The Government Is On Cap Tables Now. Deeptech Founders Should Plan For It.
In May 2026 the Commerce Department committed $2.013 billion to nine quantum companies and took equity in every one. Washington now holds stakes in more than 26 companies. Here's what government capital actually costs a founder, and what to do at seed stage when you don't qualify.
On May 21, 2026, the Commerce Department signed letters of intent with nine quantum computing companies worth about $2.013 billion. IBM took the biggest piece, $1 billion to stand up a superconducting quantum foundry subsidiary. GlobalFoundries got $375 million. Atom Computing, D-Wave, Infleqtion, PsiQuantum and Quantinuum got $100 million each. Diraq got $38 million. Every award carried the same condition: a minority, non-controlling equity stake for the federal government.
That's not a grant. That's a cap table entry.
And it isn't a one-off. Since January 2025, Washington has been buying. $8.9 billion into Intel for a 9.9% position in August 2025. Roughly 15% of MP Materials in July 2025. Five percent of Lithium Americas from the Energy Department, and 10% of Trilogy Metals for $35.6 million. By mid-2026 the tally sat between $23.9 billion across 26 confirmed deals and roughly $27 billion across thirty-odd, depending on who's counting.
So if you're building in semiconductors, quantum, critical minerals, energy, or anything defense-adjacent, you've probably asked the obvious question. Can I get some of that? Here's the practical answer, including the part most people won't say out loud.
So what did Washington actually buy?
The deals don't share a structure. They share a direction.
| Company | Date | What the government took | Sector |
|---|---|---|---|
| MP Materials | July 2025 | ~15% via preferred convertible stock and warrants | Rare earths |
| Intel | August 2025 | 9.9% at $20.47/share, plus a 5-year warrant at $20 | Semiconductors |
| Trilogy Metals | October 2025 | 10% for $35.6M, warrants on 7.5% more | Critical minerals |
| Lithium Americas | October 2025 | 5% of the company, plus 5% of the GM joint venture | Lithium |
| Nine quantum firms | May 2026 | Minority non-controlling stakes across $2.013B | Quantum |
The differences matter more than the headline percentages. Intel's stake was explicitly passive: no board representation, no governance rights, no special information rights. MP Materials was a different animal. The Department of Defense became the largest shareholder, and the package came bundled with a long-dated price floor on rare earth oxide, an offtake commitment, and a $150 million loan for heavy separation at Mountain Pass.
There's no template. Each was negotiated separately, by different agencies, under different legal authorities: CHIPS Act money for Intel and the quantum cohort, Defense Production Act authority for MP Materials, Energy Department programs for Lithium Americas.
It's not a grant, and calling it one will cost you
Here's the thing founders get wrong. They hear "federal award" and mentally file it next to a research grant. Non-dilutive. Free. Light on strings.
It's none of those things anymore.
When Commerce takes a minority equity stake as a condition of the money, that's dilution. Real dilution, at a valuation the government negotiated. D-Wave said plainly that all $100 million of its award came in as an equity investment. And the equity is the cheap part. The expensive part is everything stapled to it.
Intel's warrant is the cleanest example. Five years, 5% of the company at $20 a share, exercisable only if Intel stops owning at least 51% of its foundry business. That isn't a financing instrument. It's a strategy lock.
I've watched founders treat a government term sheet like a rebate check and skip straight past the covenants. Don't. Read the domestic manufacturing commitments, the milestone conditions, and the change-of-control language before you read the headline number. Those clauses will shape your next four years more than the dollar figure will.
Does a federal stake help or hurt your next round?
Both. Which one wins depends on who you're pitching.
On the upside, a government position is the loudest validation signal available in hard tech. Trilogy Metals shares more than tripled the morning its stake was announced in October 2025. MP Materials gained roughly 224% across 2025. Quantum names jumped on the May 2026 news before a dollar moved. Investors read a federal check as evidence that the technology is real and there's a buyer waiting at the end of the road.
On the downside, plenty of funds now treat a government line item as something to underwrite rather than celebrate. What they worry about, roughly in order:
- Exit optionality. A foreign acquirer is effectively off the table, and CFIUS review shifts from a risk to a certainty.
- Strategy flexibility. Domestic manufacturing commitments are expensive and hard to unwind if your unit economics move against you.
- Political durability. Administrations change. A deal signed under one set of priorities can become an awkward asset under the next.
- Structure. Preferred convertible stock plus warrants plus milestone conditions is not a cap table most seed funds enjoy diligencing.
Neither reaction is wrong. Just know which room you're in before you lead with it.
Your AI co-founder is ready when you are.
Foundra turns everything in this article into an actual plan. Validation, customers, pricing, launch. In one place, in your voice, in an afternoon.
Get started→$39/month. Cancel anytime.
How long does this take, and what it does to your runway
Quarters. Sometimes years. Never weeks.
The May 21, 2026 quantum announcement was letters of intent, not wire transfers. Definitive agreements and actual disbursement follow on their own clock. Between the first serious conversation and money in the bank, plan on nine to eighteen months for anything with an equity component, and assume it can die at any point for reasons that have nothing to do with your technology.
Which leads to the one runway rule that matters: never put government equity in your base case. Model it as upside in a separate scenario, and know which month you'd need a bridge if it slips two quarters. You can map that in a spreadsheet, in Notion, or in a planning tool like Foundra that walks first-time founders through financial projections section by section. The tool isn't the point. Having the second scenario written down is.
Who is this available to? Probably not you, yet
The short answer is that almost nobody at pre-seed or seed qualifies, and pretending otherwise burns months you don't have.
Look at who got picked. IBM. GlobalFoundries. Intel. PsiQuantum, which had already raised well over a billion dollars privately. MP Materials, which runs the only currently active rare earth mine in the country. Trilogy Metals, sitting on Alaskan copper and cobalt.
The common thread isn't stage or story. It's that each one controls an asset the government can't conjure up somewhere else: a fab, a mine, a magnet line. If your company is four people and a promising simulation, you don't have that asset yet. You might in six years. Build toward it. Fund the next twenty-four months some other way.
What to actually do at seed stage instead
Federal money is still very available to small deeptech companies. It just doesn't come with an equity stake attached, which is a feature.
- SBIR and STTR. NSF restarted its programs in 2026 with a $250 million commitment: Phase I up to $305,000, Phase II up to $1.25 million, and a new Strategic Breakthrough tier reaching $30 million for select Phase II companies. Run the full sequence and you're looking at potentially north of $31 million without a point of dilution. Start with the project pitch, not the full proposal.
- DIU and Other Transaction Agreements. The Defense Innovation Unit had $979 million in FY2026 and awards prototype agreements in 60 to 90 days, fast by any procurement standard. OTAs let the government negotiate terms, pricing, and data rights closer to commercial norms. One catch: DIU funds adaptation of things that already work, not pure research.
- DOE, ARPA-E, and the national labs. Cooperative agreements, CRADAs, and facility access. Cheaper than building your own.
- Customer-funded development. A program office paying you to build the thing beats any award. It's revenue, it's validation, and it doesn't dilute.
- State and regional programs. SBIR matching funds, manufacturing tax credits, land and facility deals. Unglamorous, and often the difference between a pilot line and a slide.
Two things to avoid. Don't hire a grant consultant before you've read a solicitation yourself, and don't reshape your roadmap to fit a topic call. Companies that chase award language end up with a scrapbook of small wins and no product.
How to become the kind of company Washington can invest in
If these deals are a five to ten year target rather than a this-year plan, a few things are worth doing now.
Keep your supply chain and manufacturing intent domestic where you reasonably can, and document the decisions as you make them. Watch your foreign limited partner exposure, because a fund with the wrong LPs can complicate a future review in ways that are painful to unwind. Get comfortable with government contracting mechanics early, even at tiny dollar amounts. And build toward owning something physical and hard to replicate, because that's what every one of these deals was really buying.
None of it guarantees anything. It's just the set of conditions under which the conversation becomes possible.
Frequently asked questions
Can a pre-seed startup get a US government equity investment?
Realistically, no. Every announced deal so far has gone to a company with a fab, a mine, a production line, or hundreds of millions already raised. Focus on SBIR, STTR, and DIU prototype agreements, which are built for small companies and don't take equity.
Is government equity investment dilutive?
Yes. Commerce conditioned its quantum awards on a minority, non-controlling equity stake, and D-Wave confirmed its full $100 million came in as an equity investment. Treat it as a priced round with an unusual investor and unusual covenants, not as a grant.
Does the government get a board seat?
It depends on the deal. Intel's was passive, with no board representation and no governance or information rights. Other arrangements included warrants, offtake agreements, and price floors, so read the specific documents rather than assuming a standard.
Will taking government money scare off venture investors?
Some, not most. The bigger concerns are exit optionality, since a foreign acquirer becomes very hard, and the ongoing cost of domestic manufacturing commitments. In hard tech, the validation signal usually outweighs the friction.
How long does the process take?
Plan on nine to eighteen months from first serious conversation to funds received, and remember that announcements are often letters of intent rather than closed deals. Keep the money out of your base case model entirely.
You just read the theory. Ready to build the thing?
Foundra is your AI co-founder. It turns an idea into a validated business plan, a go-to-market, and your first 10 customers. In an afternoon, not a semester.
$39/month. Cancel anytime. Works in 20 languages.