The Fusion Report (usually) tries to stay away from politics, but the intersection of politics, energy, and the economy made it incredibly hard to do in 2026. In 2024, Donald Trump won the presidency specifically because of the economy, particularly because of a US rate of inflation of 9.1% in June 2022, and a record price of a dozen Grade A large eggs of $4.82 in January 2023. Now, the headline consumer price index (CPI) in September was 3.7%, up from 3.4% in August. Worse yet, the price of gasoline is up 34.2% on average since the beginning of Donald Trump’s second term, due to a very unpopular war with Iran started and waged by the Trump administration. And while Donald Trump is not running for president, a lot of people in the Republican Party are running for re-election in 2026, and the odds don’t look good for them. The question is, what does all this mean for energy policies, particularly for fusion energy?

Source: NBC News Decision Desk 2026 race ratings

Current Expectations For The US Congressional Elections

As of October 23, Democrats are strongly favored to regain the US House of Representatives in the November 3, 2026 congressional elections, while control of the Senate remains more uncertain. Cook Political Report’s latest assessment projects Democratic gains of 12–22 House seats and now rates Democrats as favored in the 218 districts needed for a majority, with another 17 races classified as toss-ups. That suggests a potentially comfortable Democratic majority, although Republican-friendly district boundaries could limit seat gains even in a strong national showing. The political environment is unfavorable for Republicans: President Trump’s low approval ratings, frustration over consumer prices and stronger Democratic voter enthusiasm have made the elections increasingly a referendum on his administration. Nevertheless, Republicans’ campaign-spending advantage and voters’ continuing reservations about Democrats mean a landslide is not assured.

The US Senate offers Republicans a stronger defensive position, but a Democratic takeover has become a credible possibility rather than a long shot. Republicans hold a 53–47 majority, meaning Democrats need a net gain of four seats to take control. UBS’s October 16 assessment gave Democrats perhaps slightly better than even odds, with their path depending on defending vulnerable seats—including Michigan and Georgia—winning North Carolina, and capturing at least three additional Republican-held seats from a competitive group including Maine, Ohio, Iowa, Alaska, Kansas and Texas. Many of those contests remain within polling margins of error, making turnout and polling accuracy decisive uncertainties.

Energy Policy Changes If Democrats Win Both Houses of Congress

If Democrats win both chambers, U.S. energy policy would shift toward protecting clean energy, expanding the grid and scrutinizing the Trump administration—but not a wholesale reversal of current policy. The key constraint is the presidential veto: Trump would remain president through January 2029, and ordinary Democratic majorities could not restore clean-energy incentives or impose new restrictions on fossil fuels without his agreement or enough Republican support to override him. The most immediate effect would therefore be to block further legislative rollbacks, strengthen oversight and gain leverage in spending negotiations.

The Impacts Of A Win On The Commercialization of Fusion Energy

A Democratic victory in both chambers in 2026 would likely give fusion energy a stronger congressional platform for funding and commercialization, rather than trigger a fundamental policy reversal. Senate Democrats already include fusion in their clean-firm-power agenda, alongside nuclear fission and geothermal energy, and have proposed financing mechanisms to accelerate deployment. A concrete legislative vehicle is the bipartisan American Leadership in Fusion Act, introduced in September 2026 by Representatives Zoe Lofgren, a Democrat, and Jay Obernolte, a Republican. It proposes $10 billion in direct investment, including $3.8 billion for major materials and fuel-cycle testing facilities, $3 billion for a new milestone-based demonstration program and $2 billion for the existing milestone-based development program. Democratic control could improve the prospects for advancing this package or incorporating parts of it into other legislation, although those investments remain proposals—not enacted funding.

The most consequential potential impact would be greater funding certainty for the infrastructure and public-private partnerships needed to move fusion beyond successful experiments toward commercially viable power plants. Supporting shared testing facilities, materials research, fuel-cycle development and milestone-based demonstrations could help private companies address engineering bottlenecks that individual developers struggle to finance alone; DOE explicitly states that its commercialization roadmap depends on future partnerships and congressional appropriations. However, a Democratic Congress would still need agreement from President Trump on new legislation and spending, and fusion’s existing bipartisan backing means election results alone would not determine its trajectory. The likely opportunity is therefore a negotiated expansion of commercialization support—not a guaranteed funding surge or an immediate acceleration of deployment, which would still depend on technical progress.

Conclusion: What Changes If The Election Is Not Clear-Cut?

A Democratic House and Republican Senate would mean divided congressional control—not an unclear result, but a clear division of power. Democrats could block Republican legislation, control House committee investigations and demand concessions in spending negotiations, but they could not enact their own energy agenda without Senate approval and President Trump’s signature. Republicans would retain control of Senate committees and the confirmation process for executive-branch and judicial nominees. For energy policy, the likely outcome would be more bargaining and fewer sweeping statutory changes: neither a full restoration of Democratic clean-energy policies nor an unrestricted continuation of Republicans’ legislative agenda. Existing laws and executive policies would not automatically change because the House changed hands.

The implications for fusion is that it could still advance under this arrangement because it already has bipartisan congressional support. The American Leadership in Fusion Act, introduced by Democrat Zoe Lofgren and Republican Jay Obernolte, illustrates the potential for agreement on research infrastructure and commercialization support even when broader climate policy remains contentious. A Democratic House could advocate funding for testing facilities and public-private demonstrations, but the final amounts would require agreement with the Republican Senate and Trump. The principal risk would therefore be budget uncertainty and legislative delay—not necessarily a partisan reversal against fusion. New program authorizations would also need to be distinguished from actual appropriations: permission to establish or expand a program does not itself guarantee the money.

In short, the positive direction that’s fusion Is trending towards is not likely to change, whether the Democrats just win the House or they win both the House and the Senate. But the most important outcome is neither a House nor a Senate win by the Democrats or Republicans; it’s a demonstrated achievement of Qeng>1, which would still leave a long ways towards practical fusion-powered electricity (to date, NIF is the only facility that has achieved scientific gain, or Qsci>1, and no facility has yet achieved Qeng>1); this is what everyone in the fusion world has been working and waiting for. In that sense, the commercial fusion industry for better or worse has its future in its own hands.

Things You Gotta Know

Commonwealth Fusion Systems Signs Letter of Intent for Up to $85 Million CHIPS Award to Advance Semiconductors for Fusion Power Plants
Commonwealth Fusion Systems announced that it signed a letter of intent with the U.S. Department of Commerce for up to $85 million through the CHIPS Research and Development Office to develop semiconductors for fusion power plants. The proposed award would fund R&D on silicon carbide power electronics for magnet power supplies, solid-state radio frequency generators for plasma heating, AI applied to SPARC and ARC including a SPARC digital twin, faster control electronics, and radiation-hardened electronics for neutron-exposed systems. The letter is one of four totaling up to $389 million, alongside up to $175 million for Helion, up to $73 million for Vertical Semiconductor and up to $56 million for Sibelco.

Blue Laser Fusion Closes $34 Million Capital Raise and Alternative Public Offering to Advance Directed-Energy Defense and Fusion Energy
Blue Laser Fusion, founded by Dr. Shuji Nakamura, announced that it closed a financing and alternative public offering with total proceeds of $34 million, including an initial closing of $25 million in private placement units, completed concurrently with a reverse merger with Unite Acquisition 2 Corp. Laidlaw & Company (U.K.) Ltd. served as exclusive placement agent. The Goleta, California company plans to use the proceeds for directed-energy defense prototypes and field tests, including counter-drone and missile defense applications, along with pulsed-laser development for inertial confinement fusion and expanded manufacturing in Goleta. The company said it may close additional tranches of funding.

Helical Fusion and Germany’s Gauss Fusion Sign Joint Statement of Intent to Advance Stellarator Fusion Power Commercialization
Tokyo-based Helical Fusion and Munich-based Gauss Fusion announced a joint statement of intent to accelerate commercialization of Helical Stellarator and Modular Stellarator technologies through private-sector leadership. The statement is a non-exclusive, non-binding expression of shared intent, with each company keeping its independence in technology, business strategy and operations. The companies commit to develop, demonstrate and commercialize fusion energy, deepen dialogue drawing on Japanese and German research strengths, and share their vision with governments, research institutions, industry and investors. The initiative was referenced in an official bilateral document released for the Japan-Germany Leaders’ Meeting, and no financial terms or project timelines were disclosed. Helical Fusion, a spin-out of Japan’s National Institute for Fusion Science, runs the Helix Program targeting commercially viable fusion power in the 2030s with more than 30 industrial partners.

Proxima Fusion Signs HTS Tape MoU With Fujikura and Gyrotron MoU With Kyoto Fusioneering as Japan and Germany Deepen Fusion Ties
Proxima Fusion announced a memorandum of understanding with Japan’s Fujikura to jointly assess a manufacturing facility in Germany for high-temperature superconducting (HTS) tape, one of several Japan-Germany fusion agreements signed in Tokyo during German Chancellor Friedrich Merz’s visit to Japan. Fujikura would contribute its HTS technology and manufacturing know-how while Proxima contributes its long-term demand and technical requirements, starting with an evaluation of production in Lower Saxony, and Proxima anticipates needing about 20,000 km of HTS tape for its Alpha demonstrator and about 40,000 km for each planned Stellaris power plant. Proxima also signed an MoU with Kyoto Fusioneering to expand joint development of next-generation gyrotrons for plasma heating and fuel cycle technologies, with Kyoto Fusioneering contributing to Germany’s Fusion Hubs initiative. Kyoto Fusioneering separately signed an MoU with Darmstadt-based laser fusion developer Focused Energy to explore cooperation on fuel and target technologies, the tritium fuel cycle, materials for extreme conditions and diagnostics.

The Fusion Report has run several articles recently about the impacts of artificial intelligence (AI) on the US and global economy. With less than a week before the midterm elections, the topic has taken on political significance as well. So let’s dive into the questions of:

• What is the AI bubble (it’s not just about stock valuations);

• Who are the big players “inflating” the AI bubble; and

• When are the dangers of a bursting likely to show up in the US economy?

Art courtesy of Wikipedia

Defining the Concept(s) Of An AI Bubble(s)

An AI bubble is the possibility that the enthusiasm (real or otherwise) and market forces around artificial intelligence have pushed company valuations and investment spending beyond what future revenues and profits can justify (BTW, this article was researched and partially written using AI technology). The concern spans two related areas:

• Investors paying excessive prices for AI-related businesses;

• Companies building more computing infrastructure (chips, data centers, and power capacity) than customers will need, or is supported economically; and

• The larger economy is endangered by exposure to these forces, either by starving the capital markets, or through write-offs of AI-related corporate debts.

These forces can reinforce each other: rising valuations encourage greater spending, while ambitious spending plans strengthen expectations of extraordinary growth. A bubble would emerge if those expectations became detached from achievable financial returns, not simply because investment is large or the technology is attracting attention. In a sense, the impact will be similar to the combination of the late 1990s dot com bubble and the home mortgage interest bubble of the late 2000s.

The central distinction is that a transformative technology can still produce a financial bubble. As the internet boom demonstrated, lasting technological value does not guarantee that every company survives or that investors earn acceptable returns. The case against an AI bubble rests on genuine customer demand, expanding commercial applications, and the financial resources of major technology companies; the concern is that adoption and profitability may not grow fast enough to justify the spending already underway. If expectations disappoint, the adjustment could involve falling share prices, startup failures, infrastructure write-downs, and reduced investment rather than the disappearance of AI itself. The key question is therefore not whether AI is useful, but whether customers will generate enough revenue—and suppliers enough profit—to support the prices and investment commitments built around it.

Who Could Get Hurt By An AI Bubble?

The players in the AI ecosystem (and those who are endangered by a potential AI bubble) are more widespread and diverse than the diagram at the start of this article suggests; they range from AI developers and chip suppliers to cloud operators and their financiers to construction companies and people in the construction business, but their vulnerability differs considerably. AI developers dependent on external funding such as OpenAI, Anthropic, and AI-focused startups would face pressure if investors became unwilling to finance heavy computing costs before sustainable profits emerge. Debt-funded computing providers such as CoreWeave, Oracle, and similar companies are exposed because they must service their exponentially increasing borrowing, even if demand weakens or their hardware loses value.

Nvidia could suffer from slower chip purchases and its financial ties to customers, while Microsoft, Amazon, Google/Alphabet, Meta, and Oracle risk disappointing returns on expensive data-center investments—although their established businesses provide more protection than smaller, loss-making firms have. Even still, while their exposure may be smaller, the debts taken on by these companies in the name of AI are enormous; they have collectively added approximately $350B in debt obligations over the past five years to fund their massive artificial intelligence and data center infrastructure buildout. Beyond traditional on-balance-sheet debt, analysts and financial watchdogs like the International Monetary Fund and Moody’s Ratings highlight hundreds of billions of additional dollars in off-balance-sheet leases, special purpose vehicles (SPVs), and uncommenced commitments.

The financial fallout could also reach venture-capital investors, shareholders, banks, private-credit funds, and data-center bondholders: falling valuations would hurt equity owners, while underused facilities and financially distressed customers could leave lenders facing losses. Much of the data center boom is fueled by heavy private credit and specialized lending; if demand or profitability fails to meet expectations, defaults would ripple through financial institutions, drying up construction financing for commercial projects far beyond just tech. Even the construction industry and its employees would be hurt by a sudden freeze in data center construction, which has become a major driver of regional and national building business.

How Close Are We To The AI Bubble Bursting?

We are already seeing a significant AI-market correction and pockets of financing distress, but that is not yet proof that the entire bubble is bursting. As of October 16, 2026, the clearest warning comes from debt-heavy infrastructure providers: CoreWeave’s bonds have traded at yields of roughly 15%–16%, levels indicating serious investor concern about repayment, while a broader selloff has hit chipmakers and other AI-linked companies. The immediate trigger has included hotter inflation and higher borrowing costs—not definitive evidence that demand for AI computing has collapsed, as well as political and popular backlash against data center construction, as well as rising costs in energy and other related markets. However, there is no reliable basis that a sector-wide crash is days or months away.

The critical issue is financing: whether companies can keep funding infrastructure until customer revenues generate sufficient cash to cover operating costs, interest, and investment. Strong demand alone does not guarantee that a heavily indebted provider can earn an adequate return. Reuters’ reporting describes the current stress primarily as investors questioning how the expansion is funded, rather than a sudden disappearance of demand. According to Reuters, the most useful warning signs are:

• Credit problems spreading: more infrastructure providers struggling to borrow or refinance, rather than distress remaining concentrated in the weakest companies.

• Earnings disappointing: cloud operators and AI developers failing to deliver the revenue, margins, and cash flow investors expect. Upcoming earnings reports are becoming a key test of the investment narrative.

• Investment plans reversing: major customers cutting data-center and chip commitments, which would transmit financial stress into suppliers’ actual sales. Semiconductor demand and capital-spending plans are therefore important indicators to watch.

A sharp correction can still occur without a wholesale collapse: valuations could fall and weaker firms could fail while larger companies continue expanding. Conversely, a sustained credit squeeze combined with disappointing earnings could make the downturn self-reinforcing. The evidence currently supports describing this as an active stress test of the AI boom—not a confidently timed, inevitable crash.

Comparing An AI Bubble Burst Versus The Subprime Mortgage Crisis

So in many respects, an AI bubble will look similar to a dot-com bubble. But how would it compare to the subprime mortgage crisis? A potential AI bubble burst would resemble the 2008 subprime mortgage collapse in its underlying cycle: optimistic expectations encourage investment and borrowing, rising asset prices reinforce confidence, and disappointing returns can trigger losses, tighter credit, and forced asset sales. Before 2008, lenders and investors assumed housing prices would remain strong, allowing increasingly risky mortgages to be repackaged into securities and financed with substantial leverage. In the AI boom, the analogous assumption is that rapid adoption and future profits will justify enormous spending on chips, data centers, and supporting infrastructure. If those profits disappoint, falling valuations and infrastructure write-downs could hurt developers, suppliers, investors, and lenders simultaneously. The IMF has warned that reassessing AI’s expected returns could produce corrections in both equities and credit, reduce investment, and tighten financial conditions internationally.

The crucial difference is the mechanism of the market failures between a potential AI collapse, and the 2008 subprime market failure. In 2008, mortgage losses undermined securities held throughout the financial system, including assets treated as safe collateral for highly leveraged, short-term borrowing. When confidence evaporated, funding markets seized up, turning a housing downturn into a banking and credit crisis. An AI bust could initially look more like a technology-investment crash: shareholders absorb falling valuations, startups fail, and companies cancel expansion plans, without necessarily paralyzing the banking system. However, debt-financed data centers, private-credit exposure, and financial ties among AI companies create pathways for losses to spread beyond technology stocks. A comparable systemic crisis would depend less on how far AI shares fall than on whether those losses impair major lenders and disrupt funding markets; an AI correction alone does not establish that a repeat of 2008 is inevitable.

Conclusion: It’s Two Days Until Halloween – Which Is Scarier?

Many stock market analysts, leading investors, and industry insiders warn that the AI sector is showing late-stage bubble signs, with several projecting a potential peak or pullback in 2027 or 2028 (though opinions remain sharply divided). Both Ray Dalio (Bridgewater Associates) and Michael Burry (predictor of the 2008 subprime market disaster) warn that massive debt accumulation, combined with extended high interest rates, and disappointing AI revenues will make carrying that debt exceedingly difficult. Even Sam Altman, CEO of OpenAI, believes that AI could be a bubble similar to the dot-com boom of the 1990s. Let’s hope everybody is being overly skeptical; an AI bubble burst would be crippling to our economy, and like the 2008 subprime crisis or the dot-com bubble, take significant time and pain to recover from.

This is the sixth article in The Fusion Report series on critical supporting technologies for commercial fusion energy. In today’s article, we will explore the role that pulsed power systems play in fusion energy solutions, from inertial fusion energy (IFE) systems to magneto-inertial fusion/hybrid fusion energy (HFE) systems, to magnetic fusion energy (MFE) systems. After that, we’ll look at a particular type of pulsed power system, the Impedance-Matched Marx Generator (IMG) system, which are one of the most critical components in Z-pinch IFE systems.

Pulsed Power Systems And Fusion Energy

Pulsed power systems store energy slowly and release it in extremely short, high-power bursts, making them well suited to fusion experiments where a small fuel target must be compressed and heated before it can disassemble. In inertial confinement fusion (ICF), capacitor banks, Marx generators, and transmission lines deliver currents or voltages on nanosecond-to-microsecond timescales, concentrating energy into a tiny plasma target. This creates the extreme temperatures and pressures needed for fusion reactions, conditions analogous to those in stellar interiors.

The clearest example is Sandia National Laboratories’ Z machine, which uses the Z-pinch mechanism, in which enormous pulsed currents flow through an array of metal wires surrounding a deuterium-tritium (D-T) fuel capsule. The currents generate intense magnetic fields that drive the wires inward, producing X-rays that implode the capsule and compress its fuel to fusion-relevant conditions. Related approaches, including magnetized liner inertial fusion (MagLIF) and magnetically driven direct-drive ICF, use pulsed currents to both compress fuel and establish magnetic fields that reduce heat loss during the brief implosion.

In magnetic fusion energy machines, pulser systems—often called pulsed-power systems—provide the rapid, precisely timed bursts of electrical energy needed to create, shape, and control the plasma rather than to sustain it continuously. In tokamaks, they energize the central solenoid and poloidal-field coils to induce the plasma current, establish the magnetic equilibrium, and drive rapid transients such as plasma initiation, current ramp-up, and disruption mitigation. In pulsed magnetic fusion concepts such as magnetized liner inertial fusion (MagLIF), a pulser delivers a multi-megaampere current to a metal liner, whose magnetic forces implode it and compress a pre-magnetized, laser-preheated fuel plasma to fusion conditions. These systems typically store energy in capacitor banks or Marx-generator stages and release it through fast switches and transmission lines on microsecond-to-nanosecond timescales; for a power plant, the key challenges are high repetition rate, component lifetime, efficiency, and recovery of magnetic energy between pulses.

Pulsed power also matters beyond target implosion. In laser-driven ICF, pulsed electrical systems energize flashlamps or other driver components; in magnetic-confinement devices, fast pulsed-power switches and power supplies can manage transient loads, mitigate fault conditions, and support rapid shutdown of plasma disruptions. For a future fusion power plant, the central engineering challenge is repetition rate and efficiency: experimental machines may fire occasionally, but commercial systems would need reliable, high-frequency pulses while recovering as much stored magnetic and electrical energy as possible.

Performance Requirements of Pulsed Power in Laser Fusion Systems

Pulsed power systems in laser fusion must store large amounts of electrical energy and deliver it to the laser amplifiers with exceptional precision. In flashlamp-pumped systems such as Lawrence Livermore National Laboratory’s National Ignition Facility (NIF), modular capacitor banks charge over tens of seconds and then discharge through spark-gap switches into thousands of flashlamps, pumping the neodymium-doped glass amplifiers. The NIF power-conditioning system stores up to roughly 400 MJ, and supplies more than 34 kJ to each of nearly 8,000 flashlamps, with modules operating at roughly 24 kV and peak discharge currents around 25 kA. The pulses must be tightly synchronized across all beamlines because the final target shot depends on many laser beams arriving together with closely matched energy and timing.

For a commercial laser-fusion plant, the requirements become substantially more demanding than for a single-shot experimental facility. A power plant would need repetition rates on the order of several to ten shots per second, rather than the infrequent shots typical of present facilities, while maintaining high electrical-to-laser efficiency, component lifetime, and thermal management. This favors more durable and efficient pump sources—especially high-power laser diodes—along with solid-state or other long-lived switching technologies, robust capacitor or energy-storage modules, and power electronics capable of absorbing and recycling energy between shots. In short, the pulsed-power system must combine megajoule-scale energy storage, microsecond-scale controlled release, beam-to-beam uniformity, and the reliability and repetition rate required for continuous electricity production.

Impedance-Matched Marx Generators (IMGs) and Inertial Fusion

An IMG for a commercial IFE system must deliver extremely high peak power to a fusion target with minimal energy loss, precise timing, and enough durability to operate as part of a power plant. Unlike a conventional Marx generator followed by pulse-forming lines, an IMG is designed so that its internal electrical impedance matches the load impedance, allowing electromagnetic waves to add constructively and maximizing power transfer to the target. For pulsed magnetic inertial fusion, this generally means producing multi-megavolt, multi-megaampere-class pulses with a rise time near 100 ns; published requirements identify currents above 1 MA, voltages above 1 MV, energy-delivery efficiency above 85%, and repetition rates above 0.1 Hz as baseline driver targets. IMG architectures are attractive because they can reach roughly 90% efficiency, avoid the additional pulse-compression hardware that lowers conventional Marx systems to about half that efficiency, and reduce voltage reversal that can damage components.

For commercial operation, however, the decisive requirements extend beyond a single high-power pulse. The generator must fire reliably and repeatably at a rate compatible with plant economics—often framed as at least 0.1 Hz initially, with much higher repetition rates ultimately needed for meaningful electricity output—while surviving tens of thousands to millions or billions of shots without frequent replacement. It must also synchronize many switches and stages to nanosecond-level jitter, because the combined modules must deliver a coherent pulse to a common load; demonstrated prototype systems have targeted roughly 2 ns timing jitter across dozens of stages. Finally, a commercial IMG must be modular, mass-manufacturable, low-cost, thermally robust, and efficient enough that the fusion energy produced exceeds all energy drawn from the grid and stored in the capacitors—not merely the electrical energy delivered in one pulse.

One of the most notable examples is the Pacific Fusion Demonstration System, which utilizes an array of IMGs as the pulsed-power driver for their MagLIF system, replacing lasers with rapid electrical pulses to compress and heat fusion fuel. In its architecture, many low-inductance capacitor-and-switch “bricks” are arranged in modular IMG stages; when triggered, each stage launches a power pulse that propagates down the module and triggers subsequent stages, stacking the pulses without requiring separate pulse-sharpening hardware. This impedance-matched, traveling-wave design is intended to transfer energy more efficiently to the load than a conventional Marx generator, while supporting a modular, mass-manufacturable plant architecture.

In the Demonstration System, multiple IMG modules would combine their outputs through magnetically insulated transmission lines to deliver more than 60 MA to a centimeter-scale MagLIF target in a roughly 100-nanosecond pulse. The current drives a cylindrical metal liner inward, compressing a pre-magnetized and laser-preheated D-T plasma; the company’s system model indicates the driver could deliver more than 8 MJ to the target and about 1 MJ to the fuel. The central commercial premise is that IMG modules can provide high efficiency, fast rise time, and repeatable high-power pulses while being produced at scale, enabling a pulser-based IFE plant rather than a single-shot experimental accelerator.

Conclusion: Why Pulsed Power and IMGs for Commercial Fusion

Pulsed power is the enabling technology for commercial pulser-based inertial fusion energy (IFE): it stores electrical energy over longer intervals and releases it in a short, precisely synchronized burst that magnetically implodes and heats a fusion target. In this role, the driver must deliver multi-megaampere, multi-megavolt currents on roughly 100-nanosecond timescales, with high efficiency, low timing jitter, and repeatable pulse shape shot after shot. IMGs are a next-generation approach to meeting those requirements because they launch and combine electromagnetic waves internally, eliminating much of the lossy pulse-compression hardware used in traditional Marx systems and enabling energy-delivery efficiency near 90%.

For commercial fusion, however, the challenge is not only producing one powerful pulse but doing so economically at power-plant repetition rates—often on the order of 1 Hz—over a multi-decade plant life. IMGs address this by using modular, potentially mass-manufacturable capacitor-switch “bricks,” reducing stored-energy requirements, and improving efficiency, but they still require major advances in capacitor and switch lifetime, thermal management, triggering reliability, and component cost. In short, IMG-based pulsed power could make IFE more compact and efficient than legacy architectures, but its commercial viability depends on scaling from thousands of demonstration shots to millions of reliable, low-cost, high-efficiency pulses.

Things You Gotta Know

Helion Signs Letter of Intent for Up to $175 Million CHIPS Award to Advance U.S. High-Voltage Semiconductor R&D for Fusion Energy
Helion announced that it signed a letter of intent with the U.S. Department of Commerce for up to $175 million through the CHIPS Research and Development Office to develop high-voltage semiconductors for its pulsed-power fusion systems, matched by $50 million in private investment for $225 million in total. More than $100 million would fund semiconductor and advanced-packaging R&D in Everett, Washington, with more than 100 jobs, and about $80 million would support tooling and process work with an industry partner in Burlington, Vermont. Commerce would receive a $175 million non-voting interest in Helion based on its most recent Series G valuation upon a definitive award agreement.

Israel’s Energy Ministry Publishes Draft National Fusion Strategy Weighing Up to NIS 2 Billion in Public Investment
Israel’s Ministry of Energy and Infrastructure released its first national fusion strategy for public comment, laying out three investment packages that range from NIS 50 million to 100 million (roughly $13 million to $27 million) over five years to NIS 1 billion to 2 billion (up to about $647 million) over seven to ten years. The middle package, NIS 500 million to 1 billion (up to about $322 million), would build a research center and open national facilities such as the SARAF accelerator at Soreq to private industry, while the largest would add a coordinating authority and a possible experimental reactor built with an international partner. The ministry warns that inaction in the coming years risks missing the window in which global fusion supply chains are forming, and Israeli compact fusion developer nT-Tao said it welcomes the strategy.

Cambridge Heavy Industries Raises $6.2 Million Seed Round to Advance Modular, Benchtop-Scale Fusion System Technology
MIT Nelson Lab spinout Cambridge Heavy Industries announced a $6.2 million seed round led by Azolla Ventures, with participation from Overlay, the Mass Clean Energy Center, the Neta Foundation, the Rothrock family and P3A. The Cambridge, Massachusetts company builds compact fusion machines that use commercial lasers, reports a prototype that reached 350 GPa using tens of millijoules of energy, and plans to start by producing tritium, which it says sells for about $33 million per kilogram. The funding will go toward building its first compact reactor in Massachusetts, buying an ultrafast laser, optics and neutron diagnostics, and expanding the team.

NeoVolta Launches a Third-Party-Ownership Financing Platform to Drive Residential and C&I Energy Storage Revenue Growth
San Diego-based NeoVolta announced an initial $40 million third-party-ownership financing platform for residential storage and solar-plus-storage, backed by a tax equity partner and expected to support about 700 projects. Luminia, NeoVolta’s financing and deployment partner, helps manage the platform, and the California pipeline includes contracted projects for income-qualified homeowners eligible for the state’s Self-Generation Incentive Program rebates. Commercial and industrial offerings are expected to follow as NeoVolta’s Pendergrass, Georgia manufacturing facility ramps, and its NV Wave residential battery will be added starting in November.

Startorus Fusion Magnet Sets Record 63 T/s Field Change Rate in Cryogenic Testing
Shaanxi Startorus Fusion Technology announced that it completed cryogenic testing of the central solenoid model magnet for its planned CTRFR-1 spherical tokamak, reporting a magnetic field change rate of 63 T/s that it describes as a global record for known fusion superconducting magnets. The test also reached a 10 T peak field at 20 K with a cryogenic current of 6,350 A, and next steps include tests above 12 T and AC loss measurements. CTRFR-1 is targeted for 2028, with mass manufacturing of poloidal field model coil magnets in Shanghai slated to begin in 2027 and a pilot plant targeted for 2032.

Stellios Technologies Emerges From Stealth With $5 Million Pre-Seed Round for High-Power Microwave Systems and Compact Fusion
Stellios Technologies emerged from stealth with an oversubscribed $5 million pre-seed round backed by Sofreh Capital, Wave Function Ventures, Cubit Capital, Mana Ventures, Atoms & Bits, Output Capital and Erik Bruckner, and says it held $2 million in signed customer purchase orders before raising outside capital. Its first products are megawatt-class high-power microwave systems for counter-drone and counter-electronics missions, with industrial uses that include plasma heating for fusion research. Founder and CEO Richard Starson, a Caltech PhD who co-founded Escape Dynamics, is also pursuing an aneutronic compact fusion research effort called Stellios Core, which remains a research goal rather than a demonstrated product.

The Fusion Report has written a number of pieces on artificial intelligence (AI) and AI data centers over the past year, and one topic which is particularly timely is the question of, “what is powering these data centers?” While the use of clean (‘green’) energy for AI data centers originally had gotten significant headlines, the reality is that most of the power for AI data centers isn’t renewable energy sources. The purpose of this article is to look into what are the likely sources of power, now and in the future, for data centers in the United States, and whether that power will be from the grid or behind the meter.

Sponsored

Today: Powering AI Data Centers Is A Mix of Sources

Electrical power in the United States today is a mix of sources. Depending on what data you look at, there is variation between the time scale (2023 is the latest data from the US-EIA), and what is being measured (usage vs. generating capacity). The largest source of electrical power in the US is based on natural gas, contributing roughly 41% of all electricity generated. The second largest source is renewables, which are roughly one-quarter of all electricity, followed by nuclear power at 18% and coal at 17%. Petroleum and other gases are less than 1%. Within renewables, wind power today provides 10%, followed by solar at 8.6%, hydropower at 5%, and biomass and other renewable sources at 1%.

Since US AI data centers in 2026 are powered primarily by the regional electric grids to which they connect, their actual fuel mix varies by location rather than by a single dedicated “AI” supply. At the national level, natural gas is the largest source: the Energy Information Administration projects it will provide about 40% of U.S. electricity generation in 2026. Nuclear supplies roughly 18%, renewables about 25%, and coal about 16%; renewables combine wind, solar, hydropower, biomass, and geothermal generation. Recent estimates specific to US data centers likewise put natural gas above 40% of their electricity supply, followed by renewables at about 24%, nuclear at about 20%, and coal at about 15%. The reality is that even if a hyperscaler has signed a renewable-power contract, its servers generally still consume the instantaneous grid mix unless matched by dedicated or time-aligned clean generation.

US AI Data Center Electricity by Fuel Mix, 2030

By 2030, U.S. AI data centers are likely to be supplied by a more diverse electricity mix than today, but natural gas is still expected to be the largest incremental fuel source. AI facilities need very high load factors and near-continuous power, while grid interconnection and transmission additions often take years; as a result, gas-fired combined-cycle plants, gas turbines, and—in some cases—on-site gas generation are expected to provide firm capacity for much of the new load. The International Energy Agency projects that roughly 15–27 GW of on-site natural-gas capacity could be serving data centers globally by 2030, mostly in the United States. In one U.S. base-case projection, gas generation associated with data centers rises from just under 100 TWh in 2024 to well above 200 TWh by 2030. Gas’s role does not mean every AI campus will burn gas on site: most will still draw from regional grids, whose marginal supply is often gas-fired, particularly during periods when renewable output is low or demand is high.

The biggest factor making forecasting of the fuel types to power AI data centers in the next few years difficult is the speculative nature of US plans for building new power plants, especially “behind the meter” ones. According to the Global Energy Monitor, the United States has 378GW of new gas generation plants under proposal, more than the next nine countries combined. These projects range in planning levels from the early planning stages to construction, but many of these projects don’t yet have vendors named for gas turbines or other critical components; moreover, only about 30% of these early projects historically actually reach conclusion. The one confounding issue for US data centers is that, despite the large supply of natural gas in the United States, the data centers will be forced to compete with an export market in Asia and Europe, which is very hungry for liquefied natural gas.

Additionally, by 2030 behind-the-meter power plants for US AI data centers will have become popular as a way to bypass interconnection agreement delays. These systems will take the form of campus-scale microgrids, built next to or within large data-center campuses to bypass long interconnection queues and provide highly reliable power; examples of these include Elon Musk’s xAI facility in Memphis, Tennessee, the OpenAI/Stargate facility in Abilene, Texas, Meta’s Prometheus data center in Ohio, and others. The core of these microgrids will most often be modular natural-gas generation—either gas turbines for large blocks of capacity or reciprocating engines for faster, phased deployment—supplemented by medium-duration batteries, switchgear, controls, backup generation, and a grid connection where available. Some sites will also add solar, fuel cells, and potentially geothermal or other local resources, but these will generally complement rather than replace firm generation because AI loads operate around the clock. Forecasts suggest behind-the-meter solutions could supply roughly 25–33% of incremental data-center demand through 2030, while industry estimates indicate that 38% of data centers may use on-site generation at least in part; however, the exact share will depend on regional permitting, gas-pipeline availability, emissions rules, and the speed of transmission and utility interconnection upgrades. Nuclear solutions may appear in a limited number of projects, but widespread small-modular-reactor (SMR) deployment is more likely to be a post-2030 development.

Conclusion: Powering AI Data Centers in 2035 and Beyond

Beyond 2035, AI data centers will most likely continue to be powered by a portfolio of electricity sources, rather than a single fuel. Wind, solar, hydropower, and—in some regions—geothermal are poised to provide most of the incremental energy because they are generally the fastest-growing and often lowest-cost new generation options. For AI operators, however, buying annual renewable energy is not sufficient: training clusters and inference services need electricity every hour, and their concentrated demand can be measured in hundreds of megawatts or even gigawatts at individual campuses. That makes transmission, grid-scale batteries, long-duration storage, demand flexibility, and geographically diversified renewable procurement as important as the generating fuel itself. The International Energy Agency projects renewables to meet most additional global data-center electricity demand through 2035 and expects them to remain the dominant source of incremental supply afterward; its base case still has total data-center demand above 1,300 TWh globally in 2035.

The balancing and reliability layer will determine how clean this mix becomes. Existing and new nuclear power—including life extensions, large reactors where available, and potentially small modular reactors later on—offers firm, low-carbon output that matches data centers’ around-the-clock load, so it is likely to gain strategic importance after 2035. Natural gas will nevertheless remain a substantial bridge and backup fuel, especially in U.S. regions with constrained transmission or slow clean-power permitting; gas plants may increasingly operate fewer hours but provide capacity during low-wind, low-solar periods. Additionally, fusion power is also a potential dark horse in the AI Data Center electricity race, though even an early rollout of fusion is not likely before 2040 because of the new technology’s inherent technical and schedule risks. The most plausible post-2035 outcome is therefore renewable-heavy electricity backed by nuclear, hydro, storage, and flexible gas—rather than an entirely renewable or entirely nuclear AI grid. Even in the IEA’s scenarios, fossil fuels still account for roughly 40% of data-center electricity generation in 2035, underscoring that policies, grid build-out, and the commercial success of firm clean technologies will shape the longer-term result.

Today, The Fusion Report carries an announcement from UK artificial intelligence (AI) startup Zenithon. The company is a UK-based AI startup which incorporated in 2025, and raised a $10M initial fund last month. The company builds physics-grounded machine learning models that make simulation for fusion, aerospace, and semiconductor development dramatically faster. It was founded by Alex Higginbottom, who left a University of Manchester PhD in AI for complex systems to pursue the venture.

Sponsored

The Focus of Zenithon’s AI Models

Zenithon’s AI models focus on “world models for extreme physics”: machine-learning systems designed to predict how highly complex physical systems behave under demanding engineering conditions. Rather than generating text or images, the company’s models act as fast surrogates for computationally expensive physics simulations, allowing engineers to rapidly test how changes to a design affect performance. Zenithon says its models can evaluate roughly one million design points in the time a conventional simulation takes to evaluate one.

The company’s initial technical focus is on fusion energy, where modeling plasma behavior, heat and particle transport, and reactor performance remains a major bottleneck. Its founders’ background includes machine-learning surrogates for gyrokinetic transport calculations, which describe how heat and particles move through a fusion plasma. By reducing simulation times from months to seconds, Zenithon aims to help fusion developers screen more reactor configurations, optimize components, and shorten design cycles.

Beyond fusion, Zenithon is extending the same approach to spacecraft, rockets, and semiconductor manufacturing. These fields share a common problem: expensive experiments, long simulation runtimes, and enormous design spaces. Zenithon’s models are intended to learn from both simulation data and real-world experimental results, providing predictions alongside uncertainty estimates so engineers can explore unfamiliar designs while retaining a measure of confidence in the output.

Why Could This Be Important To Commercial Fusion Energy?

Zenithon’s modeling capabilities could matter to fusion because they target one of the field’s central bottlenecks: the time and cost of simulating turbulent, magnetically confined plasma. Conventional high-fidelity simulations can take hours, days, or even months per design case, which limits how many reactor geometries, magnetic configurations, and operating scenarios engineers can realistically test. Zenithon claims its AI “world models” can evaluate about one million design points in the time a traditional simulation evaluates one, potentially turning plasma modeling from a months-long research exercise into a near-real-time design tool.

Fusion devices such as tokamaks and stellarators involve tightly coupled physics: magnetic fields, plasma turbulence, heat and particle transport, wall interactions, and engineering constraints all affect performance. Zenithon’s founders began with machine-learning surrogates for gyrokinetic transport calculations, which model how heat and particles move through a fusion plasma. If such surrogates are accurate enough, engineers could rapidly screen candidate designs, identify promising operating regimes, and reserve expensive high-fidelity simulations for the most credible options.

The company says its models are trained on both simulation data and real-world experimental results, and provide uncertainty estimates with predictions. That could make AI outputs more useful for engineering decisions: instead of a black-box prediction, teams could see where the model is confident and where additional simulation or experimental validation is needed. In fusion, where prototype-scale devices are extraordinarily expensive and experimental time is scarce, this kind of uncertainty-aware modeling could help prioritize diagnostics, campaigns, and design changes. Examples of these models’ output can be seen at zenithon.ai.

Conclusion: Some Important Caveats to Zenithon’s Claims

The potential for Zenithon’s models to enable breakthroughs in fusion is significant, but they depend on validation. Independent reporting notes that Zenithon’s early published work described validation against unseen simulation data and near-real-time gyrokinetic-turbulence predictions, while the company’s broader “million-times faster” claim and use of live experimental data at production scale remain claims to be demonstrated at commercial scale. Secondly, the ability of the speed-ups to be important is somewhat dissipated by the fact that the building of fusion hardware that can make use of those breakthroughs still takes a significant amount of time, something which modeling has limited impacts on. In short, Zenithon could be important if its models preserve enough physical fidelity while delivering orders-of-magnitude speedups—because that combination would let fusion developers explore far more of the design space before committing to costly hardware.

This is the fifth article in The Fusion Report series on critical supporting technologies for commercial fusion energy. In today’s article we looked at how commercial fusion energy machines heat plasma high temperatures and how they measure the temperature and diagnose the state of the plasma. But first, let’s start with which types of fusion systems need these high-energy plasma heating and diagnostic systems, and which ones don’t.

Why Both MFE and HFE Systems Need High-Energy Plasma Heating

All fusion energy systems, whether magnetic fusion energy (MFE), inertial fusion energy (IFE), or hybrid fusion energy (HFE) systems, utilize plasma diagnostics systems because fusion occurs only within a narrow and rapidly changing range of plasma conditions that cannot be observed directly. Diagnostics provide the measurements—such as temperature, density, pressure, composition, confinement, magnetic structure, fusion-product yield, and interactions with reactor walls—needed to determine whether a device is approaching the conditions for net energy production or losing performance through turbulence, impurities, fuel imbalance, or instability. They are equally essential for real-time control: a small disturbance can degrade confinement or, in some concepts, damage plasma-facing components, so operators and automated systems need fast, reliable signals to adjust heating, fueling, magnetic fields, compression, or shutdown systems. Across magnetic-confinement, inertial, magneto-inertial, and other fusion approaches, diagnostics also validate simulation models and scaling laws, turning individual experimental shots into engineering knowledge for the next machine. In effect, diagnostics are the plasma’s instrumentation system: without them, developers could neither prove fusion performance nor operate a future plant safely, reliably, and economically.

The same is not true for plasma heating systems. While IFE systems utilize the energy of the particles themselves as they collide together, MFE and HFE systems must supply that energy externally. Magnetic-fusion systems heat plasma through a combination of internal electrical heating and external energy injection. In tokamaks, an induced electric current initially heats the conducting plasma through resistive, or ohmic, heating; however, this becomes insufficient at the higher temperatures needed for deuterium-tritium fusion. Devices therefore add high-energy neutral-beam injection, in which fast neutral atoms penetrate the magnetic field and transfer energy through collisions, and radio-frequency heating, which uses electromagnetic waves tuned to resonate with electrons or ions—such as electron-cyclotron and ion-cyclotron heating. Once fusion reactions become significant, energetic alpha particles (the nucleus of a helium-4 atom, also known as ‘fusion ash’) can also deposit energy in the plasma, helping sustain its temperature. Hybrid systems, often called magneto-inertial or magnetized-target fusion systems, combine magnetic insulation with rapid compression: they first form and sometimes preheat a magnetized plasma using electrical currents, particle beams, lasers, or radio-frequency power, then compress it with a rapidly imploding metal or liquid liner, plasma shell, laser pulse, or other driver. Compression raises both density and temperature while the embedded magnetic field reduces heat losses, allowing hybrid concepts to pursue fusion conditions with shorter confinement times and, potentially, lower driver requirements than purely inertial approaches.

Plasma Heating Systems: Ohmic Heating, Neutral Beam Injection, RF/Microwave Heating, and Laser Heating

Ohmic heating is the first heating mechanism used in many magnetic-confinement fusion devices, particularly tokamaks. A transformer-like central solenoid induces an electric current in the plasma; because the plasma has finite electrical resistance, that current produces ‘joule heating’, much like an electric heating element. Ohmic heating is useful for initiating the plasma discharge and raising temperature during startup, while the plasma current can also contribute to magnetic confinement. Its effectiveness falls as the plasma becomes hotter, however, because plasma resistivity declines with temperature. Consequently, ohmic heating alone cannot bring a reactor-scale deuterium-tritium plasma to the roughly 100-million-degree conditions needed for substantial fusion power, making auxiliary heating systems essential.

Neutral beam injection (NBI) heats plasma by accelerating hydrogen or deuterium ions to very high energy, neutralizing them so they can cross the confining magnetic field, and firing the resulting fast-atom beam into the plasma core. Once inside, the neutral atoms are ionized, magnetically trapped, and slowed by collisions with plasma electrons and ions; their directed kinetic energy is thereby converted into plasma heat. NBI is also valuable because it can drive plasma current, modify the current profile, and impart rotation that can improve stability and confinement. It is typically the highest-power auxiliary heating method in large tokamaks: ITER’s two planned heating neutral-beam injectors are designed to provide 16.5 MW each, or 33 MW combined, using 1 MeV deuterium beams.

RF and microwave heating systems deliver electromagnetic energy at frequencies selected to resonate with charged particles moving in the magnetic field. Ion-cyclotron resonance heating uses radio waves, typically in the tens-of-megahertz range, to transfer energy mainly to ions; electron-cyclotron resonance heating uses millimeter-wave microwaves—170 GHz in ITER—to heat electrons locally and precisely. Lower-hybrid waves, generally in the gigahertz range, are particularly useful for driving current as well as heating. Because wave absorption can be targeted by particle species, location, and magnetic-field conditions, RF systems support not only bulk heating but also current-profile control and the suppression of localized instabilities. ITER plans 20 MW each of ion-cyclotron and electron-cyclotron heating, alongside its neutral beams, illustrating why practical magnetic-fusion plants will rely on complementary heating systems rather than a single technique.

Lasers can heat plasma in both MFE and HFE systems, but they play very different roles. In conventional MFE devices—such as tokamaks and stellarators—lasers are not normally a primary bulk-heating technology; the main tools are ohmic heating, neutral-beam injection, and resonant RF or microwave systems. In HFE systems, especially magneto-inertial or magnetized-target fusion concepts, lasers can be much more central: they may preheat a magnetized fuel plasma before rapid compression raises its density and temperature to fusion-relevant conditions. An example is the Sandia National Laboratories MagLIF concept: an axial magnetic field magnetizes the fuel, a laser preheats the cylindrical fuel plasma to roughly 100–300 eV, and a pulsed-power-driven metal liner implodes to compress and further heat it. The laser supplies an initial temperature and pressure boost, while magnetic insulation and liner compression help retain and amplify that energy. Other approaches utilize particle beams, plasma guns, or other approaches in a similar manner to help heat the plasma.

Plasma Diagnostic Systems: Seeing What’s Going On

Plasma diagnostic systems are the measurement infrastructure that enables scientists and operators to understand, control, and improve fusion plasmas. Because a fusion plasma is far too hot and energetic for conventional probes to contact directly, diagnostics infer its condition from magnetic fields, electromagnetic radiation, injected laser or microwave signals, escaping particles, and fusion neutrons. Together, these instruments measure essential quantities including electron and ion temperature, density and pressure profiles, plasma current, shape and position, magnetic equilibrium, rotation, radiation losses, fuel ratio, helium-ash concentration, and impurities released from plasma-facing components. ITER, for example, is designed to use roughly 60 diagnostic instruments to measure 101 parameters; magnetic sensors determine plasma current, energy, shape, and position, while Thomson-scattering and interferometry systems measure temperature and density profiles.

Diagnostics are also indispensable for real-time control, machine protection, and verification of fusion performance. Neutron cameras, spectrometers, and flux monitors measure neutron production and thereby provide a direct indication of fusion power, while spectroscopy and bolometry identify impurity sources and map radiated energy losses in the core and divertor. Fast signals from these and other systems feed plasma-control algorithms that can adjust heating, fueling, magnetic fields, and exhaust conditions, or initiate protective actions when an instability or off-normal event threatens the device. Over repeated experiments, diagnostic data test physics models of transport, turbulence, stability, and plasma–material interactions, helping developers translate short experimental pulses into the validated operating knowledge required for reliable fusion power plants. ITER’s control architecture is specifically designed to process diagnostic measurements in real time, including calculating electron temperature from Thomson-scattering data for use in plasma operation.

Conclusion: Plasma Heating and Diagnostics Are Critical For Fusion

Plasma heating and plasma diagnostics are complementary foundations of every fusion-energy system: heating supplies the energy needed to raise fuel to fusion-relevant temperatures—typically on the order of 100 million degrees Celsius for deuterium–tritium reactions—while diagnostics determine whether the plasma is actually reaching, sustaining, and safely operating within those conditions. Ohmic current, neutral beams, RF and microwave waves, lasers, compression, or other drivers must deliver energy efficiently without causing excessive instability or damaging plasma-facing components. Diagnostic instruments then measure temperature, density, confinement, fuel composition, impurities, radiation losses, plasma position, instabilities, and fusion-neutron output, providing the feedback needed to optimize heating and control the machine in real time. This closed loop is essential: without adequate heating, fusion conditions cannot be achieved; without diagnostics, developers cannot verify performance, identify losses, protect equipment, or convert experimental results into a reliable, commercially viable power-plant design.

The Fusion Report today brings a significant piece of engineering and physics news regarding tokamaks – one of the most utilized machines for commercial fusion energy. While many companies have built fusion solutions using tokamaks, most notably Commonwealth Fusion Systems (CFS), they have historically been troubled with plasma instabilities. A particular instability, known as edge-localized modes (ELMs), are very problematic, as they release both heat and high-energy particles from magnetic confinement. Work by the DIII-D National Fusion Facility regarding the use of negative triangularity (reshaping the magnetic confinement region so that the curve of the ‘D’ is facing inward instead of outward) eliminates ELMs, resulting in a more stable plasma.

DIII-D will collaborate with fusion engineering company Kyoto Fusioneering to manufacture the first wall tiles needed to build the negative triangularity divertor, and work with Columbia University to use the new divertor to tackle remaining issues including core-edge integration, power exhaust diagnostic composition, and the performance of plasma-facing components. “High neutron loads and heat fluxes are among the most critical challenges to the viability of fusion power plants. By leveraging Kyoto Fusioneering’s engineering and manufacturing capabilities to develop and test novel divertor concepts at DIII-D, we aim to contribute to the early realization of commercial fusion power,” said Bibake Uppal, president of Kyoto Fusioneering America. Below is the press release regarding the use of negative triangularity approaches to provide for more stable plasma confinement.

Press Release: DIII-D Team Wins Prestigious Plasma Physics Award for Demonstrating Viability of Negative Triangularity for Commercial Fusion

The race toward commercial fusion energy requires scientists and engineers to continue rapidly solving key challenges and optimizing solutions for fusion science and technology. The goal is the development of a device that operates continuously with high performance and stability. A multi-institutional team from the U.S. Department of Energy’s DIII-D National Fusion Facility recently won the John Dawson Award for Excellence in Plasma Physics Research from the American Physical Society for substantially advancing the negative triangularity approach for fusion.

Development of Negative Triangularity at DIII-D

Tokamaks, donut-shaped devices that contain 100,000,000oC plasmas with magnetic fields to drive fusion reactions, are one of the leading designs for commercial fusion pilot plants (FPPs). Most development of operational approaches for tokamaks has focused on optimizing operation in positive triangularity, a D-shaped plasma with the curve facing the outer edge of the device. Positive triangularity plasmas can achieve high confinement, an important feature for power generation; however, these plasmas have notable challenges related to power handling that have yet to be solved for commercial power generation. Operation with lower confinement improves heat flux and power handling issues, with the trade-off of lower performance that would ultimately mean less power output for electricity generation.

An alternative solution is to move away from the positive triangularity shape altogether, operating the plasma in a reverse-D shape called negative triangularity (NT). Building on foundational work performed at TCV, Max Austin, research scientist at University of Texas at Austin, and Alessandro Marinoni, assistant professor at University of California, San Diego, led initial NT experiments at DIII-D. They showed high performance without the development of instabilities that cause large releases of particles and heat (known as edge-localized modes, ELMs). These initial experiments motivated a dedicated campaign to fully explore NT, developing strongly shaped diverted plasmas under FPP-relevant conditions.

“The high degree of flexibility at DIII-D – not only in our ability to adjust our research program to meet priority needs but also in the ability to change physical and operational features of our tokamak – is essential for supporting innovation in fusion science and technology. We can make changes and comprehensively measure their effects on commercially relevant fusion plasmas, accelerating progress toward identifying the best solutions for fusion energy,” said Richard Buttery, director of the DIII-D National Fusion Facility. “In the case of negative triangularity, we could rapidly implement a change to the device to support a focused campaign that dramatically advanced our understanding of this unique approach.”

Negative Triangularity Campaign

Beginning in 2022, the team exploited the flexibility of DIII-D as a research platform to embark on a dedicated diverted NT campaign. Filippo Scotti, research scientist at Lawrence Livermore National Laboratory, worked with the General Atomics Engineering team for DIII-D to replace the standard tile armor in the lower divertor with specialized tiles designed for NT operation. Although DIII-D was designed for operation in positive triangularity, dedicated efforts led by Al Hyatt, chief physics operator on the DIII-D Operations team at General Atomics, and supported by Oak Nelson, associate research Scientist at Columbia University’s School of Engineering and Applied Science, developed an approach to operate the tokamak in NT scenarios with the new armor.

The dedicated NT campaign was led by Carlos Paz-Soldan, associate professor of Applied Physics at Columbia University’s School of Engineering and Applied Science, and Kathreen Thome, staff scientist at General Atomics, with Austin serving as the campaign deputy. They spent weeks exploring diverted plasmas with strong NT shaping to validate this approach for tokamak operation. The plasmas produced during this campaign consistently exhibited high confinement and strong stability over a vast operational space.

“The campaign was incredibly exciting – we produced high-performance plasmas on DIII-D for weeks without seeing a single ELM. No one had ever accomplished something comparable before,” said Thome. “This campaign was the most fun, and maybe also the most stress, I’ve ever experienced in the control room. NT is a potentially revolutionary way to design and operate a tokamak for fusion energy, and helping lead this effort to demonstrate the commercial potential of this unique scenario was an extraordinary moment in my career.”

Key experimental contributions were also made by Scotti, Nelson, and Marinoni. Scotti conducted the first experiments evaluating heat exhaust, achieving detached divertor conditions during NT operation of a tokamak for the first time and motivating the design of a closed pumped NT divertor, an effort led by Scotti, for installation on DIII-D in the near future. Building on Marinoni’s earlier work identifying a mechanism associated with ELM inhibition in NT plasmas, Nelson confirmed its occurrence across a much broader range of conditions. Marinoni also explored how the observed NT confinement would extrapolate to an FPP. The work led by these scientists, showing simultaneous achievement of high performance in the plasma core with mitigated heat flux at the divertor, demonstrates that NT operation can achieve key features required for future FPPs.

“The US Department of Energy’s Fusion Science and Technology Roadmap calls for advancing confinement approaches that combine high performance, stability, and effective power exhaust. DIII-D’s negative triangularity research is making important progress on that challenge, while growing engagement with the private sector is helping move promising concepts toward commercial fusion,” said Matthew Lanctot, acting director for Fusion Energy Research Division, DOE Office of Science.

Moving Negative Triangularity Forward

The NT research performed at DIII-D has established momentum for the development of NT-based machines around the world. Through 2026, DIII-D team members continued to advance the NT approach, optimizing plasma shape and further understanding NT confinement and stability properties in dedicated experiments. In the near future, DIII-D will be installing an NT-optimized divertor through a public-private partnership with Kyoto Fusioneering to advance this operational approach for commercialization. The innovative research performed at DIII-D today continues to develop the solutions needed to mature fusion into the energy source of the future.

###

About the DIII-D National Fusion Facility.

DIII-D is a highly flexible, comprehensively diagnosed magnetic fusion research facility and has been the site of numerous pioneering contributions to fusion energy science. As a fusion testbed enabling critical science and technology advances, DIII-D continues the drive toward practical fusion energy with critical research conducted in collaboration by nearly 1000 scientists representing over 125 institutions worldwide. Research at DIII-D, a U.S. Department of Energy Office of Science User Facility, is open to all interested parties. For more information, visit d3dfusion.org.

Why DIII-D’s Negative Triangularity Work Is Important

The classical tokamaks utilize a D-shaped plasma space, and the curve of the ‘D’ faces outward from the central solenoid. The issue that such tokamaks have historically run into is plasma instability that causes a release of both high energy particles and heat from the magnetic confinement regime, which are known as edge-localized modes (ELMs). Negative triangularity, which puts the curve of the D on the inside of the tokamak rather than the outside, has been shown by DIII-D’s researchers to neutralize ELMs, negating one of the most difficult issues that have plagued tokamaks for commercial fusion energy. Moreover, the DIII-D researchers were able to reconfigure their tokamak for negative triangularity through modification of the existing first wall. Finally, the inclusion of Kyoto Fusioneering as part of the team to productize the negative triangularity divertors and related equipment potentially enables other existing tokamaks to be similarly configured for negative triangularity, solving a critical issue for commercial fusion energy development.

The Fusion Report today has a couple of stories about Type One Energy. First off, we will cover Type One Energy’s announcement today of the successful close of their $200M Series B funding round, led by Series A round investors Breakthrough Energy Ventures and Clutterbuck Capital, and joined by new investors Lowercarbon Capital, Siemens Energy Ventures, and SiteGround Capital. The Series B close comes shortly after Type One Energy received their initial fusion operating license for Project Infinity from the State of Tennessee. Following that, we will be speaking with Type One Energy’s Dr. John Canik, the co-founder and Chief Science and Engineering Officer (CSEO) of the start-up. But first, here’s an overview of Type One Energy’s approach to commercial fusion energy and the Series B funding press release.

Type One’s Commercial Approach to Fusion, Series B Funding Round

Type One Energy is pursuing commercial fusion through stellarator technology: a magnetic-confinement machine with complex, externally generated magnetic fields designed to hold a fusion plasma in a stable, continuous configuration. The company’s strategy, branded FusionDirect™, emphasizes using established materials and engineering practices, extensive industrial partnerships, and a staged demonstration path rather than waiting for wholly new enabling technologies. The company’s first power plant, Infinity Two, is a deuterium–tritium stellarator designed to produce roughly 800 MW of fusion power and deliver 400MWe of continuous, dispatchable electricity to the grid. Type One Energy sees the stellarators’ inherent steady-state operating capability and avoidance of the large plasma currents used in tokamaks can reduce disruption risk and better suit utility-scale, high-capacity-factor generation. The market acceptance of this is demonstrated by Type One Energy’s successful Series B funding round which closed today.

“The breadth and quality of investors in this funding round demonstrates growing support for our strategy to industrialize the commercial deployment of fusion energy,” said Christofer Mowry, Chief Executive Officer of Type One Energy. “The Series B financing enables us to remain focused on advancing our stellarator technology and Project Infinity design activities, while working with experienced industrial partners to deliver the first commercial fusion power plant at TVA’s Bull Run site.”

That approach is being put into practice through Project Infinity at the Tennessee Valley Authority’s (TVA) Bull Run Energy Complex in Clinton, Tennessee, near Oak Ridge. The first phase, Infinity One, is a 30% scale, non-power-producing stellarator prototype and workforce-training platform scheduled for commissioning in 2029; it is intended to validate key plant engineering, operations, maintainability, and reliability features before deployment of the electricity-generating plant. The subsequent Infinity Two project is planned as a 400MWe commercial fusion plant developed with TVA and engineering partner AECOM, with full startup targeted for 2034, subject to final approvals. Construction on both Infinity One and Two is scheduled to begin in early 2027. Reusing the former coal-power-plant site gives the project access to existing energy generation infrastructure and supports TVA’s need for future carbon-free power.

“Fusion is approaching the point where deployment, not discovery, defines the challenge,” said Carmichael Roberts of Breakthrough Energy Ventures. “Type One Energy brings together the necessary business and technical leadership, optimized stellarator technology and industrial relationships that will help move fusion from breakthrough science to commercial power plants. We are pleased to deepen our support as the company advances Project Infinity and its international partnerships.”

Interview With CSEO John Canik

Dr. John Canik is the co-founder and Chief Science and Engineering Officer (CSEO) of Type One Energy. John co-founded the company in February 2019 with fellow University of Wisconsin-Madison (UW-M) fusion energy scientists Chris Hegna, David Anderson, Paul Harris, and Randall Volberg. They were funded in 2023 by the Bill Gates-backed venture capital (VC) company Breakthrough Energy Ventures, enabling John to leave Oak Ridge National Lab and join Type One Energy full time.

The Fusion Report’s Interview with Dr. John Canik

The Fusion Report: There are significant positives and negatives for utilizing a stellarator versus a tokamak for commercial fusion. Which do you think are the biggest pluses and the biggest challenges of a stellarator?

John Canik: The big advantage of Stellarators is that they lend themselves very well to stable, steady-state operations. What you really want in a power plant is something that is just on and running and not really anything interesting happening in the interior of it. Both stellarators and tokamaks are shaped like a donut; they have components of the magnetic field that go the long way around the donut and components that go the short way around the donut. The fundamental difference is that in a tokamak, they make the field that goes the short way around the donut by running a current in the plasma itself. In the stellarator, you do all of that using the magnets that are on the outside of the plasma. So, they’re all funny shapes, they’re all complicated 3D structures. But in a stellarator, all of the current that makes the entire magnetic field is in superconducting magnets, not in the plasma. As a superconducting magnet, like an MRI, you just turn it on, and in principle, it can stay on. This is ideal when you are trying to run a power plant and you just want heat to come out steadily.

The Fusion Report: A lot has been written about the complex geometries that are required in a stellarator. How do you approach designing your geometries, and then validating them?

John Canik: As you said, the big challenge with the stellarator is that it has a complicated 3D geometry, making it harder to design. It takes more complicated simulations, and more complex tools to do those simulations. You have a larger design space to navigate, which can be a challenge. One of the reasons we launched Type One Energy in this modern era is that we have much, much better computational tools and platforms and codes. That enables us to do a lot of that design work in a far better and more thorough way than has ever been done before in stellarators.

Part of this design optimization is optimizing the shape of the magnets themselves, which is the source of this manufacturing complexity. One of the roots of improvement is the ability to optimize the shape of the magnet itself, such that it relaxes the tolerances involved in making that magnet. It can get quite substantial. Take the Wendelstein 7-X stellarator in Germany as an example. Had they done a computational study with modern tools, they would have been able to relax the tolerances on the W7-X magnets by something like a factor of 2 to 3. Which, when it comes to manufacturing, that’s just a gigantic improvement.

The Fusion Report: Can you tell me about Type One Energy’s experience building and operating stellarators?

John Canik: The subscale prototype that we’re building at Bull Run will be the first company-built stellarator, but our team has enormous experience building stellarators. One of our other co-founders, Dave Anderson, would say he has built two and a half stellarators in his career, including designing and building a machine called the Helically Symmetrical Experiment or HSX, which was the world’s first fully optimized stellarator. It was built at the University of Wisconsin-Madison. Thomas Sunn Pedersen, our CTO, designed and built a kind of a university-scale stellarator called the CNT at Columbia University in New York City and then he was also one of the directors at W7-X for over 11 years. We also have other folks that have been part of the design and construction of multiple fusion facilities—both stellarators and tokamaks. So, our team has enormous experience doing a lot of the things that we are setting out to do at Type One Energy.

The Fusion Report: Do you have a price per MWh target that you’re trying to achieve and you’re designed with the TVA, and if so what is it?

John Canik: In more general terms, we expect that fusion will certainly be able to get to significantly lower cost of electricity than fission, in part due to not having all the regulatory burdens. Eventually, we expect to be cost-competitive with fossils. But it’s true that the first fusion plant will be more expensive than the ultimate product. The first of kind is always more expensive than the nth of a kind.

The Fusion Report: What is it like working with a TVA on a project like this? What makes them an ideal utility partner from your perspective?

John Canik: For us, TVA is the ideal partner. Our business model is to be a technology OEM. We’re not intending to be vertically integrated. We don’t want to be the ones ourselves that build and own and operate power plants. That’s for our customers to do — and TVA fits this mold perfectly. They have incredibly deep experience building and operating big power plants. So, they are consistent with our business model, and of course, we always want to learn from their experience. They’ve built big commercial power plants before and we can learn from that.

The Fusion Report: Type One Energy just announced a $200 million Series B funding round, which is impressive but relatively small compared to the nearly $4B that Commonwealth Fusion Systems (CFS) has raised to date. Do you think that puts you at a disadvantage, and how do you overcome it?

John Canik: It is true we have not raised as much as, for example, CFS. We have not raised $4 billion, and it’s super impressive that they’ve done that. But it’s also kind of consistent with our business model. We’re not looking to raise the sort of money you need to be vertically integrated and build all the capability to do these things ourselves. We’re focusing on the fusion power core, the design work, and being able to deploy at scale. At the end of the day, we’re looking to the utility as our end customer. We don’t want to be the ones that build and operate the plants, in part because we don’t want to have to put up the front-loaded cost to do that. So our business model is inherently very capital efficient. That’s a feature, not a bug.

The Fusion Report: What’s the next big milestone you plan to achieve?

John Canik: Our focus right now is on what’s called Infinity One. This is the 30% scale prototype that we are planning to build at Bull Run. Our near-term milestones are focused on the design of that and getting through some of the design gates. We’re getting close to the start of construction at Bull Run, so preparing for that is the focus for us at the moment.

Type One’s Approach And Funding, and Why They’re Important

Type One Energy’s approach is important from two perspectives. First of all, they have taken an OEM approach and partnered with a large power company in the TVA, with extensive experience building power plants. Secondly, they waited until 3-D modeling and advanced computing tools were far enough along to enable the complex design of the stellarator’s high temperature superconductor (HTS) magnets, reducing the design and manufacturing risk of their program. The result is tighter designs, better room for tolerances, tighter budgets, and the ability to make it all work.

Type One Energy’s Series B funding success is equally important; by utilizing an OEM-focused model, they have spread the risk and the cost across multiple industrial partners whose business it is to build power plants. This significantly reduces the total capitalization required by Type One Energy, and better scales the speed with which they can commercialize fusion energy and go to market. Thanks again to Type One Energy, CEO Christofer Mowry, and CSEO John Canik for their time on this interview.

Things You Gotta Know

Commonwealth Fusion Systems Orders More Than 10,000 Kilometers of HTS Tape From Fujikura for ARC Power Plants
Commonwealth Fusion Systems agreed to purchase more than 10,000 kilometers of high-temperature superconducting (HTS) tape from Japan’s Fujikura Ltd., an order CFS calls the largest of its kind for HTS tape. The tape will go into ARC power plants, starting with the first grid-scale plant at the Fall Line Fusion Power Station in Chesterfield County, Virginia.

DIII-D Team Wins the APS John Dawson Award for Advancing Negative Triangularity Toward Commercial Fusion
A multi-institutional team at the U.S. Department of Energy’s DIII-D National Fusion Facility in San Diego won the American Physical Society’s John Dawson Award for Excellence in Plasma Physics Research for its work on negative triangularity, a reverse D-shaped plasma that pairs high confinement with strong stability. A dedicated campaign that began in 2022 produced high-performance plasmas for weeks without a single edge-localized mode (ELM), the bursts of heat and particles that can damage a device. DIII-D will next install an NT-optimized divertor through a public-private partnership with Kyoto Fusioneering.

ENN Reports Hydrogen-Boron Fusion Reactions on Its EXL-50U Spherical Torus
China’s ENN Group said it achieved hydrogen-boron fusion reactions on its EXL-50U spherical torus, which it calls the first time a commercial fusion company has done so on its own device. The team reported a reaction rate above 100 million reactions per second by combining high-energy neutral beam injection with radiofrequency waves. ENN is now building its next-generation Helong-2 device, expected to be completed in late 2027, with a goal of lighting the first lamp with hydrogen-boron fusion by 2030.

Japan’s Helical Fusion Plans Preliminary Power-On Tests of Its Helix HARUKA Demonstration Reactor in 2027
Helical Fusion said it expects to begin preliminary power-on tests in 2027 at Helix HARUKA, its demonstration reactor under construction at the National Institute for Fusion Science site in Toki. The company then plans to add a second coil and other major components, including a liquid-metal wall blanket, before testing the completed integrated facility around 2030, ahead of Helix Kanata, a larger reactor designed for 50 MW of net power.

Governor Newsom Signs SB 925, Directing a California Strategic Plan for Fusion Energy Development
California Governor Gavin Newsom signed SB 925, authored by Senator Jerry McNerney, which requires the state to create a strategic plan for fusion energy development, including a commercial fusion strategy and support for in-state fusion manufacturing. The plan builds on California assets such as the DIII-D National Fusion Facility in San Diego and Lawrence Livermore National Laboratory’s National Ignition Facility. Newsom also signed AB 2647, which directs an assessment of new nuclear power plants to help meet the state’s goal of 100 percent zero-carbon electricity by 2045.

Italy’s Senate Gives Final Approval to a Framework for Nuclear Fission and Fusion Energy
Italy’s Senate gave final approval, 81 to 51 with seven abstentions, to legislation directing the government to develop regulatory frameworks and a national program for nuclear fission and fusion energy, along with research and investment incentives. The government now has twelve months to issue the implementing decrees.

The Fusion Report recently had an opportunity to interview Dr. David Kirtley, the founder and Chief Executive Officer (CEO) of Helion Energy. Prior to founding Helion Energy in 2013, David served as a principal investigator, fusion research leader, and propulsion lead at Mathematical Science Northwest (MSNW). Before that, David was a research scientist at the Air Force Research Laboratory at Wright-Patterson Air Force Base in Dayton, Ohio, while concurrently working as a graduate student researcher at the University of Michigan Plasma Dynamics Laboratory. Before we get into the interview, we’ll discuss Helion Energy’s long-term approach to commercial fusion energy, which is fairly unique.

Helion Energy’s Long-Term Approach to Fusion Energy

Helion Energy is pursuing a compact, pulsed form of magneto-inertial fusion rather than the large, steady-state tokamak approach used by many public fusion programs. Its machines create two field-reversed configuration (FRC) plasmas—self-organized rings of hot plasma confined by magnetic fields—then accelerate them toward one another, merge them, and rapidly compress the resulting plasma with magnetic fields. The compression is designed to raise plasma conditions to fusion energies and temperatures, where a fusion reaction between deuterium and helium-3 (D-He3) can occur. While D-He3 requires higher temperatures than deuterium-tritium (D-T) fusion, its main reaction is aneutronic (i.e., it produces no neutrons). Also, rather than requiring a plasma to remain self-sustaining through ignition, Helion aims to repeat short pulses at high frequency; in principle, output can be adjusted by changing the pulse rate.

A second defining feature is Helion’s use of direct electricity recovery (DER), rather than a conventional thermally-based electricity generation cycle (steam turbines plus generators). D–He3 fusion produces energetic charged particles, allowing the expanding post-pulse plasma to push against the generator’s magnetic field and induce current in surrounding coils—an effect Helion compares with regenerative braking in an electric vehicle. This is meant to avoid the conventional chain of fusion heat, steam production, turbine rotation, and electrical generation, reducing conversion losses and plant complexity. The fuel cycle is also intended to reduce neutron production compared with deuterium–tritium fusion, potentially easing some material degradation, material activation, and shielding challenges. However, Helion’s approach (like all other commercial fusion energy companies) must still demonstrate reliable high-repetition operation, favorable net plant energy balance, durable components, and a practical helium-3 fuel supply at commercial scale.

On the positive side, Helion signed a power-purchase agreement in 2023 to provide Microsoft with electrical power. Helion plans to supply electricity from its planned Orion plant in Chelan County, Washington, beginning in 2028, with at least 50 MW delivered after an initial ramp-up. Constellation Energy is designated as the project’s power marketer, meaning it will handle the commercial/grid-market role associated with the electricity transaction. The project is intended to connect to the regional grid near Microsoft’s central Washington data-center operations, rather than being a dedicated behind-the-meter supply line. Helion also has a 2023 collaboration with steelmaker Nucor to deploy a proposed 500 MW fusion power plant supplying zero-carbon electricity directly to a Nucor steelmaking facility in the United States. Helion has characterized this as a customer agreement and development collaboration. Nucor also invested in Helion as part of the arrangement.

The Fusion Report’s Interview With David Kirtley (9/24/26)

The Fusion Report: So let’s get the obvious question out of the way. What caused the delay of Helion’s schedule, and how comfortable/confident are you with your new schedule?

David Kirtley: This wasn’t so much a delay as a misunderstanding, which I have seen incorrectly reported, even when we first announced the PPA. We didn’t have a milestone that said, “On day one, we are going to go to full power and produce energy,” or anything along those lines; a lot of work has to be done before you are ready to get to that point. A first of a kind technology takes some time to commission. That was built into the agreements with Microsoft and internal plans; we build the buildings, we build the technology, we start operating, and with time, we can ramp up to full power.

The Fusion Report: You have chosen deuterium-helium3 (D-He3) fusion as your long-term approach because it is aneutronic, and because it enables direct electricity conversion. How far are you proving D-He3 fusion’s viability, as well as the ability to breed helium3?

David Kirtley: Our deuterium-helium3 (D-He3) fusion validation starts out with deuterium-deuterium (D-D) fusion and with deuterium-tritium (D-T) fusion. We have been a leader in D-D fusion, first demonstrating D-D neutron production achieving it in 2011. Our goal is not to show that D-D fusion was realistic (the energy output is not high enough given the temperatures required); our goal, which was successfully achieved, was to show that helium3 can be made from D-D fusion. Earlier this year, we were the first private company to successfully demonstrate D-T fusion, proving the operation of our Polaris fusion machine. This was preparation testing for our D-He3 testing, first on Polaris, and then on Orion, once we achieve the required temperatures and pressures.

The Fusion Report: One of the coolest things about your approach is the use of direct electricity conversion (DEC). How efficient do you think DEC will be in converting fusion energy into electricity, and how much better is it than the conventional thermal approach to electricity generation?

David Kirtley: Great question. Our theoretical research shows that we should be able to get up to 90% efficiency in converting charged particles (which is primarily where the energy from aneutronic fusion goes into) directly into electricity. Even if we were to get half that in practice, it would still be significantly better than thermal cycles using heat exchangers, steam turbines, and generators. That would be a game changer for our industry, especially given that our DEC system is solid state, as opposed to the spinning machines required in a thermal cycle system, which would also require significant maintenance over its lifecycle.

The Fusion Report: Your Orion fusion machine is fairly small, targeting 50 MWe. I assume that you plan on making larger follow-on reactors. Is there a practical limit to an FRC machine’s effective size/power output, and if so what is it?

David Kirtley: It’s actually interesting in that a 50MWe power plant is probably the lower limit that an FRC machine can be done at, which is one of the reasons we chose it as our initial milestone. As you can imagine, efficiency also scales with size, so the larger the system gets, the more efficient energy-wise and cost-wise the system gets. So to answer your question, we expect to build larger systems which would be significantly more efficient from a cost and energy perspective as they scale.

The Fusion Report: What will the next big milestone be for Helion Energy and for Orion?

David Kirtley: There’s three areas that we’re really concentrating on in the near future:

We’re still learning a lot from Polaris and we’ll continue utilizing it while Orion is under construction.

We’re completing Orion’s infrastructure in Malaga, Washington. We’re completing the last building and getting the substation ready for operation.

On the business side, we’re focused on streamlining the manufacturing of the fusion equipment – things like how reliably and how fast we can build them. I’m looking at build rates every day, making sure we can build the Orion equipment quickly and reliably.

Helion Energy’s Funding To Date

Helion Energy has raised $1.5 billion in invested capital to date, making it the second-highest private lifetime funding of any commercial fusion energy company globally. Helion’s funding culminated in an oversubscribed $500 million Series G round which closed in September 2026. The Series G was initially announced at $465 million in June, led by Thrive Capital, and placed the Everett, Washington fusion developer at a $15.5 billion post-money valuation—almost three times its $5.4 billion valuation after the $425 million Series F in January 2025. Helion’s investor base includes financial and strategic backers, including Lightspeed Venture Partners, SoftBank Vision Fund 2, Mithril Capital, Capricorn Technology Impact Funds, Good Ventures, Nucor, and new Series G participants such as Lux Capital, Peak XV Partners, BoxGroup, Alta Park Capital, Anti Fund, and Bill Ford, the executive chair of the Ford Motor Company. Earlier financing included a $500 million Series E in 2021, alongside up to $1.7 billion in milestone-linked commitments, although those commitments should not be treated as cash raised unless and until the relevant conditions are met.

Future financing options are likely to combine further late-stage private equity with project-specific and strategic capital rather than relying on a single source. Helion can return to existing and new crossover, pension, sovereign-wealth, and growth-equity investors—groups it said participated in the expanded Series G—as it funds construction of Orion, scales production at its Omega capacitor factory, and works toward supplying Microsoft under its agreement for at least 50 MW. The company could also pursue customer-backed project finance, strategic investments from industrial power buyers or equipment partners, government grants and incentives, equipment or manufacturing financing, and eventually an IPO if technical milestones and commercial revenue substantially reduce perceived risk. Its prospective $1.7 billion of contingent Series E commitments remains a potentially meaningful source of capital, but it is conditional, while the feasibility and cost of financing a first-of-a-kind fusion plant will depend heavily on Helion demonstrating repeatable performance, regulatory progress, construction execution, and credible delivery against its power-purchase agreement commitments.

Conclusion: Helion’s Outlook On Commercial Fusion Energy

Helion Energy successfully tested D-T fusion early this year, achieving a record temperature of 150 million °C. After an extended campaign of more D-T fusion testing occurs and higher temperatures are reached, Helion Energy will be ready to perform D-He3 testing on Polaris, and then on Orion. The Orion testing will include electricity generation, initially connecting to the grid by the end of 2028, and achieving 50MWe of output after a ramp-up and testing period. We wish Helion Energy good luck on their testing campaign and achieving this historic milestone!

This is the fourth article in The Fusion Report series on critical supporting technologies for commercial fusion energy. In today’s article, we’ll look at the means to convert particle momentum and/or thermal energy from fusion reactions into electricity. But first we’re going to cover some basics about different approaches to fusion, because the type of fusion approach chosen impacts the method for performing energy conversion and fuel management.

Different Types of Fusion: Deuterium/Tritium and Aneutronic Versions

The most basic form of fusion for commercial fusion energy is deuterium-tritium (D-T) fusion. D-T fusion combines two isotopes of hydrogen: deuterium (one proton and one neutron) and tritium (one proton and two neutrons), to form helium (two protons and two neutrons with 3.5 MeV of kinetic energy), and a high-energy neutron with 14.1 MeV of kinetic energy. D-T fusion is the most popular candidate for commercial fusion energy because it requires the lowest fusion temperature (100M-150M degrees Celsius, or 13.6keV), plus the right combination of plasma pressure and time (known as the triple product).

Other relatively attractive candidates for commercial fusion are forms of aneutronic fusion, which produce no high-energy neutrons, an advantage we will touch on later. The aneutronic fusion requiring the lowest ignition temperature (58keV, or several hundred million degrees Celsius, about 2X what is needed for D-T fusion) is deuterium-helium3 (D-He3) fusion. D-He3 fusion produces a He4 (normal helium) nuclei and a proton, with a combined energy of 18.4MeV, and with no high-energy neutrons. The biggest supporter of D-He3 fusion is Helion Energy (#2 in total funds raised, with $1.5B in private funding), though others have shown interest.

Proton-boron11 (p-B11) fusion is attractive because it doesn’t require He3 (which is rare on Earth, but not on the Moon), but it requires nearly double the ignition energy (123keV) of D-He3. This is equivalent to an ignition temperature of well over a billion degrees Celsius, with some experts believing that temperatures of 2.3 to 4.6 billion degrees Celsius are required to yield a substantial net energy gain. With that said, there are at least three companies planning on utilizing proton-boron fusion, including TAE Technologies (#3 in total funds raised, with $1.3B in private funding and plans for an IPO), Marvel Fusion of Germany (#9 in total funds raised, with $415M in private funding), and HB11 Energy of Australia (#25 in total funds raised, with $32M in total funding).

The Impact of Fuel Choice on Fusion Energy Conversion to Electricity

Historically the conversion of energy (usually thermal energy to electricity, but also inertial energy to electricity) has been through rotating machinery – turbines and generators. In a D–T fusion reaction occurring in an inertial fusion energy (IFE) or magnetic fusion energy (MFE) machine, the 14.1 MeV neutron (roughly 80% of the total energy) enters the thermal blanket, where the neutron slows through collisions and deposits its kinetic energy into the blanket material as heat. In an MFE machine, the 3.5-MeV alpha particle produced by D-T fusion remains confined by the magnetic field and transfers its energy back to the plasma, helping sustain the extreme temperatures needed for continued fusion. In an IFE machine, the alpha particle is absorbed by the thermal blanket just like the neutron is.

The resulting heat is transferred to the coolant in the thermal blanket (usually water, helium, molten salt, or liquid lithium-lead), and then into a power-conversion system much like those used in other thermal power plants. In a steam-cycle design, the coolant produces high-temperature steam that spins a turbine connected to an electrical generator; gas-turbine or combined-cycle-like systems could also be used with sufficiently hot coolants. As such, a D–T fusion plant’s principal electricity pathway is indirect: fusion energy becomes neutron energy, then blanket heat, then mechanical turbine power, and finally electricity. This is largely the same mechanism that has been used for over 140 years.

For aneutronic fusion, there are a couple options: i) you can use the same indirect thermal cycle approach with a thermal blanket, steam turbine, and a generator; or ii) you can use direct energy conversion (DEC; first demonstrated a few months ago by Realta Fusion), because the energy released from the fusion appears principally as the kinetic energy of charged products (notably alpha particles), rather than in fast neutrons. Magnetic fields can guide those charged particles out of the fusion region and into an electrostatic or electromagnetic converter. In an electrostatic design, positively charged reaction products move against a deliberately imposed electric potential, slowing as they transfer kinetic energy to the electric field; electrodes then collect the resulting voltage and deliver current to an external circuit. Other candidate schemes include traveling-wave converters and magnetic-expansion systems that use changing electromagnetic fields to recover energy from directed charged-particle flows. The approach depends on specifics of the fusion mechanism used, whether MFE, IFE or a hybrid approach. However, while DEC has been demonstrated, it is not “production-ready” yet today.

Fuel Management Systems for Fusion

Fuel management in a D-T commercial fusion energy electrical plant depends on whether it is an IFE or an MFE/hybrid fusion plant. For MFE and hybrid fusion plants, fuel management is a tightly controlled, largely closed-loop process designed to deliver the correct fuel mixture to the plasma while minimizing the on-site inventory of radioactive tritium. Deuterium and tritium are stored, metered, analyzed, and injected into the fusion machine either as gas for edge fueling or as high-speed frozen pellets that penetrate farther into the plasma. Only a small fraction of injected fuel fuses during a pass, so the divertor and vacuum-pumping system continuously remove unburned deuterium and tritium, helium “ash” produced by fusion, and contaminant gases. A tritium-processing plant then purifies this exhaust, separates hydrogen isotopes to restore the required D–T ratio, and returns recovered fuel to storage and reinjection. For an IFE system, the machine must simply inject the frozen fuel pellets (for direct-drive IFE machines) or fusion-hohlraum combination pellets (for indirect-drive IFE machines) into the machine, at rates of up to 10 Hertz. The energy, in the form of high-speed neutrons and alpha particles, is absorbed by the thermal blanket, while the helium ash and other “burned” products (the fuel pellet shell, etc.) are evacuated between shots.

The fuel-management system in D-T fusion energy plants must also recover and manage tritium bred in the lithium-containing blanket surrounding the plasma. Fusion neutrons react with lithium to create replacement tritium, which must be extracted rapidly from the blanket medium, processed to remove impurities, measured for inventory control, and transferred into the fuel cycle. This requirement makes tritium self-sufficiency—not simply fuel injection—a core plant-design challenge: the breeding blanket, extraction equipment, pumps, isotope-separation units, containment barriers, and de-tritiation systems must collectively recover more tritium than the plant consumes and loses. Because tritium is radioactive and a concern for nuclear proliferation, fusion facilities use multiple physical containment barriers, continuous monitoring, air and water cleanup systems, and carefully managed storage inventories to limit worker exposure and environmental releases. Advanced concepts such as direct internal recycling aim to send purified D–T exhaust back to pellet-fueling systems without full isotope separation, potentially reducing processing equipment and the total tritium inventory held in the plant. Note that tritium recovery and management is not an issue that aneutronic fusion plants need to worry about.

Downstream, the fusion exhaust is separated by physical form and chemical species rather than discarded as a single waste stream. A representative IFE fuel-cycle scheme uses cryo-pumping to capture volatile gases, separates and vents helium ash, recovers deuterium and tritium for purification and reuse, and routes target-derived particulates to collectors. Solid or liquid debris can settle by gravity into collection regions, while captured particulate material may be oxidized or otherwise processed to release retained hydrogen isotopes before the non-fuel residue is disposed of or recycled. In liquid-wall concepts, the circulating protective liquid itself acts as a debris collector: it absorbs deposited material and transports it to external processing equipment. The goal is high recovery of valuable target materials and fuel, minimal tritium retention in waste, and a sufficiently clear chamber between pulses to sustain dependable power-plant operation.

Conclusion: Power Conversion and Fuel Management is Complex

One of the reasons that power conversion and fuel management for fusion energy is so complex is the unforgiving environment in which fusion is required to operate. Whether IFE, MFE or a hybrid fusion system, success requires a nearly perfect debris-free environment. At the same time, the tritium component in D-T fuel must be strictly controlled and accounted for. It is in many ways a chemistry problem as much as it is a physics problem or an engineering problem, and is an area that continues to develop as fusion energy nears production levels.

Things You Gotta Know

House Science Leaders Introduce the Bipartisan American Leadership in Fusion Act to Put $10 Billion Behind Fusion Commercialization
House Science, Space, and Technology Committee Ranking Member Zoe Lofgren (D-CA) and Research and Technology Subcommittee Chair Jay Obernolte (R-CA) introduced the American Leadership in Fusion Act, which would provide $10 billion in direct investments to accelerate the commercialization of fusion energy. The bill allocates $3.8 billion to build major new test facilities for fusion materials and fuel cycle challenges, $3 billion in forward funding for a new milestone-based demonstration program, and $2 billion for the existing Milestone-Based Fusion Development Program, plus $800 million for materials and fuel cycle research, $300 million for small and mid-sized test facilities tied to the 2026 DOE Fusion Science and Technology Roadmap, and $100 million to build supply chains for critical fusion technologies. Representatives Lori Trahan (D-MA), co-chair of the bipartisan Congressional Fusion Energy Caucus, Don Beyer (D-VA), and James Gallagher (R-CA) are co-sponsors. The Fusion Industry Association, which called late last year for a one-time $10 billion injection of public capital into partnerships with the private fusion industry, has endorsed the bill along with Commonwealth Fusion Systems.

nT-Tao and Orion Nuclear Energy Sign MoU to Bring 20 MW Compact Fusion Power to U.S. Data Centers and Critical Infrastructure
Israeli compact fusion company nT-Tao and Orion Nuclear Energy Corporation, a U.S. developer of advanced nuclear power projects, announced a strategic collaboration under a Memorandum of Understanding to identify, evaluate, and develop U.S. deployment opportunities for nT-Tao’s Compact Fusion System, an on-site system designed to deliver up to 20 megawatts of continuous electricity from a footprint small enough to sit at a customer’s own facility. Orion will serve as U.S. project development and commercialization partner, identifying customers and power offtakers, evaluating sites, structuring financing, and engaging utilities, infrastructure providers, and regulators, while nT-Tao contributes system design, performance data, and its development roadmap. Orion Chairman and CEO John Aycoth said Orion is already siting advanced nuclear projects alongside data centers and industrial loads across six states, and that a 20 MW system fills the gap between microreactors in the one to ten megawatt range and utility-scale SMRs that start well above fifty. The system is built on nT-Tao’s MEGA pulsed-power platform; its C3 prototype achieved first plasma earlier this year and was recently upgraded with a 50 MW pulsed-power subsystem. The companies plan to identify a limited number of initial U.S. opportunities for further evaluation.

ENN Breaks Ground on Helong-2, a Hydrogen-Boron Fusion Platform Targeting First Fusion Electricity in 2030
China’s ENN Group held a groundbreaking ceremony for Helong-2, its Spherical Ring Hydrogen-Boron Plasma Platform, at the group’s fusion R&D center in the Langfang Economic and Technological Development Zone in Hebei Province. Built by China Construction First Group Construction and Development Company, the project comprises more than ten core systems and is expected to be ready for equipment installation by the end of June 2027, with construction and commissioning complete and experiments under way by the end of 2027. Helong-2 is ENN’s third-generation device after Xuanlong-50 and Xuanlong-50U, and it anchors a three-step commercialization plan: achieve a hydrogen-boron fusion reaction on Xuanlong-50U in 2026, generate the first hydrogen-boron fusion electricity on Helong-2 in 2030, and reach low-cost generation and a demonstration reactor stage before 2035. ENN Chairman Yu Jianchao called Helong-2 a key engineering platform on the path to a fusion smart power plant. Hydrogen-boron, or proton-boron, fusion produces helium nuclei rather than high-energy neutrons, an approach also pursued by TAE Technologies in the U.S. and Marvel Fusion in Germany.

Proxima Fusion and Lower Saxony Plan Europe’s First Large-Scale Fusion-Grade HTS Tape Factory
Munich stellarator developer Proxima Fusion and the German State of Lower Saxony signed a memorandum of understanding to build Europe’s first large-scale production facility for fusion-grade high-temperature superconducting (HTS) tape. Around €140 million in investment is planned for the first two project phases running through the end of 2029, to be financed equally through private capital and public funding, with Lower Saxony contributing €21 million, just under 30 percent of the proposed public share. The agreement covers site assessment, a milestone-based support and funding structure, permitting, and partnerships with companies, municipalities, and research institutions across the state. Proxima expects to need about 20,000 kilometers of HTS tape for its Alpha demonstrator and a further 40,000 kilometers for its Stellaris power plant, and says it is in discussions with several international partners about bringing established HTS manufacturing expertise to Lower Saxony. HTS tape is currently manufactured at industrial scale primarily in Asia, and the plant would also serve grid, medical imaging, and other high-field magnet applications.

On several occasions, The Fusion Report has run stories about the human side of the challenges of building a successful commercial fusion energy industry. The travails of recruiting and hiring physicists, engineers, and even trades such as machinists, plumbers, and electricians to the commercial fusion industry are well-known. Traditionally, the approach has been to hire from the outside, and then integrate (assimilate?) the new workers into the ‘fusion mindset’. But when does it make sense to do the reverse – to hire those from the outside and utilize their mindset and skills to drive new solutions? That question will be the subject of today’s article.

Why Did Inertia Hire Sam Crisafulli From ASML? Skills and Mindset

An example of a successful hiring mindset is Inertia’s recent hiring of Sam Crisafulli as their Chief Systems Engineer. While Sam’s resume is certainly impressive (BSEE, PhD in Systems Engineering, 29 years of senior/executive experience including 12 years with ASML’s Plasma Generation and Energy Control Group), he didn’t have experience in a national laboratory such as Lawrence Livermore National Laboratory’s (LLNL) National Ignition Facility (NIF), which is where a number of Inertia’s founders worked. So why did Inertia hire Sam?

Inertia hired Sam as their Chief Systems Engineer because bringing a fusion power plant from demonstrated science to a deployable product requires tightly integrating complex physics, hardware, controls, manufacturing, and plant-level engineering. At ASML, Sam led a critical subsystem supporting the company’s extreme-ultraviolet lithography technology—an environment that demands exceptional systems engineering, reliability, precision, and coordination across highly interdependent technical disciplines. At Inertia, his role is to build the engineering architecture that connects the company’s science and technology into a commercial fusion power plant, helping turn individual technical advances into a machine that can be built, operated, maintained, and ultimately scaled.

What Skills Cross Over Into Fusion Companies Exceptionally Well?

Many of the capabilities needed for a successful commercial fusion energy company are already mature in adjacent industries. The limiting challenge for building a successful fusion energy company is not only achieving fusion conditions; it is combining a fusion core with dependable power conversion, fuel handling, maintenance, controls, manufacturing, and grid operation into a plant that can be built repeatedly and operated economically. That makes experience from semiconductor equipment, aerospace, fission and process industries, power systems, and high-volume advanced manufacturing highly transferable. Fusion organizations increasingly need systems engineers who can manage interfaces and requirements across mechanical, electrical, software, thermal, vacuum, cryogenic, and safety systems—not just plasma physicists. Some of those critical skills include the following:

The most valuable cross-industry capability is systems engineering: defining what each subsystem must do, managing interfaces, testing them against requirements, and designing the whole plant for reliability and maintainability. For example, a blanket, heat-extraction loop, tritium-processing system, magnets, vacuum vessel, pulsed-power equipment, diagnostics, controls, and turbine island may each work independently; systems engineering determines whether they can operate together safely and deliver net electricity at high availability. Fusion’s plant-engineering challenge explicitly extends “beyond the fusion engine” to integrating the complete facility. And those are exactly the skills that Sam Crisafulli brings to Inertia (as an example of skills crossover).

What Does That Mean For Fusion Energy Hiring Strategies?

As commercial fusion companies move from scientific and engineering demonstrations towards commercial deployment, their workforces should look less like physics-only research programs, and more like the complex industrial enterprises they aspire to be. That means growing demand for electrical, mechanical, materials, manufacturing, and controls engineers, as well as machinists, welders, electricians, supply-chain professionals, technicians, operators, and project managers. Even “soft” skills such as product management, product marketing, government relations, and regulatory affairs have important places in a company that plans to design and build commercial fusion energy power plants. And while partnering with engineering, procurement, and construction (EPC) companies such as Bechtel, Fluor, Kiewit, and Black and Veatch (among others) is a smart strategy to reduce the massive skillsets (and capital) required to build power plants, a successful commercial fusion energy company still should end up looking a lot more like an original equipment manufacturer (OEM) such as GE Vernova, Siemens Energy, or Mitsubishi Power, and a lot less like a Department of Energy (DoE) national laboratory. In many ways that leap is at least as hard as the leap from conquering fusion in a laboratory to making fusion repeatable, because it means that the DNA of the company must change.

The Fusion Report has had its numbers of disagreements with Elon Musk, many of which we have covered in our articles. From putting data centers into space to (what we see as) missteps with Tesla, not to mention the joke known as the “Department of Government Efficiency”, we have disagreed with Elon more than we agreed with him. Also, let’s remember that Elon’s company Tesla is in the solar and battery energy storage system (BESS) business, with revenue in Q2 2026 of $3.1B, a capex spend of more than $25B/year on solar cells/panels and Megapack (battery wall) fabrication facilities, and a debt bet of up to $30B. However, we do agree strongly with Elon on the potential for solar energy/BESS technology, especially the relative ease of deploying solar and BESS, assuming that geography is well suited to solar, such as the US southwest (California, Arizona, Nevada), followed by Texas and the US southeast (Florida, North Carolina, Virginia, Georgia). However, we do disagree with Elon’s view that solar energy can replace the use case for fusion energy, so let’s jump right into it.

Elon Musk’s (And TFR’s) Views On Commercial Fusion Energy

In Elon Musk’s view, he doubts that commercial fusion energy will be economically competitive, though he does not necessarily oppose the concept of fusion energy research. Elon has stated that commercial fusion reactors will be extremely expensive and difficult to build, fuel, maintain, and operate, and that their electricity would likely be more expensive than solar power paired with BESS technology. Musk’s preferred alternative is to harness the “enormous, free fusion reactor in the sky” (the Sun) through solar generation, storage, and transmission infrastructure. In 2021, he also said that fusion energy was “not a necessity,” as Musk postulated that very safe nuclear fission reactors (which are neither cheap nor truly clean) could provide reliable and clean energy. More recently, he has criticized efforts to build small terrestrial fusion reactors as wasteful unless they are explicitly treated as scientific experiments, suggesting that he views fusion as scientifically impressive but unlikely to become the lowest-cost commercial energy source (which isn’t necessarily the true test for fusion energy).

In his analysis, Elon is right about a few things (everything else equal):

• From a build-time standpoint, nothing beats solar energy plus BESS. One gigawatt of solar/BESS takes 2-4 years to permit, develop, and build. By comparison, a 1GW natural gas plant takes 3-6 years to permit, develop, and build, while a 1GW nuclear power plant takes 10-15 years.

• Solar (and other renewable energy sources) is “greener” than any other power source on the planet, including fusion energy or nuclear fission. With that said, almost nothing is “truly green”, as even hydroelectric power requires significant amounts of concrete to build the dams, water sluices, and other infrastructure.

• Commercial fusion energy has a significant number of technical and schedule risks/unknowns to be solved, such as materials durability, tritium breeding and fuel supply, component replacement, construction cost, and regulatory requirements (these are expected to be resolved within the next 10 years).

Now for the part of the discussion where we diverge with Elon:

• The biggest challenges with solar energy includes that only some of the US is good for solar energy generation, the amount of land solar energy generation requires, and it needs high-voltage transmission lines from the generation site to the urban areas where the customers are located. A number of sources estimate that the amount of land required per GW of solar in areas highly favorable to solar tends to be between 5,000 and 10,000 acres (8 to 16 square miles), with the amount of land required being inversely proportional to the favorability of that land for solar energy production (less favorable equals more land).
  

• Overshadowing the actual land cost (which is typically leased at a cost of $500-$2,000/acre/year) is the actual cost for transmission lines if there are not ones that are close to the project and/or do not have extra capacity. With a typical site distance of 20 to 50 miles from an urban center, the cost of the transmission lines can reach close to a billion dollars, assuming that there are even utility corridors to run the lines.

While these costs do not make a solar energy solution “dead on arrival”, they can significantly extend the permitting times compared to a “brownfield” (existing site) conversion of a non-renewable power plant to energy sources such as fusion energy, nuclear fission or natural gas, making the solar solution less advantaged. And it doesn’t make areas with low solar production more desirable, an issue that Elon Musk did not address, which is why the EIA’s Annual Energy Outlook forecast puts solar at roughly 20% of US generation by 2050 – an important amount, but only part of the total electricity that the US needs.

The Use Case for Fusion Energy: Greenfield Vs. Brownfield

The US Department of Energy (DoE) describes fusion as having the potential to deliver abundant, reliable, zero-carbon primary energy, while ITER identifies grid and industrial baseload power as a central application. The “official“ primary use case for commercial fusion energy is to provide large amounts of reliable, zero-carbon electricity alongside variable sources such as wind and solar power. In the “official” use case, fusion energy can therefore serve as a baseload power source for regions that need dependable generation but have limited access to hydroelectric resources or suitable sites for conventional nuclear plants.

The larger concern limiting fusion energy’s use case in brownfield applications, is its commercial risk and schedule. Given the number of old electrical plants being retired, a brownfield utility site such as a coal plant or old nuclear fission facility would have valuable grid access and local workforce advantages. It is also likely that brownfield applications would roughly double the number of applications or opportunities for commercial fusion energy power plants. However for many utility companies, fusion remains an unproven emerging technology with a number of unresolved questions, making it critical that early applications of fusion energy deliver the answers required: costs, material survivability and overall schedule risk mitigation.

It Isn’t Solar vs Fusion; It’s How Quick Fusion Can Get To The Finish

In the US markets, commercial fusion energy’s greatest enemies are essentially geography and time for a number of important reasons:

• Baseline electricity in the United States is very cheap, with an average price of $0.18/kWh, vs $0.38-$0.43 per kWh in Germany. Even within the United States, prices vary significantly by state, with places like California and Hawaii having prices similar to those of Western Europe.

• Over the next 10 years, the cost of solar cells and battery energy storage system (BESS) technologies will drop significantly. At the same time, the efficiency of solar cells has increased from roughly 15%-17% a decade ago to 20%-22% today. These cost and technology advances make solar energy even more cost-advantaged vs commercial fusion energy.

• Massive overhauls are occurring in US electrical transmission systems over the next decade, increasing both capacity and reliability. This eliminates or significantly reduces one of the major hurdles to renewables in general, and solar energy in particular.

• Next-generation nuclear fission reactors like Small Modular Reactors (SMRs) are currently receiving far more near-term commercial interest and real-world utility partnerships than commercial nuclear fusion is, primarily due to nuclear fission’s technical maturity advantages, known regulatory pathways, and the Trump administration’s support for fission energy over the potential of fusion energy.

While these factors do not mean a “death knell” for commercial fusion energy, they likely mean that its ability to be taken seriously as a true mass-market energy resource in the US over the next decade is going to be challenging until the technological and economic hurdles for commercial fusion energy have been successfully overcome; effectively Elon Musk has not been proven wrong yet 😊. Let’s hope at least a few companies within the fusion industry have a chance to do so.

Things You Gotta Know

Inertia Enterprises Says Ignition-Benchmarked Simulations Show Its Target Design Delivers More Than 25x Gain and 250 MWe to 1.3 GWe of Grid Electricity
Inertia Enterprises announced that a high-fidelity virtual shot of its commercial Laser Indirect Drive target design, run with Lawrence Livermore National Laboratory on the same design codes the lab used to achieve ignition at the National Ignition Facility, shows the design can produce enough fusion energy to generate between 250 MWe and 1.3 GWe of grid electricity. Driven with 10 MJ of laser energy, the baseline design yields about 260 MJ per shot for a target gain of roughly 26, compared with the roughly 8 MJ and 4x gain NIF has demonstrated, and early optimizations push modeled output above 400 MJ for gains up to 43x. At the planned 10 shots per second, the baseline corresponds to about 290 MWe of net electricity with a 12 percent efficient laser or 450 MWe at 15 percent, while the 43x case reaches roughly 1.05 GWe to 1.3 GWe. The design keeps NIF’s proven physics but scales up the target and fuel mass, swaps the gold-lined depleted uranium hohlraum for lead, assumes rougher fast-formed deuterium-tritium ice, and still ignites with full yield after accounting for injection and pointing errors and ablator voids. Co-founder and Chief Scientist Annie Kritcher called it a significant technical achievement; the result completes the 25x gain task on Inertia’s Top Ten Things list for phase one, and the underlying paper is on arXiv and being prepared for peer-reviewed submission.

Commonwealth Fusion Systems Runs SPARC’s Full Cryogenic System at Its 8 Kelvin Operating Temperature
Commonwealth Fusion Systems announced that the cryogenic system for its SPARC tokamak in Devens, Massachusetts has run at its full operating temperature of 8 kelvin, colder than the surface of Pluto. The plant circulates supercritical helium to chill SPARC’s toroidal field magnets to 8 K and its poloidal field and central solenoid magnets to 15 K, with 80 K coolant for supporting non-superconducting magnets. To commission the system before the tokamak is finished, CFS built a bypass pipe it calls a turnaround tool for practicing cooldowns, plus a blowdown system that stores 8 K helium in tanks to flush through SPARC during its most powerful 10 second pulses, followed by a roughly four hour recool. The run is a step toward an integrated dry dress rehearsal that links cryogenics with magnet power supplies and radio frequency heating without the tokamak running. CFS says the plant is essentially the equipment its ARC power plant will need, and the milestone follows the company’s $1 billion raise this summer that brought total capital raised to roughly $4 billion.

UKAEA and PPPL Sign Joint Declaration to Link Their SUNRISE and STELLAR-AI Fusion Supercomputers Across the Atlantic
The United Kingdom Atomic Energy Authority and the US Department of Energy’s Princeton Plasma Physics Laboratory signed a Joint Declaration of Intent at the Global Fusion Policy Summit in London to explore linking UKAEA’s SUNRISE supercomputer with PPPL’s STELLAR-AI platform as the SUNRISE-STELLAR-AI Federation. SUNRISE is the UK’s first AI supercomputer dedicated to fusion, backed by £45 million from the UK government and anchoring the UK’s first AI Growth Zone at UKAEA’s Culham Campus; STELLAR-AI is PPPL’s AI and high performance computing platform, operated with support from Princeton University. The federation would train the same AI models on data from UKAEA’s MAST Upgrade in Oxfordshire and PPPL’s NSTX-U in New Jersey, two compact spherical tokamaks of similar design, because models trained on a single machine struggle to predict another. Next steps under discussion include digital twins of both facilities and moving computing jobs between the platforms, with the aim of accelerating designs such as the UK’s STEP plant and PPPL’s Spherical Tokamak Advanced Reactor. The declaration builds on a memorandum of understanding the two laboratories signed in June.

nT-Tao Deploys a 50 MW Pulsed-Power Subsystem to Push Its Compact Fusion Prototype Past 1.5 Million Degrees
Israeli compact fusion company nT-Tao announced that it has developed and deployed a proprietary 50 MW pulsed-power subsystem with real-time nonlinear control, expanding its MEGA (Modular Energy Generator Architecture) platform. The subsystem combines 50 compact Power Electronic Generators of roughly 1 MW each, arranged as ten parallel branches of five modules and designed to operate at 50 to 2,000 kHz, a roughly five-fold scale-up from the earlier 12 module, approximately 10 MW generation. It builds on nT-Tao’s recently proven real-time nonlinear controller, which maintains resonance and maximizes power transfer as the electrical characteristics of the system and the plasma change during each pulse. Co-founder and CTO Boaz Weinfeld said the hard part is not generating 50 MW but controlling 50 modules in real time so they act as a single tuned system, and that building from proven 1 MW modules keeps the footprint compact and makes scale-up a matter of repeating a unit. nT-Tao says the subsystem will drive its C3 prototype series past the 1.5 million degrees Celsius already demonstrated, toward a 20 MW on-site fusion system for distributed baseload power.

We are nearing the start of fiscal year 2027 (FY-2027), which begins on October 1^st, which is important for two reasons: 1) we’re a stone’s throw away from 2028, which was when Helion Energy was supposed to achieve full operation of their 50MWe fusion power plant for Microsoft; and 2) the FY-2027 Department of Energy (DoE) budget kicks off, with the question of how much money is being earmarked for commercial fusion energy development. These are the key questions that The Fusion Report will explore in today’s article.

Helion (and Microsoft) Unsurprisingly Make A Slide To The Right…

Earlier this summer at the 2026 FusionX Americas Conference in Boston, Helion Energy quietly moved the full operation date for their Malaga, Washington Orion 50MWe commercial fusion energy power plant, which will supply a Microsoft hyperscale data center with electricity. Delivering the (unsurprising) bad news was Director of Public Affairs Jackie Siebens, who spoke for the company at the conference where she stated that, “We will not, in no universe, have 50 megawatts hit full operation in 2028.” Rather, Jackie stated that the initial operating date for the Orion plant will begin by the end of 2028, but that Helion does not expect to start full power generation at the plant until 2029 or 2030. For most people in the industry, or those who are following it, the news is not particularly a surprise – Helion’s milestones have always been particularly aggressive, in an industry it has not necessarily been known for conservativism.

Though full operation by 2029 or 2030 would still be an industry-leading accomplishment, it comes at a bad time for Helion Energy who just a week before the FusionX conference closed a $465M funding round, with a total lifetime funding raise of slightly over $1.5B. While this is the second most funding that a company has raised in the commercial fusion energy industry, it is significantly behind the nearly $4B raised by number one commercial fusion energy company Commonwealth Fusion Systems to date. To complete their Orion plant, it is likely that Helion will have to raise significantly more capital (perhaps on the order of $2B to $3B) or raise equivalent venture debt. In any case, that fundraising will be harder after this announcement that would have been without it.

DoE: More Money Than Last Year, But Only NNSA Gets An Increase

The FY2027 DoE budget request is higher in aggregate than FY2026, rising from $49.1 billion enacted in FY2026 to $53.9 billion in FY2027—an increase of about $4.8 billion, or nearly 10%. However, this is not a broad-based expansion of DoE’s civilian energy mission: the increase is driven overwhelmingly by a proposed $7.4 billion (29%) increase for the National Nuclear Security Administration (NNSA), particularly nuclear-weapons modernization and production activities. In contrast, non-NNSA DOE programs would decline by roughly $2.6 billion (11%), with proposed reductions affecting the Office of Science, ARPA-E, nuclear-energy programs, grid and energy-security activities, and many clean-energy, efficiency, transportation, and demonstration programs.

The request also proposes new or reorganized accounts—such as Baseload Power, Artificial Intelligence and Quantum, Critical Minerals and Energy Innovation, Hydrocarbons and Geothermal Energy, and Fusion—while relying in part on transfers or rescissions of previously appropriated Infrastructure Investment and Jobs Act (IIJA) funds. In essence, the FY2027 increase reflects a major reprioritization toward nuclear security, weapons infrastructure, baseload generation, critical minerals, and AI computing capacity rather than an increase in civilian energy innovation spending.

The FY-2027 DoE Budget Impact on Fusion Energy Programs

For fusion energy, the DoE’s FY2027 budget request for Fusion Energy Sciences (FES) is $755.25 million, down $50.41 million, or about 6.3%, from the FY2026 enacted level of $805.66 million. The headline reduction largely reflects a proposed $93 million cut to the U.S. contribution to the International Thermonuclear Experimental Reactor (ITER), while funding for the rest of the FES portfolio would increase by roughly $43 million. The FY2027 request includes $135 million for public-private partnership activities, including the Fusion Milestone Program, INFUSE, private-facility research, and the Fusion Bridge program, indicating continued support for commercialization-oriented work despite the overall decline. The FY-2027 DoE budget also proposes $10 million for a new DoE Office of Fusion to coordinate department-wide fusion activities. In short, FY2027 would shift fusion funding away from the international ITER commitment and toward domestic and private-sector-facing efforts, but total FES funding would remain below FY2026 appropriations and well below the $1.114 billion authorized for FY2027 under the CHIPS and Science Act.

How Does the FY-2027 DOE Budget Compare to Other Countries?

In comparison, the United Kingdom has committed more than £2.5B (roughly $3.4B at recent exchange rates) over five years for fusion research and commercialization, implying an average annual commitment of more than £500M, or approximately $650M–$700M. On this annualized basis, the US DoE FY-2027 FES request is modestly larger, but the two countries are operating at broadly comparable scales of public support. The comparison is not exact: the DoE figure is a single-year congressional budget request focused on FES, while the UK figure is a multi-year package that includes the UK Atomic Energy Authority’s (UK-AEA) research activities, the STEP pilot-plant program, facilities, workforce, and commercialization measures.

The more important distinction is strategic direction.

The US FY-2027 request preserves and modestly expands domestic (non-ITER) fusion work, including $135M for public-private efforts such as the Fusion Milestone Program, INFUSE, private-facility research, and Fusion Bridge. However, the UK program is increasing support under a defined multi-year strategy that aims to turn the country’s research base into a domestic fusion industry and deliver the STEP spherical-tokamak prototype plant. Thus, the United States retains a formidable fusion science and private-company ecosystem, but its proposed FY-2027 funding provides less multi-year certainty than the UK’s commitment. The UK’s approach is more visibly oriented around a nationally sponsored pilot plant and commercialization pathway, whereas the U.S. approach relies more heavily on a distributed network of national laboratories, universities, private firms, and competitive R&D partnerships.

Similarly, Germany has pledged more than €2B through 2029 (roughly €400M, about $430M-$470M per year if distributed evenly) to expand fusion research, infrastructure, pilot projects, and commercial readiness. Germany’s program is therefore smaller in yearly dollar terms but is moving sharply upward from previous annual public fusion-research funding of about €150M and is explicitly organized around a national Fusion Action Plan and the objective of building a German fusion power plant. Japan is more difficult to compare because its support is divided among ministries, national laboratories, ITER obligations, and broader innovation programs rather than reported as a single DOE-style annual fusion appropriation. Nevertheless, Japan remains a major fusion power through its leadership in magnetic-confinement research, its collaboration with Europe on the JT-60SA tokamak, and its updated national Fusion Energy Innovation Strategy, which emphasizes commercialization, public-private coordination, and greater investment. Reported Japanese commitments of more than ¥100B (approximately $660M) support fusion research and startup activity across future programs, placing Japan in a broadly comparable funding range, though not on a directly equivalent annual basis.

Conclusion: Is the US DoE Moving Forward or Backward on Fusion?

Ultimately, the U.S. fusion energy budget represents a complicated “one step forward, two steps back” paradox when measured against broader energy commitments. On one hand, targeted strategic shifts—such as the creation of a standalone Office of Fusion and record commercialization injections like ARPA-E’s $135 million initiative—mark a significant forward leap toward treating commercial fusion as a near-term national priority rather than a distant science project. On the other hand, the broader financial reality tells a story of retreat; the overall federal request for the Fusion Energy Sciences (FES) program has dropped beneath previous years’ appropriations and remains significantly short of the funding goals authorized by the CHIPS and Science Act. Compared to trillions allocated globally toward mature renewable technologies, or the aggressive multi-billion dollar state-backed budgets of international competitors like China, the current sub-billion dollar U.S. commitment risks falling behind. While the administration’s focus on slashing regulatory friction and boosting public-private partnerships provides a vital foundational blueprint, without a dramatic scale-up in base capital investment, the current budget moves the US forward conceptually but leaves it poorly equipped to lead the global commercial fusion race.

This is the third article in The Fusion Report series on critical supporting technologies for commercial fusion energy. In today’s article, we’ll look at a class of technologies called divertor and heat exhaust systems. But first, let’s look at which of these systems are required for what types of fusion machines.

Magnetic Fusion Energy (MFE) Machines – Tokamaks, Stellarators

The classical definition of divertors are deliberately sacrificial exhaust regions of magnetic fusion energy (MFE) machines: magnetic field lines guide helium “ash,” fuel ions, and heat out of the plasma core onto specially cooled targets. In massive tokamaks, such as the International Thermonuclear Experimental Reactor (ITER), the usual arrangement is an axisymmetric single-null or double-null X-point divertor, typically positioned at the bottom (and, for double-null operation, also the top) of the torus. The X-point separates confined field lines from open lines that intersect divertor targets, allowing particles and heat to be removed while helping shield the main chamber wall. For example, ITER’s divertor uses tungsten armor because of its exceptionally high melting point, and is designed to extract fusion heat and ash while limiting plasma contamination.

In commercial reactor-scale tokamaks, the central problem is that a compact conventional divertor can concentrate enormous heat fluxes onto a small wetted area. That has motivated “advanced” tokamak layouts: snowflake divertors use a near-second-order magnetic null to broaden and split the exhaust pattern, while long-leg concepts such as Super-X route exhaust along a longer path to a larger-radius target, increasing flux expansion and giving the plasma more opportunity to radiate and cool before reaching material surfaces. These approaches have been tested or developed in devices including the Swiss TCV, the Princeton Plasma Physics Laboratory (PPPL) National Spherical Torus Experiment (NSTX), the General Atomics DIII-D facility, China’s Experimental Advanced Superconducting Tokamak (EAST), and the UK Atomic Energy Authority (UK-AEA) MAST Upgrade.

Stellarators require a substantially different divertor philosophy because their externally generated, three-dimensional helical magnetic fields lack the tokamak’s simple rotational symmetry and generally do not rely on a large plasma current. Rather than using a circular, axisymmetric X-point and ring-like targets, stellarators commonly employ island or helical divertors that exploit naturally occurring edge magnetic islands and direct open field lines to multiple shaped target modules around the vessel. Wendelstein 7-X is the best-known example: its modular divertor system must conform to a strongly three-dimensional field-line topology, making the geometry, target shaping, pumping placement, and heat-load distribution more complex than in a tokamak. The payoff is potentially steady-state operation and the ability to tailor the edge magnetic geometry through coil and boundary design; the challenge is ensuring that the island structure distributes heat broadly enough and does not create localized hot spots. Other fusion concepts may depart further: spherical tokamaks face severe space constraints and often pursue compact, high-flux-expansion or long-leg divertors, while some low-temperature, open-field or inertial fusion concepts do not use a tokamak- or stellarator-style divertor at all. Thus, the key distinction is not merely material choice, but the machine’s magnetic topology: relatively symmetric X-point exhaust in tokamaks versus intrinsically 3-D island/helical exhaust in stellarators.

Magnetic Mirrors

Magnetic mirrors are much closer to having a divertor analogue, because their magnetic field lines are inherently open. A mirror confines many charged particles by increasing the magnetic-field strength at each end, reflecting particles whose pitch angle is sufficiently large. Particles inside the loss cone instead stream out along the machine axis through the ends—known as end loss. In a simple mirror, this is a primary confinement limitation, not a small peripheral exhaust like in a tokamak. Consequently, mirror systems require engineered end regions to receive this escaping particle and power flow. Depending on the design, those regions may include:

• An end tank or expander, where field lines expand and the heat/particle flux can spread over a larger area.

• A cooled end plate, collector, or plasma-facing target to tolerate residual particle and thermal loads.

• Neutral pumping and fueling hardware to control the end-region plasma and remove particles.

• A direct-energy converter (DEC), which uses electrostatic fields or grids to decelerate escaping charged particles and recover part of their kinetic energy directly as electricity rather than merely turning it into heat.

That last option is distinctive: in a magnetic mirror, a stream of charged particles exiting through an end loss cone can be a usable output channel. Direct conversion has long been investigated for mirror end losses, and current mirror developers are pursuing it as a way to improve plant efficiency. It cannot capture neutron energy, which still has to be recovered in a surrounding blanket, but it may recover some energy carried by escaping ions and electrons.

Inertial Fusion Energy (IFE) Machines

For IFE machines, a tiny target is compressed and ignited in the center of a chamber by lasers, ion beams, or another driver. The burn lasts only briefly and produces an outward, nearly radial pulse of X-rays, charged target debris, and 14.1-MeV neutrons; there is no magnetically confined edge plasma that must be led along open field lines to a divertor. Accordingly, IFE plants need a fusion chamber and first-wall protection system, but not a divertor in a classical sense. The engineering task is to survive repetitive micro-explosions (potentially several per second), while preserving clear beam paths, handling debris and tritium, recovering heat, and limiting damage to structural materials. There can still be specialized debris collectors, gas-handling equipment, vacuum pumping, liquid cleanup, and heat-extraction systems in an IFE plant, but calling those a “divertor” would be misleading.

Conclusion: Divertors and Heat Exhaust Systems Are MFE Issues

Different fusion energy systems provide different challenges, and divertors and heat exhaust systems are primarily an issue for magnetic fusion energy (MFE) approaches, as well as to a lesser extent for magnetic mirrors. The primary reason that the divertor is an issue for MFE systems is that they have to be able to exhaust burned products at the same time they are still continuously burning new hydrogen into helium, while not breaking magnetic field confinement. IFE systems can simply pump out the helium gas and any other byproducts, or capture them in a liquid first wall blanket. Magnetic mirrors provide different challenges than MFE or IFE machines, with the potentially added benefit of direct energy conversion (DEC), especially for aneutronic fusion fuels where the exhaust is primarily charged particles.

Things You Gotta Know

Princeton and PPPL Researchers Debut PACMAN, an AI Framework That Controls Fusion Plasma in Milliseconds
Researchers at the US Department of Energy’s Princeton Plasma Physics Laboratory and Princeton University tested a new AI framework called PACMAN across five experiments on the DIII-D National Fusion Facility tokamak in San Diego. The system links multiple machine learning models that read real-time plasma measurements and issue control commands in about 20 milliseconds. That is far faster than a human operator, who typically responds on the order of seconds. Across the five experiments, PACMAN let a reinforcement learning model take full control of the tokamak’s heating systems. The framework also predicted sudden bursts of energy from the plasma’s edge and detected and controlled waves driven by fast particles. It further adjusted plasma density and rotation to research targets, and it predicted a damaging instability called a tearing mode roughly 200 milliseconds before it would have occurred, giving operators time to adjust the plasma and avoid it. Each model and controller works independently, which the researchers say can be added without disturbing the rest of the system. The modular framework could extend to other tokamaks.

RTX Partners With and Invests in Xcimer Energy to Develop Advanced Directed Energy Technology
RTX Ventures, the venture capital arm of RTX, made a minority investment in laser fusion company Xcimer Energy to support development of scalable multi-kilojoule-class pulsed-laser technologies. RTX and Xcimer also signed a teaming agreement to explore integrating and containerizing Xcimer’s pulsed-laser technology into RTX’s defense systems, aimed at giving the US military new directed energy capabilities to defeat airborne threats. The companies say the core engineering challenge is shared across laser fusion and laser defense: building a laser that fires repeatedly and reliably at a low cost per joule while remaining mass-manufacturable, maintainable, and energy efficient. RTX’s investment builds on Xcimer’s laser fusion roadmap, which the US Department of Energy approved in June under its Milestone-Based Fusion Development Program.

Fusion Industry Association Gives Evidence to UK Parliament on National Policy Statement for Fusion Energy Generation
Fusion Industry Association UK Director Tris Denton gave evidence to Parliament’s Energy Security and Net Zero Select Committee as part of its inquiry into the National Policy Statement for fusion energy generation, known as NPS EN-8, which is intended to streamline the planning process and provide clarity for fusion power plants in the UK. The committee session covered power plant planning, regulation, commercial timelines, and supply chains as the UK moves toward deployment. The hearing leads into UK Fusion Week, where the FIA will join the Global Fusion Policy Summit hosted by the US Department of Energy and the UK Department for Energy Security and Net Zero on September 14 to discuss international collaboration on fusion commercialization.