Washington Took Its Quantum Stakes Up Front and Tied the Cash to Milestones

September 8, 2026 – The U.S. Department of Commerce announced final CHIPS and Science Act awards of up to $100 million each to four quantum computing companies, for research and development on manufacturing and scaling their hardware. The recipients are D-Wave Quantum, Rigetti Computing, Quantinuum and PsiQuantum. They build different kinds of machines: superconducting qubits in annealing and gate-model designs, trapped ions, and photonics.

Commerce takes a minority, non-controlling equity stake in each company as a condition of its award. It set that term in the letters of intent it signed in May and said the stakes were meant “to enhance the return for the U.S. taxpayer.”

The awards take the form of Other Transaction Agreements (OTAs), which are not governed by the Federal Acquisition Regulation. The three publicly traded recipients signed theirs on September 4 and disclosed the terms in securities filings on September 8; under each, Commerce releases money in tranches as the company meets milestones. PsiQuantum, which is private, said it had signed definitive documentation but disclosed no signing date or equity terms.

On the same day, Commerce announced a final award of up to $375 million to GlobalFoundries for R&D at its Quantum Technology Solutions business in Malta, New York. The work covers cryogenic CMOS process design kits, advanced packaging and heterogeneous integration for several quantum architectures, and GlobalFoundries can draw the money over five years as it meets milestones. Under a securities issuance agreement dated September 3, GlobalFoundries issued Commerce 9,907,399 ordinary shares at $37.85 each, about $375 million at that price.

The five awards finalize five of the nine letters of intent, worth $2.013 billion in total, that Commerce signed on May 21. On September 16 it finalized a sixth: up to $1 billion for Anderon, a new IBM subsidiary that runs a 300mm quantum wafer foundry in Albany, New York. As of September 17, Atom Computing and Infleqtion, which build neutral-atom machines, and Diraq, a silicon-spin developer with a proposed award of up to $38 million, had not finalized theirs.

Tranches and share terms in the three 8-K filings

D-Wave, Rigetti and Quantinuum each filed a Form 8-K on September 8 that describes the award’s payment schedule and the shares issued or to be issued to Commerce. In each case the shares were worth the full $100 million award at the issue price.

Rigetti agreed to issue 7,739,938 shares at an implied price of $12.92. Of its award, $43.9 million is to be made available on or as soon as practicable after the award date. Tranches of $29.9 million and $26.2 million follow only if Commerce determines, “to its satisfaction in its sole discretion,” that Rigetti has met the milestones for each.

If Rigetti misses a required project activity date, Commerce may, subject to cure and forbearance provisions, demand repayment of everything it has paid as a debt. That right ends if Commerce terminates the agreement for convenience before the last required activity date. The three funded tasks are an integrated, miniaturized readout package, a new cryostat architecture to expand cryogenic capacity, and fabrication capabilities for high-connectivity chips.

D-Wave issued 7,095,721 shares at $14.093 each. Its filing sets the price at 15 percent below the lowest closing price on three dates: when Commerce first sent a draft letter of intent, when the letter was signed, and the award date. The filing lists five tranches: $53.6 million shortly after the award date, then $9.1 million, $16.7 million and $20.4 million on three milestones, and $0.3 million at completion.

The milestones cover installing tools and equipment, fabricating prototype quantum processing units (QPUs), process integration, and calibrating and benchmarking scaled QPUs. D-Wave said the award would speed development of a 100,000-qubit annealing system and a gate-model system of 10,000 physical qubits targeting 100 logical qubits and more than one million operations.

Quantinuum issued Commerce 2,369,528 Class A shares. Its formula takes the lower of two prices: the $60 IPO price less 20 percent ($48.00), and the award-date closing price less 15 percent. The September 4 close was $49.65, so the second price applied, an implied $42.20 a share.

Quantinuum had $56.0 million made available on or about the award date, and $32.0 million and $12.0 million follow as it meets milestones. The project covers low-loss integrated photonics at 422 nanometers, application-specific control chips (ASICs) built on a high-voltage process for cryogenic operation, and optical components at the wavelengths trapped-ion systems use. Quantinuum said GlobalFoundries would fabricate its next-generation ion traps and control electronics on 300mm wafers, and Monarch Quantum would develop and manufacture lasers and optical components.

PsiQuantum said the award, together with its own co-investment, would fund R&D and domestic manufacturing of optical switches, high-temperature single-photon detectors and advanced packaging. Commerce’s announcement listed electro-optic materials, high-temperature single-photon detectors and ultra-low-loss photonic packaging.

Shares of the three public companies rose in premarket trading on September 8: Rigetti by 7.4 percent to $16.33, D-Wave by 5.4 percent to $17.48 and Quantinuum by 3.9 percent to $51.56.

My Analysis

Shares for the full award at signing, cash by milestone

Commerce owns its full stake in D-Wave, Rigetti and Quantinuum from the first day, and about half of the money it committed to them arrives only if they meet milestones Commerce must approve. The $100 million in each announcement is a ceiling on payments.

Across the three public companies, $153.5 million of the $300 million is available at the start, from 44 percent of Rigetti’s award to 56 percent of Quantinuum’s. Rigetti’s filing is the only one of the three that states how Commerce decides a milestone has been met; the decision is Commerce’s alone. The pricing formulas are written differently, but each issue price works out to 85 percent of the company’s closing price on September 4, the award date: $15.20 for Rigetti, $16.58 for D-Wave and $49.65 for Quantinuum.

Commerce agreed to limits that match its selling rights to the cash it has paid out. Without the company’s consent, it may sell only the fraction of its shares that equals the fraction of the award disbursed and not returned. If Commerce terminates a D-Wave or Quantinuum agreement for convenience, or ends Rigetti’s for any reason, the company can buy back the shares matched to the unpaid balance for $1.00 in total.

Commerce therefore owns the full block of shares from the first day and pays cash as the companies deliver. Before each later tranche it reviews the company’s progress, and Rigetti, for one, must file annual technical milestone reports. Commerce also gets a contractual measure of each company’s progress, separate from the benchmarks each vendor chooses to publish.

The disclosed milestones are manufacturing steps. D-Wave’s run from installing tools to benchmarking scaled processors, and the tasks Rigetti and Quantinuum disclosed are components and fabrication processes, from readout packages and cryostats to integrated photonics and control chips. I read that as Commerce’s working test of viability for these awards: what a company can fabricate, integrate and benchmark against written criteria.

Venture investors stage their financing the same way. Here one department approves payments to four companies that compete for the same customers and capital, and owns shares in at least three of them.

The Intel and xLight precedents

The largest precedent is Intel. On August 22, 2025, the government took 433.3 million Intel shares at $20.47 each, $8.9 billion for a 9.9 percent stake, funded by $5.7 billion in CHIPS grants that had been awarded but not paid and $3.2 billion from the separate Secure Enclave program. Senator Elizabeth Warren criticized the terms in a September 2025 letter to Commerce Secretary Howard Lutnick, arguing that the deal relieved Intel of most of its obligations while committing billions of dollars to a company whose revenue was falling.

Commerce’s CHIPS Research and Development Office set equity terms with a startup months before the quantum awards. Its first award under the Trump administration was a December 2025 letter of intent with xLight, a Palo Alto company developing a free-electron laser as a light source for extreme ultraviolet (EUV) lithography: up to $150 million in incentives, with Commerce to receive $150 million of equity. xLight finalized the award on June 2, 2026.

The Intel stake converted grants already awarded, while the quantum awards included equity from the letter-of-intent stage. The Intel purchase was $8.9 billion, 89 times any single quantum stake at its issue price. Intel’s amended agreement removed certain project milestones to release its $5.7 billion early; under the quantum OTAs, Commerce withholds later money until milestones are met.

On July 29, outside quantum, Commerce signed a $300 million silicon-photonics letter of intent with GlobalFoundries under which it will also receive equity, about 1 percent of the company. Across Intel, xLight, the quantum group and that photonics award, Commerce has moved from converting one large manufacturer’s grants into shares after the fact to writing equity into its R&D awards from the start.

Four architectures at the same $100 million ceiling

Commerce’s May announcement described the seven computing companies as a portfolio aimed at the hardest unsolved engineering problems across several modalities. Six of the seven have the same $100 million ceiling, and Diraq’s is $38 million. I read the equal ceilings as a hedge: Commerce gave the same amount to neutral-atom, superconducting, trapped-ion and photonic developers and named no favorite among them.

D-Wave builds superconducting annealers, which use a different computational model from the gate-model machines that most error-correction work targets. Its award also funds its newer gate-model program, where 100 logical qubits is a roadmap target. Rigetti, which makes superconducting gate-model processors in its own California fab, will spend its award on readout packaging, cryostat capacity and chip connectivity. Quantinuum and PsiQuantum will spend theirs on making key components at volume: photonics, cryogenic control chips and optical parts for trapped ions, and switches, detectors and packaging for photonic machines.

Foundry awards to GlobalFoundries and Anderon

Anderon’s $1 billion and GlobalFoundries’ $375 million make up $1.375 billion of the $2.013 billion package, about 68 percent. I argued in May that the money goes to factories and supply chains more than to research, and the finalized awards keep the same ratio: capacity in the United States to make quantum chips for whichever architectures reach volume.

GlobalFoundries launched its Quantum Technology Solutions unit in May to sell fabrication, cryogenic CMOS, packaging and integration to quantum hardware companies. It already makes PsiQuantum’s Omega photonic chipsets on 300mm lines in Malta, and Quantinuum has chosen it for next-generation ion traps. Commerce now owns shares in GlobalFoundries and in Quantinuum, one of its customers.

Anderon, which IBM is backing with another $1 billion of its own, says it has started running its first quantum wafers and will make them for other hardware companies as well as for IBM.

IonQ’s absence and its SkyWater foundry

IonQ, which builds trapped-ion quantum computers, was not among the nine companies that signed letters of intent in May, and I have not seen an explanation from Commerce or IonQ. On September 8 IonQ raised its 2026 revenue guidance to $450–460 million, its first forecast to include SkyWater Technology, and The Motley Fool’s analysis attributes most of the rise from the $280–290 million guided in August to the foundry. The same day IonQ launched Superion 256, its sixth-generation platform, whose first chips were fabricated at SkyWater and which it expects to deliver to customers in 2027.

SkyWater, a U.S. contract chipmaker that IonQ agreed in January to buy for about $1.8 billion and took over on July 31, has made D-Wave’s annealing processors for years. Its Minnesota fab fabricated the Advantage2 prototype used in D-Wave’s March 2025 paper in Science, under a wafer-supply agreement D-Wave signed with Cypress Semiconductor in 2012 that was later assigned to SkyWater.

A company with no award now owns a foundry that one of the four funded developers has relied on. With GlobalFoundries working for PsiQuantum and Quantinuum, three of the four funded developers work with one of two outside foundries. Commerce’s portfolio is diversified by architecture more than by manufacturer.

What the agreements give Commerce

D-Wave, Rigetti and Quantinuum describe these terms in their 8-Ks:

• Voting. Commerce does not vote its shares except on changes that would adversely affect the stock it owns and, at Rigetti and D-Wave, on mergers.
• Resale. Commerce cannot sell shares to a competitor in a private transaction, and each company has agreed to register Commerce’s shares for resale.
• Inventions. The U.S. government receives a license to inventions made with award money, and the Rigetti and D-Wave agreements give Commerce march-in rights.
• Domestic production. Each agreement includes domestic-control or domestic-production requirements. Quantinuum’s requires funded inventions to be predominantly produced in the United States, and bars transferring the resulting intellectual property to a foreign entity of concern, for the performance period and at least ten years after it.
• Repayment. At Rigetti and Quantinuum, Commerce can demand repayment of disbursed money for missed required project activities or for material breaches of the security or domestic-production terms.

Rigetti’s filing warns that Commerce’s combined role as shareholder, counterparty and regulator could complicate the company’s deals and relationships, and could lead other governments to ask for equity in return for their own support. D-Wave’s filing notes that no other agency or branch of the U.S. government has committed to support the transactions, that a future court, Congress or administration could find them “unauthorized, void or voidable,” and that the company has not finished working out how to account for them.

Other agencies also fund these companies directly: PsiQuantum signed a $125 million agreement with DARPA’s Quantum Benchmarking Initiative in July. Nothing in the filings shows a conflict between Commerce’s roles. The 8-Ks also do not say how Commerce will keep its milestone decisions separate from its interest as a shareholder, and that is the question I would put to the department.

Milestones due by 2031

All three public OTAs run for up to five years from September 4, 2026, so their milestones fall due by September 2031 unless Commerce agrees to an extension. Two company targets for fault-tolerant machines come before that date: IBM aims for a large-scale fault-tolerant computer by 2029, and PsiQuantum’s chief executive, Victor Peng, told Reuters in July that its first commercial-scale machine should be running around 2030. DARPA’s Quantum Benchmarking Initiative is testing whether an industrially useful quantum computer can be built by 2033, two years later.

Infleqtion’s final award will be the next public check on these terms. It is the only public company among the three developers still at the letter-of-intent stage, so its 8-K will show whether Commerce gives a neutral-atom developer the same deal: shares for the full award at signing, and cash against milestones.