The Week in Quantum: Foundries, MOUs, and a Nasdaq Pop

Today on Quickly Quantum: the week quantum computing traded qubit-count bragging rights for buildings, budgets, and stock tickers. Before we get into the full arc, here's the tease — a French quantum company nearly doubled its share price on day one of trading, and Saudi Arabia just added a second quantum vendor to its shopping list. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Saturday, September 5, 2026, and this is our weekly recap — the whole arc of the week, in one sitting. And there's a lot of arc to cover: IBM shipped a faster chip, two national labs plugged quantum hardware straight into their supercomputers, and Europe kept laying the industrial groundwork underneath all of it. If you only caught pieces of this week, here's where it all connects. Ready? Let's get into it.

Quick recap on IBM's Nighthawk r2, since this one already got the full treatment this week: it's a hundred-twenty-qubit processor, and its headline claim is a twenty-five-fold jump in circuit speed over the prior Heron generation, thanks to a redesigned reset mechanism — the process that clears a qubit's state between runs — which also cuts initialization error. IBM itself flags a catch: that twenty-five-times number is workload-dependent. Short circuits or low-repetition experiments barely notice the improvement; it's large-scale, repetitive workloads — the kind IBM's own error-correction research leans on — that see the real gain. Does that mean every job you run on IBM's cloud suddenly finishes twenty-five times faster? Not quite. This is a chip built for a specific class of heavy-lifting work, and the press release lets the headline number do more talking than the fine print supports.

Sticking with national labs — PsiQuantum and Brookhaven National Lab, part of the Department of Energy, announced a partnership this week using PsiQuantum's free software platform, Construct, so Brookhaven scientists can design and optimize algorithms for fault-tolerant quantum computers — machines with built-in error correction — before that hardware even exists. As @PsiQuantum put it on X, announcing the deal, they're 'preparing together for the quantum future,' with Construct supporting Brookhaven researchers on applications for the first generation of fault-tolerant systems. Worth being blunt about what this actually is: a software and algorithm-development agreement, not a hardware delivery. There's no fault-tolerant machine sitting at Brookhaven yet for anyone to actually run these algorithms on. It plugs into the Department of Energy's Quantum Genesis initiative, aiming for a scientifically relevant fault-tolerant system by 2028 — so what you're hearing about today is prep work for a machine that's still years from existing. We've said on this show before that PsiQuantum hasn't shipped a working large-scale machine yet — that's still true today, and this deal doesn't change it, it just adds Brookhaven to the waiting list of scientists lined up for whenever it does.

Over in the Gulf, Quantinuum and Saudi energy giant Aramco signed a non-binding memorandum of understanding in Riyadh this week to explore industrial quantum computing use cases and benchmark Quantinuum's trapped-ion hardware — qubits made from individually controlled charged atoms. This extends a thread you've heard on this show before: Aramco already runs a Pasqal neutral-atom machine in Dhahran, and now it's adding a second quantum vendor to the mix. Aramco clearly likes optionality — one neutral-atom vendor already running hardware, one trapped-ion vendor now in early talks. And let's be clear what a non-binding MOU actually is: a signal of intent, not a committed research budget or hardware on the ground — the companies are still at the technical-onboarding stage. Will this convert into a paid deployment the way Aramco's Pasqal relationship did? That's the question that decides whether this MOU means anything in a year.

Now for the number that stopped me cold this week: Pasqal, the French neutral-atom quantum company, went public via a SPAC merger, and its shares closed up ninety-five point two percent on day one, finishing at nineteen dollars and eleven cents. That's part of a roughly two-billion-dollar valuation deal that handed the company about three hundred sixty million dollars in cash. But that valuation comes attached to a real caveat: Pasqal is still unprofitable, with around sixteen point five million euros in 2025 revenue backing up that two-billion-dollar number. That's a classic pre-revenue quantum-stock story — the market betting on where the technology goes, not on what it's earning you today. How Pasqal's SPAC route stacks up against IQM's earlier Nasdaq listing this year is the comparison analysts will be drawing all quarter.

One story flew under the radar this week — Europe's push to turn its quantum research strength into actual industry. European quantum leaders convened at the European Parliament for a three-hour session on the forthcoming EU Quantum Act, debating how the continent moves from publishing papers to building companies, ahead of an announcement the European Commission is expected to make by the end of this year. I want to be precise about what this is and isn't: there's no legislative text yet, this was a stakeholder session, and EU tech legislation has a well-worn habit of slipping its timelines. So the honest read is that this is still a proposal being shaped, not a law with teeth. But it matters as the backdrop for everything else happening in European quantum right now — it's the regulatory conversation that companies building chip foundries and courting investors are all angling around. What concrete industrial-capacity commitments came out of this session, versus general political rhetoric about Europe's research edge? Honestly, the coverage doesn't make that clear, and I'd rather name that ambiguity than pretend otherwise. If you're a European quantum company, this is the framework you're waiting on before committing serious capital. If you're anyone else watching from outside Europe, it's a signal that the continent knows it has a manufacturing gap and is trying to legislate its way toward closing it.

That manufacturing gap has an actual fix in the works — Quantum Foundry Copenhagen and the Novo Nordisk Foundation announced plans this week for a fifty-three-hundred-square-meter quantum chip fabrication facility, opening in 2027, offering commercial wafer fabrication — the actual process of manufacturing chips on silicon wafers — to quantum vendors globally. As @HPCwire reported on X, unveiling the Copenhagen facility plans, this sits inside the Foundation's broader three-hundred-ninety-million-euro quantum commitment. What's notable is who's writing the check: not a government, not a venture fund, but a private foundation — a rare funding source for manufacturing infrastructure, and one that directly addresses the fabrication bottleneck the Quantum Act discussions flagged this same week as Europe's weak spot. The caveats here matter: this facility won't open until 2027, so it's a plan, not operating capacity yet, and outside cost estimates for the building itself — reported around sixty-two million dollars by some outlets — weren't in the official release, so don't confuse that figure with the Foundation's total three-hundred-ninety-million-euro commitment; those are two very different numbers. Still, in an industry that keeps saying its biggest bottleneck isn't qubit count but the supply chain underneath it, a foundation stepping in with real fabrication capacity is the kind of infrastructure story that matters more to you long-term than another qubit-count headline, even if it takes until 2027 to prove out. How this foundry's eventual capacity compares with the chip fabs already serving quantum hardware makers is a question we'll get an answer to once it actually opens its doors.

Closer to home — George Mason University and Oxford-based TreQ announced a seven-point-seven-million-dollar partnership this week to install the first U.S. deployment of TreQ's Open Architecture Quantum framework — a modular, vendor-agnostic system designed to swap in different quantum hardware over time — at a Northern Virginia campus. Delivery is set for early 2027, with commissioning by late summer 2027. That's a smaller dollar figure than the Copenhagen foundry, but it's notable for a different reason: this is happening in Northern Virginia, walking distance from Washington, in a region thick with defense and government users who need a hardware testbed without building one from scratch. What vendors actually line up to plug into that testbed first? The announcement doesn't say yet. The caveat that matters most: this is a funding and partnership announcement, not an operating machine sitting in a lab you could visit today — installation is still more than a year out, and commissioning doesn't happen until deep into 2027. It's one more data point in the university-quantum-infrastructure boom running in parallel with all these corporate deals this week — a smaller check, but the same instinct: get quantum hardware physically installed before you've fully figured out what to run on it.

We covered the launch of Jülich's trapped-ion machine, JION, earlier this week — the short version is that Forschungszentrum Jülich and eleQtron officially brought it online at the Jülich Supercomputing Centre. What's new this week is the detail: JION runs on ytterbium-ion qubits, and it's now linked directly to the JUPITER exascale supercomputer through the JUNIQ platform, pairing a quantum processor with one of the world's fastest classical machines. Why does that exascale link matter? Because it means researchers can steer quantum experiments using classical supercomputing power sitting right next to it, instead of shipping data back and forth over a network. Alongside the launch came new German state funding — two streams called SQALING and Q-STAR.NRW, each funded with up to approximately twenty-five million euros, putting the combined total at around fifty million euros for scaling up trapped-ion and semiconductor quantum technology. The honest framing: this is a research-access machine tied to a state-funded consortium, not a commercial product you or I could buy time on, and those scaling programs are funding commitments, not delivered hardware yet. Still, pairing quantum and exascale classical computing under one roof is exactly the kind of infrastructure integration this whole week has been about.

Time for a quick gut check on the week overall. Strip out the not-yet-shipped installs and the non-binding agreements, and what actually shipped or traded this week comes down to two things: IBM's faster chip and Pasqal's stock debut — everything else you heard today is a promise with a date attached, somewhere between 2027 and 2028. Real hardware, real trading, and a pile of MOUs still waiting to prove themselves — that's the case this week's news actually makes for itself. So here's where the substance-versus-hype dial lands: a six out of ten. Not bad for an industry that loves a press release. If you want the daily version of all this instead of the weekly wrap, follow Quickly Quantum wherever you get your podcasts. 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!