Research Finder
Find by Keyword
Does Quantum's Winner Get Decided by the Fab, Not the Physics?
Diraq says manufacturability, not spectacle, decides quantum; we weigh the claim against IBM's 300mm fab and the 2029 fault-tolerance race.
07/06/2026
Key Highlights
- Diraq's July 1 post frames "boring" as the winning attribute: a rack-deployable silicon chip on standard CMOS, running above 1 kelvin rather than at the millikelvin floor superconducting machines require.
- The thesis rests on a division of labor, with fidelity "largely solved" above 99 percent (with imec, Diraq reported single- and two-qubit gates above 99 percent on a 300mm industrial line, Nature 2025) while qubit count stays orders of magnitude short across every modality.
- Diraq is targeting sub-$1 per physical qubit and millions of qubits per chip, backed by up to $38M under the CHIPS Act, a GlobalFoundries partnership, and DARPA QBI Stage B, with an initial system aimed at 2029 and utility scale near 2033.
- That award sits inside a $2.013B Commerce program across nine companies, two foundries (IBM at $1B, GlobalFoundries at $375M) plus seven hardware makers (D-Wave, Rigetti, Atom Computing, Infleqtion, PsiQuantum, and Quantinuum at $100M each, Diraq at $38M), spanning every major modality.
- The clearest sign the manufacturing thesis has gone mainstream is IBM's Anderon, a roughly $2B pure-play quantum foundry in Albany ($1B CHIPS plus $1B IBM cash) built to fabricate wafers for the industry, the incumbent moving to own manufacturing, not just design.
- The open question is timing, not direction: manufacturability is a genuine structural edge, but it converts to advantage only if spin-qubit control and connectivity scale as cleanly as the transistor comparison implies.
The News
Diraq used a July 1 Substack post to argue that utility-scale quantum computing will be won by the most manufacturable machine, not the most photogenic one, positioning its CMOS silicon spin qubits against the "miracles" it says exotic modalities still require. The company reframes the field's signature dilution-fridge image as fundraising theater and contends that gate fidelity is largely solved above 99 percent across serious platforms, leaving qubit count, where useful machines will need millions of physical qubits, as the axis that decides the race. Arriving weeks after Diraq's $38 million CHIPS Act letter of intent and its GlobalFoundries and NVIDIA collaborations, the essay functions as a thesis statement for the semiconductor-playbook camp rather than a product disclosure. The full post is here: Diraq's Quantum Computers Are Boring.
Analyst Take
Every hardware generation eventually meets its volume-economics reckoning, and quantum's is (finally) arriving. Diraq's post is the clearest statement yet of a view gaining ground on the semiconductor side of this industry: the winner will be set by manufacturing yield curves, not benchmark records. We find the direction persuasive.
The easy counter-read is that this is a company still early on qubit counts, arguing that the popular scoreboard misses the point and framing a physics gap as a manufacturing virtue. This objection has teeth even if it sounds self-serving. But it crystalizes something we see emerging across the sector. The leaders are not standing still on manufacturability, and quietly industrialization is the next wave. The useful tension is no longer silicon versus superconducting. It is how fast any modality can convert a fabrication advantage into logical qubits that clear real workloads.
What Was Announced
To be precise about what Diraq did and did not do: this is an argument, not a product launch. The post makes four moves. It reframes the dilution-refrigerator photograph, the gold-plated chandelier that fronts most quantum coverage, as a fundraising device that equates difficulty with progress. It argues that gate fidelity is, in the sense that matters, already adequate across serious platforms, since several can now operate accurately enough to run error correction. It relocates the real contest to qubit count, where Diraq concedes it does not lead today, contending that only a design built for volume manufacturing can climb from thousands toward the millions of physical qubits utility scale appears to require. And it claims a structural moat: silicon spin qubits are patterned with the same CMOS processes used for mainstream logic, run warm enough (above 1 kelvin) to use commodity cooling, and install into a data-center rack beside classical compute.
The supporting record is more than rhetoric. With imec, Diraq reported single- and two-qubit gate fidelities at 99.921% on devices pulled from a 300mm industrial wafer line rather than hand-picked laboratory parts (Nature, 2025). The company’s NVIDIA collaboration is designed to automate device tuning and compress experiment cycles, folding quantum bring-up into the accelerated-computing stack instead of treating it as a separate lab craft. The framing is careful, and worth respecting. Diraq is not claiming the lead. It is claiming that the lead, measured the popular way, is being measured on the wrong axis.
Market Analysis
Here the picture complicates in Diraq's favor and against its framing at once. The manufacturability thesis is not a fringe position; it is the direction the leaders are already building toward, and none more literally than IBM. Alongside its 300mm processor fabrication, IBM is standing up Anderon, a standalone, pure-play quantum foundry meant to manufacture wafers for the wider industry, its bid to become a kind of TSMC for quantum. That is the incumbent embracing Diraq's core premise, that scaling is an engineering and packaging problem, while declining to concede that only silicon spin can solve it. Paired with its qLDPC error-correction work and real-time decoding on classical silicon, IBM's path to the 2029 Starling system looks like manufacturing discipline applied to superconducting hardware. That is the what's-next worth watching.
Policy is hedging the same question, funding two foundries and seven hardware makers across superconducting, trapped-ion, photonic, topological, and silicon-spin designs, a bet spread over modalities rather than on silicon alone. Peers sharpen it further. Quantinuum's Helios, validated with Sandia this quarter, posted two-qubit fidelity near 99.9 percent, a useful check on the fidelity-is-solved claim: solved enough to run error correction, less so as a competitive axis, where real headroom remains. And the capital keeps arriving; analysts tracking the internal quantum market at $60 billion to $100 billion by 2035. What that money buys is a shift we recognize from years spent buying enterprise iron, when proprietary, hand-tuned machines gave way to volume parts once the economics tilted. Diraq is betting silicon tilts them first. The leaders are betting they can industrialize their own modalities before it does.
Looking Ahead
The key trend we'll be monitoring is convergence, not combat. Over the next several quarters, the signal that matters will not be another qubit-count headline but evidence that manufacturability translates into logical qubits at workload scale. For Diraq, that means the GlobalFoundries relationship yielding devices at volume with the fidelity its imec results suggest, and its 2029 milestone holding. For IBM, it means Kookaburra and Cockatoo showing that modular error correction behaves as designed on the path to Starling, now with Anderon underneath as the manufacturing base. The near-term returns, though, are likely to arrive through hybrid quantum-classical HPC, where quantum processors run beside GPUs and CPUs rather than replacing them, which makes the "sits next to classical compute" posture Diraq and IBM share less a design preference than a commercial requirement. And watch cost. If silicon spin nears its sub-dollar-per-qubit target while exotic approaches stay capital-heavy, boring wins stops being a manifesto and becomes a procurement argument.
Stephen Sopko | Analyst-in-Residence – Semiconductors & Deep Tech
Stephen Sopko is an Analyst-in-Residence specializing in semiconductors and the deep technologies powering today’s innovation ecosystem. With decades of executive experience spanning Fortune 100, government, and startups, he provides actionable insights by connecting market trends and cutting-edge technologies to business outcomes.
Stephen’s expertise in analyzing the entire buyer’s journey, from technology acquisition to implementation, was refined during his tenure as co-founder and COO of Palisade Compliance, where he helped Fortune 500 clients optimize technology investments. His ability to identify opportunities at the intersection of semiconductors, emerging technologies, and enterprise needs makes him a sought-after advisor to stakeholders navigating complex decisions.