Washington Owns $300M of Quantum Rivals; IonQ Bets $1.8B on Its Own Fab

Today on Quickly Quantum: what happens when the U.S. government becomes a shareholder in three rival quantum computing companies on the very same day — and across town, IonQ bets its whole roadmap on owning the chip factory instead of renting one. Before that, in the headlines: GlobalFoundries lands its own nine-figure quantum manufacturing check, and a Cleveland Clinic, RIKEN and IBM team lands in the finals for computing's most prestigious prize. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Wednesday, September 9, 2026. Two stories today genuinely both deserve the top slot, and they're pulling on the exact same thread — how America actually builds these machines at scale. Let's get into it.

First up: GlobalFoundries. Quantum Zeitgeist reports the chipmaker just secured three hundred seventy-five million dollars over five years from the Commerce Department to expand its Quantum Technology Solutions business — a separate award from the deals you're about to hear about with the hardware companies themselves. The money's aimed at cryogenic CMOS — chip electronics built to work at the ultra-cold temperatures quantum machines need — plus advanced packaging and what's called heterogeneous integration, basically stitching different chip types together on one platform. Nicholas Sergeant, who runs Quantum Technology Solutions at GlobalFoundries, called it, quote, 'another important milestone for Quantum Technology Solutions and our efforts to build a scalable domestic quantum manufacturing ecosystem,' end quote. Quantum Zeitgeist also notes GlobalFoundries has a separate three hundred million dollar letter of intent with the CHIPS office for silicon photonics work. Now, it's worth flagging this is reported so far by one trade outlet, so we haven't independently confirmed the exact award terms.

Next: a genuine feel-good item for the field. A team spanning Cleveland Clinic, RIKEN, and IBM just advanced to the finals for the twenty twenty-six ACM Gordon Bell Prize — the top honor in supercomputing. What'd they do? They simulated the largest biologically meaningful molecule ever modeled with a quantum computer, a protein system with twelve thousand six hundred thirty-five atoms. The trick was combining IBM's Quantum Heron processors — you can picture them as the quantum half of the equation, using up to ninety-four qubits and running nearly six thousand quantum operations — with three of the world's fastest supercomputers: Fugaku, Miyabi-G, and RIKEN's newer system, ROQUO. The quantum piece handled the hardest part of the physics, and the classical supercomputers reassembled everything else. The goal is modeling how potential drugs bind to protein targets, cheaper and with fewer manual data transfers than doing it purely classically. The winner gets announced at the SC26 conference in Chicago, November fifteenth through twentieth. It's a nice reminder that not every quantum story is about qubit counts — sometimes it's about getting a hard number crunched. And speaking of getting hard numbers crunched at scale, that's exactly the fight brewing in our two lead stories today.

Our first lead today: the government just went shopping for quantum computers, and it didn't just write checks — it took equity. According to reporting from Yahoo Finance and Asianet Newsable, D-Wave, Rigetti, and Quantinuum each landed one hundred million dollars in CHIPS Act research money — three hundred million dollars total — with the U.S. government taking a minority, non-controlling equity stake in every single one of them. Now, if you're new to this corner of the industry, here's the quick primer: these are three of the most established American quantum hardware companies, and Washington didn't put its money behind just one of them. It bought a slice of three separate, competing bets, all at once.

That's the tension worth sitting with. Is this smart industrial policy — spreading early bets the way a venture fund would across a portfolio, on the theory that at least one pays off? Or is Washington quietly taking on the commercial risk of a technology that's still years from paying for itself, and doing it with public money instead of private capital? Taking equity in three direct competitors on the same day reads like a hedge to me — rather than guessing right now which company eventually wins, the government is positioning itself to own a piece of whichever one does.

And it's worth pausing on why that distinction matters. A grant is money tied to deliverables — build this, hit this milestone, here's your check. An equity stake is different. It means the government now has a financial interest in these companies' valuations, their future fundraising rounds, even what happens if one of them gets bought out. That's Washington becoming, in a pretty literal sense, a shareholder in the outcome, not just a funder of the research.

My honest skepticism here: none of this is commercial revenue. Nobody closed a sales deal today, and R&D money is not the same thing as a working product in a customer's hands. What actually matters, a year or two out, is whether these three companies can point to hardware that justifies the money — and today's reporting doesn't spell out exactly what technical milestones this funding is tied to or when they're due, which is itself worth flagging rather than glossing over. The number worth remembering isn't the size of the check Washington wrote today. It's whatever these companies can actually demonstrate down the road.

Now, there's a genuinely open question nobody's answered yet: what happens to the government's stake if one of these three companies gets acquired, or taken private? Does that equity get bought out at a premium? Does it just sit there? Today's reporting doesn't address it, and it's worth sitting with before this becomes a template for how the U.S. funds strategic technology going forward, because right now it's a first at this scale for the sector.

And it raises a related question the coverage doesn't settle either: does taking a stake in three competing approaches at the same time suggest Washington itself doesn't yet know which one is going to win? You could read that as the honest, humble version of industrial policy — spread the bet, let the technology sort itself out. Or you could read it as a sign that nobody in the room had enough conviction to pick a horse. I don't think today's reporting gives us enough to settle that either way, and I'd rather say that plainly than pretend it does.

Our second lead swings from three companies renting fabrication capacity to one company that just bought its own factory outright. IonQ held its first investor day since closing its roughly one-point-eight-billion-dollar acquisition of SkyWater Technology, and CEO Niccolo de Masi framed Tuesday's event, held at the New York Stock Exchange, as a six-announcement day. The headline hardware reveal was Superion 256 — IonQ's first two-hundred-fifty-six-qubit trapped-ion processor built fully in-house at SkyWater's fab. The system incorporates what IonQ calls Electronic Qubit Control technology, obtained through its acquisition of Oxford Ionics — replacing much of the laser equipment traditionally used to control trapped-ion qubits with electronics integrated directly into the chip. The pitch, if you're wondering why that matters, is that a system built this way is easier to standardize and manufacture at scale, and it's designed to fit into a regular data-center rack rather than needing a custom facility. IonQ says the chip went through six tapeouts — that's the point where a finished chip design gets sent to the foundry to actually be fabricated — in just the first half of this year, and that working directly with its own fab cut its design cycle from nine months down to two, while letting it run twelve times more wafer lots over six months than at its previous foundry. Chris Ballance, IonQ's president of quantum computing, said the platform is meant to simplify the hardware and let each generation build directly on the last one's manufacturing process. One detail worth noticing: IonQ sold the first Superion 256 system during the first quarter of 2026, before it had even publicly named the platform — the customer was the University of Cambridge, as part of a wider agreement covering quantum computing, networking, and intellectual-property development. Customer deliveries for everyone else aren't scheduled until 2027. So today's announcement is real manufacturing progress — customers just aren't holding these qubits yet. De Masi's quote captures the pitch well: quote, 'Superion is the result of two strategic acquisitions coming together to deliver this historic milestone. Oxford Ionics enabled IonQ to control natural, trapped-ion qubits using standard electronics, and SkyWater unlocked the ability to manufacture at semiconductor costs and scale,' end quote. IonQ also announced it's building a bigger machine in parallel rather than waiting for the first one to ship — Superion 10K, aiming for roughly ten thousand physical qubits and built to run the company's Walking Cat fault-tolerant architecture, published back in April, which lays out how the system would detect and correct its own errors. That version adds cryo-CMOS electronics — chip circuitry designed to work at the extreme cold quantum systems need — and IonQ showed early test chips at the event. The other three announcements matter for different reasons. IonQ raised its full-year twenty twenty-six revenue guidance to a range of four hundred fifty to four hundred sixty million dollars. That raise mostly reflects folding SkyWater's own existing revenue into IonQ's books starting July thirty-first — not new sales of quantum computers, so don't read too much into that number if you're tracking real growth. IonQ also announced an eight-point-one-eight-million-dollar quantum-security agreement with a company called Congruity360, tied to the threat quantum computers eventually pose to elliptic-curve signatures, the encryption securing much of today's internet traffic. The detail that ties this back to our first lead: SkyWater launched its own quantum foundry service, and its first customer is Qolab — a superconducting quantum computing company. The fab IonQ bought to build its own trapped-ion machines is now also manufacturing chips for a rival architecture. Does that make SkyWater a neutral foundry serving the whole industry, or does IonQ now have a window into a competitor's manufacturing roadmap? Neither company has addressed that directly. And what's missing from today's announcements entirely: any gate fidelities or error rates for the Superion prototypes — the numbers that actually tell you whether more qubits adds up to a more useful computer.

Now, put these two stories side by side, and the shared thread is obvious: everybody in quantum hardware right now is fighting the exact same bottleneck — manufacturing qubits reliably at scale. Washington is spreading its bet, taking equity in three competing companies at once. IonQ chose depth over breadth, buying the factory outright and betting that owning the whole stack, design through fabrication, wins over renting capacity from a partner. GlobalFoundries shows up in both stories too — it just landed its own CHIPS money, and through SkyWater, it's now fabricating chips for Qolab, a rival superconducting quantum computing company to IonQ. Neither strategy has proven itself yet: no fidelities from IonQ, and none of today's CHIPS-funded companies has yet delivered the kind of hardware milestones this funding is meant to unlock. Time for the Hype Check. I'm putting this pairing at a five: real money, real fabs, real manufacturing moves on paper, but every headline number today describes a plan or a roadmap, not a delivered, working machine customers are running today.

If today's episode helped you make sense of who's actually building these machines, follow Quickly Quantum wherever you listen so tomorrow's episode lands automatically. This has been Quickly Quantum, an AI-voiced podcast, created and built by a real human using today's cutting-edge technology. Nothing you heard on this show is financial advice. I'm Brian Lampert, and I'll catch you all tomorrow — take care!

I also host Concrete Compute: a daily briefing on the AI buildout. The datacenters, the megawatts, and who actually pays for them. Find it wherever you get your podcasts.