Nine Stocks, One Buyer
On May 21, the Department of Commerce signed nine letters of intent worth $2.013 billion - and in exchange, the United States government will take a minority, non-controlling equity stake in every company on the list. Not loans. Not grants with reporting requirements. Stock.
The list: IBM, GlobalFoundries, Atom Computing, D-Wave, Infleqtion, PsiQuantum, Quantinuum, Rigetti, and Diraq. The tape reacted the way it always does when Washington says the word quantum: D-Wave closed +33.4%, Rigetti +30.6%, and IonQ - which isn't even one of the nine - finished +12.2%.
We used Barebone to pull the Commerce announcement, the IBM foundry term sheet, the Quantinuum S-1, and the Q1 filings of every public name on the list. The real story isn't the size of the check. It's the shape of it: just under half the money goes to a single company, 68 cents of every dollar goes to manufacturing rather than computing, and the post-announcement pop - up to +65% in three sessions - was already leaking air by May 27.
We'll show you the receipts, including the ones that should make a quantum bull uncomfortable.
What $2 Billion Actually Bought
Here is the full program, from the Commerce Department's May 21 announcement:
| Company |
Proposed award |
Role |
Qubit approach |
| IBM (Anderon) |
$1,000M |
Foundry |
Superconducting first, other modalities later |
| GlobalFoundries |
$375M |
Foundry |
Superconducting, trapped ion, photonic, topological, silicon spin |
| Atom Computing |
$100M |
Computing |
Neutral atom |
| D-Wave Quantum |
$100M |
Computing |
Superconducting (annealing + gate-model) |
| Infleqtion |
$100M |
Computing |
Neutral atom |
| PsiQuantum |
$100M |
Computing |
Photonic |
| Quantinuum |
$100M |
Computing |
Trapped ion |
| Rigetti Computing |
up to $100M |
Computing |
Superconducting |
| Diraq |
up to $38M |
Computing |
Silicon spin |
The IBM line isn't really an award to IBM. The $1 billion funds Anderon, a standalone company IBM is spinning up in Albany, New York - America's first purpose-built, pure-play quantum chip foundry. IBM matches the government's billion with $1 billion of its own cash plus IP, equipment, and staff. Anderon will run a 300-millimeter quantum wafer line - superconducting qubits first, serving multiple hardware vendors - with chips for fault-tolerant systems targeted for commercial deployment by 2029.
GlobalFoundries' $375 million buys the second leg: a US-based fab that can manufacture chips across essentially every qubit technology in the table above.
Add it up and the two foundries take $1.375 billion of the $2.013 billion - 68%. The seven companies that actually design and sell quantum computers split the remaining $638 million.
A caveat before anyone gets carried away: these are letters of intent, not closed deals - Commerce describes the funding as planned, and the Rigetti and Diraq awards are explicitly "up to" figures.
IBM, GlobalFoundries, D-Wave, and Rigetti trade on US exchanges today. Quantinuum filed its S-1 on May 8 to list on Nasdaq under the ticker QNT. The rest are privately held or still on their way to public markets.
The Taiwan Scar
Why would a government buying into the computers of the future put two-thirds of its money into factories instead?
Because Washington has made this mistake before. America invented the integrated circuit, then spent four decades letting the manufacturing migrate overseas - mostly to one island a hundred miles off the Chinese coast. By the time advanced chips became the most strategically important objects on Earth, America couldn't make them at scale. The 2022 CHIPS Act was a $52.7 billion apology for that, and this quantum program is funded out of the same law.
The lesson Commerce appears to have internalized: owning the design was never the problem. America never lost the science. It lost the making - and with it, the leverage.
So the quantum program inverts the old pattern. Fund the fabs first, while the industry is still small enough that $1.4 billion buys the entire manufacturing layer. If quantum computing works, every American quantum company may end up running its wafers through Albany.
The signal has been priced as nine stock picks. It's actually one supply chain.
Five Horses, One Stable
The second pattern in the awards is harder to see and more interesting. Nobody - not IBM, not Google, not the Commerce Department - knows which physical technology will produce a useful quantum computer. A qubit, the quantum version of a bit, can be built from superconducting circuits, trapped ions, neutral atoms, photons, or electron spins in silicon. Five approaches, five different bets on physics.
Commerce funded all five.
| Approach |
The idea |
Funded |
Dollars |
| Superconducting |
Circuits chilled to near absolute zero |
IBM, D-Wave, Rigetti |
$1,200M |
| Neutral atom |
Atoms held in place by laser arrays |
Atom Computing, Infleqtion |
$200M |
| Trapped ion |
Charged atoms in electromagnetic traps |
Quantinuum |
$100M |
| Photonic |
Particles of light as qubits |
PsiQuantum |
$100M |
| Silicon spin |
Electron spin inside silicon transistors |
Diraq |
up to $38M |
| All of the above |
Multi-modality foundry |
GlobalFoundries |
$375M |
"The CHIPS R&D Office is taking a portfolio approach to strengthen and accelerate U.S. leadership across multiple quantum modalities." - Bill Frauenhofer, executive director, CHIPS R&D Office
That's a fund manager's sentence, not a procurement officer's. Superconducting - the most mature approach, the one behind both IBM's and Google's flagship machines - gets roughly 60% of the program. The most speculative approach, silicon spin, gets a $38 million lottery ticket. The government isn't trying to pick the winning horse. It's buying the stable.
And the omission is as telling as the picks: IonQ, the best-known pure-play quantum stock in the market, got nothing. Its slot - trapped ion - went to Quantinuum instead. The market bought IonQ anyway. Hold that thought.
The Number Behind the Panic
The standard pitch for quantum computing is a maze. A classical computer solves a hard problem the way you'd solve a maze: one path at a time, backtrack, try again. A quantum computer exploits superposition - a qubit holding a blend of states at once - to weigh enormous numbers of paths simultaneously. For a narrow class of problems, that's not an incremental speedup; it's a different universe of speed.
Google made the point theatrically in December 2024 with Willow, a 105-qubit chip that ran a benchmark in under five minutes that Google estimated would take the world's fastest classical supercomputer ten septillion years. The benchmark was contrived - chosen because quantum machines excel at it. The pace is still the point.
But the workload that actually moves governments is uglier than benchmarks: a sufficiently powerful quantum computer running Shor's algorithm could factor the large numbers underlying RSA encryption - the math protecting bank transfers, military communications, and most of the world's stored secrets. Hence "harvest now, decrypt later": adversaries vacuuming up encrypted data today to crack once the hardware exists. NIST published its first post-quantum encryption standards in 2024 for exactly this reason.
That's where China enters. The figure repeated in nearly every policy memo is that Beijing has committed roughly $15 billion of government money to quantum - about four times comparable US federal spending. We should be honest about that number, because almost nobody else is: it is an estimate of announced funding, it has never been officially confirmed, and Chao-Yang Lu - one of China's most prominent quantum physicists - has suggested the real figure may be about a third of what's publicized. The 4:1 ratio is soft.
Policy is being made on it anyway. The Commerce release is explicit about why this money exists: national security, technological resilience, strategic leadership. This is a defense program wearing an investment program's clothes.
The Intel Template
If the structure feels familiar, it should. In August 2025, the government converted Intel's CHIPS grants into 433.3 million shares at $20.47 - an $8.9 billion position, 9.9% of the company, plus a warrant for 240.5 million more shares. By May 6 of this year, with Intel at $113.01, that stake was up roughly +452%, a paper gain of about $40 billion. Trump said the quiet part in April: "taking pieces of the Equity for support."
The quantum LOIs are that template's first systematic reuse - not one negotiated rescue but nine simultaneous stakes across an entire industry, signed before the industry's product fully exists. The precedent is now load-bearing: equity-for-support is a repeatable instrument, and the Intel trade made it look like a brilliant one.
One difference matters, though. Intel had $13 billion revenue quarters and a turnaround story. The quantum seven are something else entirely.
A Sector Priced on Stories
Here is the part the press releases skip. These are the most recent quarterly numbers for the public pure-plays, all reported in May:
| Company |
Q1 2026 revenue |
Year-over-year |
Proposed award |
| IonQ |
$64.7M |
+755% |
$0 |
| Rigetti |
$4.4M |
~3x |
up to $100M |
| D-Wave |
$2.86M |
-81% |
$100M |
Sit with that table. D-Wave - a company whose stock added a third of its value in one session on this news - booked $2.86 million of revenue last quarter, down 81% from a year earlier. The government's proposed $100 million award equals roughly 35 quarters of that revenue. Rigetti's award is about 23 quarters of its Q1 revenue; the company holds $569 million in cash - a dilution-financed war chest equal to roughly 129 times its quarterly sales. And IonQ - the only one with revenue resembling a business, guiding to $260 - 270 million for 2026, growth assembled partly through acquisitions - got nothing.
These are pre-profit, mostly pre-product companies whose prices move on announcements, not earnings. The eight weeks around this deal prove it. Between April 1 and April 15, Rigetti rose +42%, D-Wave +52%, and IonQ +56% - then the group bled into mid-May, with Rigetti giving back -16.5% from its mid-April close by May 19. Then the award headlines hit, and the cycle restarted at higher amplitude:
| Ticker |
In the deal? |
May 21 |
May 22 |
May 22 → 27 |
May 19 → 27 |
| QBTS |
Yes - $100M |
+33.4% |
+14.2% |
-6.5% |
+51.1% |
| RGTI |
Yes - up to $100M |
+30.6% |
+19.9% |
-6.8% |
+54.3% |
| IONQ |
No |
+12.2% |
+8.1% |
+2.8% |
+35.0% |
Look at what the chart is actually saying. The two awarded names spiked more than +60% in three sessions, peaked the day after the announcement, and spent the next two sessions deflating. The company that received nothing rallied +35% over the same window and was still climbing at the May 27 close - pure sector sentiment, attached to no award at all.
Twice in eight weeks, this group has put up double-digit rallies and then stalled or retraced within days. That is what it means to own a story stock: the story arrives on a headline, and headlines have no follow-through of their own. Add the structural caveats - the LOIs aren't definitive agreements, government stakes cost existing shareholders something (Intel's deal diluted holders with 433 million new shares), and Anderon's fault-tolerant 2029 target is a roadmap slide, not a product - and the honest summary is this: Washington just validated the sector's strategic importance, and changed almost nothing about the near-term economics of any company in it.
What This Means
The framework, separated into what's real and what's hope:
The supply-chain thesis is real. Two-thirds of the money builds fabs. If quantum computing reaches usefulness on any of the five modalities, the manufacturing layer Washington just funded collects from all of them - the same logic that made TSMC the most important company of the smartphone and AI eras.
The policy floor is real; the price floor is not. A government shareholder makes bankruptcy-by-neglect less likely - procurement, follow-on funding, and political attention now flow naturally to these nine. None of that stops a story-priced stock from giving back a rally as fast as it arrived: this group did exactly that after mid-April, and was doing it again within four sessions of the award.
Watch the conversions, not the announcements. Three specific signals: whether the LOIs become definitive agreements with filed terms (Intel's deal produced an 8-K with share counts and a warrant - that's the standard of proof); whether D-Wave's $33.4 million of bookings actually converts to revenue in coming quarters; and where Quantinuum's QNT IPO prices, because that will be the first time institutions - not retail momentum - set a clearing price for a quantum pure-play.
The asymmetry is honest, at least. Buyers here are paying for a small probability of owning the next strategic industry at its founding, bundled with a high probability of violent, headline-driven round trips and continued dilution. Both halves of that sentence are true at once.
The government can afford that trade - it prints the money, holds for decades, and books national security as a dividend. The question for everyone who bought the May 21 pop is simpler, and the last week of trading already started asking it: when the next headline fades, what exactly are you holding?
Data: Barebone | Sources: U.S. Department of Commerce/NIST CHIPS announcement (May 21, 2026), IBM - Commerce Anderon foundry release, Quantinuum S-1 (SEC EDGAR), IonQ, Rigetti and D-Wave Q1 2026 results | Prices as of May 27, 2026 close
