward resolution, risk assets can breathe. If inflation cools in the next set of data, the market can return to the idea that the Fed is on hold rather than preparing to hike.

But if the opposite happens, downside risk returns quickly. Another extension of the Hormuz disruption, another hot inflation print, another move higher in the long end, or another repricing toward Fed hikes would make the crypto bounce much harder to sustain.

That is why we should not confuse a local bottom with a solved macro problem. BTC can bottom here. Alts can outperform on the upside. The Clarity Act can improve the medium-term regulatory outlook. All of that can be true while the bigger macro backdrop remains dangerous.

Looking Ahead

Next week is about confirmation.

BTC has done enough to keep the tactical bounce alive, but not enough to declare victory. The level to watch is still $78k. Hold above it, and the local bottom case remains intact. Reclaim $80k, and the path toward the mid-$80ks opens up quickly.

That is where alts should start to work. They are washed out, TOTAL3/BTC is sitting on long-term support, and positioning already reflects a lot of fear. The alt trade does not need perfect macro. It just needs BTC to stop going down.

But macro still has the final say. If yields keep pushing higher, oil stays firm, or Fed pricing moves further toward hikes, the bounce becomes much harder to sustain. A clean loss of $78k would put $65k–$68k back in play.

So the playbook is simple: hold $78k and reclaim $80k, and we lean into the bounce. Lose $78k, and we respect the risk of another leg lower.

The market has earned a bounce.

It has not earned complacency.

Share

Subscribe now

This week largely validated that framing. By the weekend, BTC was trading around $81.4k and ETH around $2.3k, while IGV finished near $91.15, S&P500 near $7398 and Nasdaq above 29,200. Both the S&P 500 and Nasdaq notched fresh record highs during the week as AI-linked stocks and strong earnings kept the broader US risk complex well bid.

The move was constructive, but it still did not look like a clean, endogenous crypto breakout. It looked much more like BTC continuing to act as the cleanest liquid proxy for improving macro sentiment, rich US equity leadership and easing near-term stress around oil. That was bullish while equities remained in gear, but it also meant crypto had not detached from the macro tape.

Crypto

From a pure market-structure perspective, BTC did the important thing. It took out $80,000 early in the week, briefly slipped back below, and then stabilised back above it. A level last seen since late January and by Sunday still closing above it. That kept the mid-$80,000s as the next obvious upside pocket rather than a distant scenario

That mattered even more because it lined up almost perfectly with the framework from last week. We had argued that above $78,000 was constructive and above $80,000 was momentum. This week delivered both. The clean takeaway was that BTC had improved its short-term structure, but it had not yet earned immunity from a broader risk wobble. The breakout looked real enough to respect, though still conditional enough not to chase blindly.

ETH, by contrast, still did not behave like leadership. It did the necessary thing by holding the shelf we had highlighted last week around the low-$2,000s and it traded near $2.36k by the weekend, but it remained well below the $3,000-plus zone that would have made the chart look genuinely convincing again. That left ETH in the same role as before: alive, stabilising, but still following rather than leading.

Under the surface, the tape was more interesting than the headline BTC move alone. In our basket work, the 50-100 cap cohort outperformed the overall market, Solana remained one of the cleaner beta leaders, and the ex-top-10 complex also participated.

At the same time, the broad top-50 basket and ETH lagged. That was actually a constructive mix. It suggested the rally broadened selectively into higher-beta pockets, but it still stopped short of the kind of indiscriminate alt frenzy that usually marks a late, frothy phase.

Cross-asset leadership and sentiment

The clearest tell remained outside crypto. US equities spent the week pressing higher, helped by a still-strong earnings tape and ongoing AI enthusiasm. S&P 500 Q1 profit growth was on track for the strongest pace since late 2021, while AI-linked names kept extending leadership. AMD alone jumped 14.9% after its forecast, the Philadelphia Semiconductor Index hit another record, and Google Cloud revenue surged 63% while investors punished Meta for pushing capex higher without providing the same clarity on monetisation.

Sentiment reflected that divergence. US Equity sentiment trackers fear-and-greed gauge sat around 67, firmly in “greed”.

while Crypto fear-and-greed index was only 52, which is merely “neutral”.

That gap is important. It indicates that the US stock market had moved much closer to enthusiasm than crypto had, even though BTC was rallying. Put differently, crypto benefited from the same cross-asset tailwind, but it still was not the primary source of the impulse.

Last week’s we flagged that BTC’s relationship with software (IGV) and equities was tightening again. This week, shows that re-coupling persisting, with IGV continuing to recover alongside BTC.

The practical conclusion stayed the same: BTC was leading crypto, but crypto was not leading macro. As long as the US growth-and-AI complex remained the market’s preferred expression of risk, BTC could keep grinding higher. If that equity leadership cracked, BTC still looked more likely to follow than to resist.

Technically, that left an important asymmetry. BTC’s 14-day RSI had risen into the mid-60s and its major long-term averages were clustered in the low-$80,000s, which suggested improving momentum without outright exhaustion. The Nasdaq looked different. RSI above 82, an unusually streched reading, which in prior instances had been followed by meaningful corrections of 15-20% usually within 30 days.

That does not mean equities have to roll over immediately. It means that the equity tape looked more crowded than BTC on a short-horizon momentum basis.

Macro

Macro still held the steering wheel, and oil remained the hinge variable. Brent settled at $114.44 on 4 May after renewed escalation around Iran and Hormuz, then dropped to $100.06 on 6 May as hopes of a temporary agreement improved, and finished 8 May at $101.29 after another volatile session in which traders again weighed ceasefire signals against renewed exchanges of fire.

That sequence mattered because it was exactly the transmission mechanism we discussed last week: lower oil eased immediate inflation pressure, which reduced pressure on rates, which helped equities, which in turn helped BTC.

The Federal Reserve had not been given a clean answer either. On 29 April it left rates unchanged and explicitly said inflation was elevated, in part because of higher global energy prices. Powell’s term as chair was due to end on 15 May, while Senate leaders were targeting the week of 11 May for a confirmation vote on Kevin Warsh. Market was simultaneously pricing stubborn inflation, a still-resilient economy and a live leadership transition at the Fed.

The data reinforced that awkward mix. April payrolls rose by 115,000 and unemployment held at 4.3%, first-quarter GDP grew at a 2.0% annualised pace, March headline PCE ran at 3.5% with core PCE at 3.2%, and ISM manufacturing held at 52.7 while input-cost pressure remained intense. Growth was still there. Inflation was still uncomfortable. That combination did not kill the rally, but it did make the rally increasingly dependent on the market’s willingness to look through valuation and policy risk.

That was why the US equity market looked a bit rich into the weekend. The S&P 500 and Nasdaq had both printed fresh highs, but valuation already sat above historical averages and the AI complex remained the dominant driver of marginal optimism. Forward 12-month P/E ratio for the S&P 500 at 21.0, above both the five-year and ten-year averages

Wall Street strategists were becoming more bullish because of AI-driven earnings momentum. Markets can of course stay expensive for longer than seems rational, but they rarely stop being expensive just because people notice.

The Week Ahead

Into the new week, the diary was set to stay heavy.

• US CPI due on Tuesday 12 May

• US PPI on Thursday 14 May.

The Federal Reserve calendar showed the weekly H.4.1 balance-sheet release on Thursday, industrial production on Friday 15 May, and a 14 May speech from Governor Michael Barr specifically on the balance sheet.

Meanwhile, Reuters reported that Chinese Vice Premier He Lifeng would meet a US delegation in South Korea on 12-13 May, ahead of a Trump-Xi summit in Beijing scheduled for 14-15 May.

Geopolitics also remained live rather than solved. Russia and Ukraine had agreed to a US-backed three-day ceasefire for 9-11 May, but both sides quickly accused each other of violations. US and Iran exchanged fire earlier in the week even as Trump insisted the ceasefire still held. In other words, the market had compressed some of the war premium, but it had not eliminated the geopolitical risk that fed into oil, inflation expectations and broader cross-asset volatility.

Crypto also had an identifiable policy catalyst on the horizon. The Senate Banking Committee formally scheduled an executive session for 14 May to consider H.R.3633, the Digital Asset Market Clarity Act of 2025. That mattered because it offered the market a real regulatory tailwind.

But it was still only a committee step, not the end of the process, so it should have been treated as incremental progress rather than a final unlock.

Closing view

The simplest way to frame the week was that last week’s levels worked. BTC did close above $78,000, it did take out $80,000, and it did end the week in a position where the mid-$80,000s remained the next zone to test. That was constructive. ETH, meanwhile, remained a support-hold chart rather than a leadership chart. Under the surface, selective higher-beta participation improved, but the move still looked more like an organised extension than the start of a broad alt euphoria phase.

The bigger point, though, was unchanged. Macro still drove the tape. US equities were richer, more euphoric and more obviously overextended than crypto, and that mattered because BTC was still trading in sympathy with that broader risk complex. If equities kept levitating and oil stayed contained, BTC probably still had room to carry momentum higher. If the Nasdaq’s extreme stretch finally triggered the sort of correction that overbought conditions often invite, we still doubted BTC would withstand it in isolation. The market had improved, but it had not become independent.

Last week’s closing instruction was effectively to watch oil, watch $80,000 and watch macro. This week, $80,000 gave way. Oil mattered every day. Macro still mattered most.

Share

Subscribe now

That bid is easiest to see in Strategy. On 27 April the company disclosed another 3,273 BTC purchase, taking holdings to 818,334 BTC, while its STRC preferred currently shows an 11.5% dividend rate and roughly $8.5 billion of notional value on Strategy’s own investor pages. That does not explain every basis point of BTC strength, but it does explain why dips have kept finding a buyer even when the macro backdrop has stayed noisy.

Crypto

Since the geopolitical unrest broke out at the end of February, BTC is up roughly +19.5%, versus +14.1% for Top 10, +13.6% for the Top 50, and +13.5% for the Top 100. That matters. The move has breadth, but leadership is still concentrated in the benchmark asset. This is not a market where the lowest-quality beta is leading. It is a market where BTC is absorbing the marginal dollar first, then selectively passing momentum down the curve.

That leadership also fits the structure of the move. The war premium has compressed, oil has backed off from the worst levels, equities have recovered, and BTC has responded like the most liquid high-beta macro asset in crypto. The important point is that BTC is outperforming crypto, but the driver is still mainly outside crypto. It is not yet a purely endogenous crypto cycle. It is a macro/equity risk-on move with BTC as the cleanest liquid proxy.

Technically, the chart is now at the important part.

BTC is pressing into the high-$78k area. A daily and weekly close above $78k would be constructive because it confirms acceptance above the upper end of the current support zone. The next obvious trigger is $80k. If $80k is taken out with strength, the chart opens quickly into the mid-$80s. That is the first upside pocket to watch.

Above that, the next resistance bands are more meaningful. The $88k–$90k area is where price starts to run into the lower edge of the prior distribution. The downside levels are just as important. As long as BTC holds above $76k–$78k, the short-term structure remains constructive. A move back below $75k would weaken the breakout and suggest the market has slipped back into the prior range. Below that, $70k–$72k is the first real support area, followed by the mid-$60k region. The broader range low still sits closer to $60k–$62k. In short: above $78k is constructive, above $80k is momentum, below $75k is caution.

Under the surface, the strongest sectors since the war began have been privacy coins and perp DEXs. That also makes sense. Privacy has a native narrative that does not depend on AI capex or U.S. equity earnings. Perp DEXs are the cleaner high-beta expression of renewed speculative activity: more volatility, more leverage, more 24/7 price discovery. If BTC is the reserve asset of this move, perp DEXs are one of the cleaner beta extensions.

ETH is more mixed. It has done the important thing for now, which is hold the $2.2k support shelf. That keeps the chart alive. But ETH is not leading. It needs to reclaim higher levels before the setup becomes convincing. The first job is to keep holding above $2.2k. Below that, $2.0k–$2.05k becomes the obvious risk zone. On the upside, $2.5k would help confirm stabilisation, but the real level is still around $3.0k–$3.1k. Until ETH can reclaim that area, it remains a support hold rather than a leadership chart.

The cross-asset charts are the real tell. BTC has outperformed both Nasdaq and the S&P 500 since the war began, but it is increasingly moving with equities again. The brief decoupling between IGV and BTC in early April has reversed. IGV has caught up, and BTC’s 20-day correlation with software is now around 0.71, up from basically zero in March. BTC/SPX correlation is even higher, around 0.92, from -0.8 in March.

BTC is leading crypto, but crypto is not leading macro. The main driver is still an equity story. BTC is acting like the cleanest liquid expression of risk appetite, institutional balance-sheet demand, and improving macro conditions. That is bullish while equities are bid. It also means BTC remains vulnerable if the equity market starts questioning valuations, AI capex, or the oil/Fed chain again.

Macro

Oil is still the anchor variable.

Brent traded above $115 during the week before fading, and by Sunday Brent for July was back near $106.3 after reports that Iran had submitted a proposal involving reopening Hormuz and lifting the U.S. blockade. OPEC+ also agreed to a symbolic 188k bpd quota increase for June, its third monthly increase since the Hormuz closure, although the physical disruption means the headline quota increase matters less than whether shipping actually normalises.

The chain is still the same. Lower oil reduces immediate inflation pressure. Lower inflation pressure reduces the urgency of a hawkish Fed. Lower rates pressure supports equities. Equities support BTC. When oil moves lower, the whole chain loosens. When oil reverses higher, the same chain tightens again.

That is why this week should not be framed as “risk is gone”. It is better framed as “the stress premium compressed”. The market removed enough war premium to let risk breathe, but not enough to make macro irrelevant. Hormuz, oil, and Fed expectations are still the macro transmission mechanism.

The Fed is trapped inside that same chain. On 29 April, the FOMC held the target range at 3.5%–3.75% and explicitly noted that inflation is elevated, partly reflecting higher global energy prices. The statement also said Middle East developments are contributing to a high level of uncertainty around the outlook. Powell’s press conference added the same point: headline PCE was 3.5% year-on-year in March, core PCE was 3.2%, and near-term inflation expectations have risen because of oil.

The macro data did not give the Fed a clean answer either. Q1 real GDP grew at a 2.0% annualised pace, up from 0.5% in Q4. March PCE was hot, with the headline index at +3.5% year-on-year. ISM manufacturing held in expansion at 52.7, but input costs hit a four-year high. Growth is still there. Inflation is still awkward. That combination keeps the Fed on hold and keeps the market dependent on whether oil continues to cool.

Equities are the other side of the story. The S&P 500 and Nasdaq closed at record highs on Friday, helped by earnings and the dip in crude. S&P 500 Q1 earnings growth is tracking around 27%–28%, the best pace since Q4 2021, while the forward 12-month P/E has moved to 20.9, above both the five- and ten-year averages. That is a strong earnings tape, but it is not a cheap tape.

The U.S. tech trade is back, but the market is becoming much more selective. Google was rewarded because Cloud revenue surged 63% and investors could see a clear monetisation path. Meta was punished because the capex number went higher while the payoff stayed more uncertain. Microsoft and Amazon remain in the same AI infrastructure race. Across Amazon, Google, Meta, and Microsoft, 2026 capex plans are now above $700bn. This will soon eat up all free-cash flow.

This is the equity filter BTC is trading through. Huge AI spend is acceptable when the revenue engine is visible. It is not acceptable when the market has to rely on faith. That is why Google ripped and Meta dumped. BTC benefits from the broader reopening of the growth trade, but if investors begin to question the AI capex spiral, BTC will feel that too.

Closing view

The setup into 4 May is constructive, but still conditional.

BTC remains the leader. A close above $78k keeps the short-term structure intact, while $80k is the trigger. If that level breaks cleanly, the mid-$80s come into view quickly. ETH is not leading, but holding the $2.2k shelf keeps the chart alive.

Under the surface, the tape is still selective. Most names have underperformed BTC, and the outperformance has been concentrated in pockets rather than broad-based beta. Privacy coins and perp DEXs have led, but this is not yet an indiscriminate alt market. For now, BTC remains the cleanest expression of the move.

The caveat is that this is still mainly an equity and macro trade. If oil keeps easing, equities stay bid, software continues to work, and the Strategy bid remains intact, BTC has room higher. If oil reverses, the Fed hardens, real yields rise, or rich equity multiples get questioned, BTC will not trade in isolation.

Watch oil. Watch $80k. Watch macro. Everything else follows.

Share

Subscribe now