Commerce Is Paying Fusion Companies to Make Their Own Chips, and Taking Equity in Return. What the $389 Million CHIPS Round Means for Applicants

When Congress passed the CHIPS and Science Act in 2022, the research money was meant for semiconductor research: national advanced packaging programs, manufacturing institutes, and a national semiconductor technology center. Four years later, the largest award in Commerce's latest CHIPS R&D round goes to a company building a fusion power plant.

On October 22, 2026, the Department of Commerce announced four letters of intent for up to $389 million in CHIPS R&D awards, all aimed at what Commerce calls America's "energy semiconductor supply chain":

• Helion: up to $175 million "to develop, scale, and establish domestic capabilities in high-voltage power semiconductors and advanced energy power modules designed for high-peak-current pulsed-power applications in fusion power generation."
• Commonwealth Fusion Systems (CFS): up to $85 million for semiconductors used in "radio-frequency plasma heating, power electronics, embedded electronics for fusion diagnostics, and radiation-hardened electronics."
• Vertical Semiconductor: up to $73 million for vertical gallium nitride (GaN) power devices, including a demonstration of "the world's first 2.5 MW super intelligence data center rack."
• Sibelco: up to $56 million to develop purification and finishing technology for ultra-high-purity quartz used in crystal-growth crucibles and other semiconductor hardware.

Each award comes with the same condition: "The Department will receive a minority, non-controlling and non-voting equity stake in each company as a condition for making the awards, enhancing the return for the U.S. taxpayer."

These are letters of intent, not final awards. Even so, the round shows plainly how CHIPS R&D money is now awarded, and companies considering an application should look closely at it.

The Deals in Detail

Helion's own announcement, and its CEO's interview with GeekWire, add detail the Commerce release doesn't include. The award would fund a three-year program. More than $100 million would support R&D and manufacturing at Helion's Everett, Washington, site, focused on high-voltage semiconductors and advanced packaging. About $80 million would fund next-generation device development with GlobalFoundries. Helion is adding $50 million of private money, for a $225 million total program. Once a definitive agreement is signed, Commerce would receive a $175 million non-voting interest in Helion, which Helion says is priced at its Series G valuation. In other words, the equity stake equals the award amount.

The technical case is specific. Helion's pulsed magneto-inertial design uses large electrical pulses to compress plasma and recovers electricity directly from the magnetic field rather than through a steam turbine. CEO David Kirtley said the semiconductors needed to "power, control, and harvest energy" from those magnets currently come mainly from overseas suppliers, and that no U.S. supplier meets the required performance. Helion expects to create more than 100 jobs in Everett. It's still targeting 2028 for initial operation of Orion, its commercial plant in Malaga, Washington, which has a power-purchase agreement with Microsoft. Some chips from the CHIPS program may end up in Orion late in construction.

CFS's $85 million covers a broader set of electronics: RF power generators for plasma heating, fast embedded controllers for plasma management, radiation-hardened electronics, and semiconductor technology for robotic maintenance systems. CEO Bob Mumgaard described the opportunity as developing "semiconductors and power electronics unique to the fusion industry."

The Pattern: Themed Rounds With Equity Attached

Commerce's press release was filed under the name "Energy Wave," and the word fits. CHIPS R&D money has been going out in themed rounds, and each one has come with federal equity:

• December 2025: a letter of intent with xLight for $150 million toward free-electron lasers for EUV lithography, with $150 million of equity. The award was finalized in June 2026.
• May 21, 2026: letters of intent with nine quantum companies, including an award of up to $1 billion tied to an IBM-linked quantum foundry, each with an equity condition. Several were finalized in early September. According to a public tracker of federal equity stakes, the D-Wave deal took government shares at a 15% discount to market.
• October 22, 2026: the four-company energy semiconductor round.

Over this period, the instrument behind these awards changed. The CHIPS R&D Office's Broad Agency Announcement, 2025-NIST-CHIPS-CRDO-01, was amended in April 2026 into a two-track program, with an R&D Project path and an Investment Fund path. It tells applicants on either path that they "may be required to issue — and should anticipate issuing — equity, warrants, licenses to intellectual property, royalties or revenue sharing." The minimum request is $10 million, and the BAA accepts submissions on a rolling basis through September 30, 2029. NIST paused submissions on September 15, 2026, for system upgrades and said it expected to reopen the portal in November. Applicants should check the current status at nist.gov/chips before planning a submission.

The October round confirms what the BAA language suggested. Equity isn't a negotiating fallback. It's a standard condition, and in Helion's case it was set at the full award amount.

Who This Round Signals Opportunity For

The most useful point for prospective applicants is how far the definition of "semiconductor R&D" now stretches. None of the four companies is a conventional chipmaker. Two are fusion developers, one is a materials supplier, and one is a device startup whose stated use is data-center power. What they share is the theme of the round: the semiconductors and materials that control, convert, and deliver power. Commerce described power electronics as "foundational to modernizing the domestic grid and powering critical energy infrastructure loads."

That suggests openings for several kinds of companies:

End users with a supply-chain gap. Helion's argument wasn't "we make chips." It was "we need a chip nobody in the U.S. makes, and building it ourselves strengthens the domestic supply chain." Companies in grid hardware, electric aviation, defense power systems, or industrial drives that depend on overseas power devices can make the same argument, especially with a foundry partner. GlobalFoundries' role in the Helion program is a model worth studying.

Wide-bandgap device developers. Vertical Semiconductor's $73 million for vertical GaN fits the round's stated focus on "higher power efficiency, enhanced grid stability, and reduced thermal losses." Silicon carbide and gallium oxide developers working at grid or data-center scale fit too.

Upstream materials suppliers. Sibelco's award for high-purity quartz is a reminder that CHIPS R&D covers inputs to the semiconductor supply chain, not only the chips. Crucibles, specialty gases, substrates, and process chemicals with foreign concentration risk all qualify.

The data-center language also deserves notice. Vertical Semiconductor's demonstration target is a "super intelligence data center rack." That term follows Executive Order 14434, which has federal agencies using "super intelligence" where they used to say "artificial intelligence." Proposals that tie power electronics to AI compute buildout match current administration priorities.

What to Weigh Before Applying

Equity changes who has to approve the application. A CHIPS R&D application now needs approval from your board, your lead investors, and probably your cap table counsel, before a white paper goes in. Helion's stake was priced at its most recent preferred round. Companies between rounds should expect negotiations over valuation. The federal stakes are non-voting and minority, but they're permanent, and future investors will see them in diligence.

A letter of intent is the start of negotiations. Helion says Commerce receives its interest "upon entering into a definitive award agreement," and that agreement hasn't been signed. In the May quantum round, some letters took about three and a half months to reach final awards, and the public record of final terms is uneven. In several cases, the finalization announcement didn't repeat the equity language. Budget for months of legal work between selection and funding.

The minimum request rules out most small firms as lead applicants. At $10 million, the BAA isn't built for seed-stage companies applying alone. Smaller device or materials companies may do better as named subcontractors or development partners on a larger company's proposal, the same way GlobalFoundries is on Helion's. They can also use SBIR and STTR programs at DOE, DoD, and NSF to build the track record a later CHIPS application would need.

Watch the politics around the equity model. Federal equity in grant recipients has become a hallmark of this administration's industrial policy, from Intel's 9.9% stake to minerals and quantum deals. The incoming Democratic Congress may scrutinize the approach in oversight. That's unlikely to reverse signed deals, but the terms of future rounds could change.

The Bottom Line

The October 22 round makes three things clear. CHIPS R&D money now flows to companies whose products use semiconductors, not only companies that make them. Commerce expects equity in return, sometimes dollar-for-dollar. And awards arrive in themed rounds that reward applicants whose proposals match the current priority. For a company with a real domestic power-electronics gap, a foundry partner, and investors willing to accept a federal shareholder, this is one of the largest pools of federal R&D money available, though unlike a grant it comes with dilution. Make sure your board accepts those terms before you prepare the white paper.