After last week's CHIPS Act wave (where the federal government announced $2 billion in Letters of Intent across nine quantum companies), this week was always going to read quieter. No headline catalyst landed and nothing on the watchlist meaningfully shifted.
But underneath the quiet, the architecture race keeps moving. Three neutral-atom milestones stacked inside 48 hours. Quantinuum's pre-IPO commercial pipeline thickened. IBM put a real dollar number on its quantum commitment.
A reader asking "did I miss anything material this week?" gets a clean answer: no, but the things that moved are the things to keep watching.
Three items worth your attention, plus a couple of flags.
Neutral-atom had three milestones in 48 hours.
Last Thursday, Infleqtion (INFQ) released a package of three results at once. The first was a record in gate fidelity (an improvement in the reliability of quantum operations). The second was new software tooling for estimating computing resource requirements for a given quantum program. The third was theoretical advances in error correction.
Logical qubits (squares) follow the solution. Physical qubits (dots) don't. Source: Pasqal.
Pasqal followed on Friday with the most concrete result of the three. Quantum error correction works by combining many physical qubits into one "logical" qubit that is, in theory, more reliable than any of the physical qubits underneath it. The catch is that for years, the logical qubits haven't actually performed better than the physical qubits they were built from — the overhead of doing the error correction has outweighed the benefit. Pasqal demonstrated that its logical qubits did outperform its physical qubits, at least on one task: solving differential equations. The choice of task matters, because differential equations are a real computational problem rather than a benchmark designed to flatter the result. This is one of the first demonstrations that quantum error correction can be net beneficial rather than net costly on a problem people actually care about.
This Thursday, Infleqtion announced a new manufacturing hub in Oxford.
None of these three is decisive on its own. Infleqtion's hardware results don't yet reach the qubit count that would justify upgrading how I'd position the name. Pasqal's result is the most meaningful of the three, but the qubit count needed to call it architecture-shifting wasn't in the disclosure.
My read: strong incremental progress, not a turning point. Neutral-atom is the architecture I currently favor on the math of how cleanly error correction scales (QuEra's published 2:1 ratio of physical to logical qubits in simulation; current hardware closer to 7:1). A week with three neutral-atom milestones and zero from competing architectures reinforces that lean at the margin.
For INFQ specifically, the catalyst that would move the position from "evaluate" to action is the first quarterly earnings print as a public company, and what that print discloses about the sensing-versus-compute revenue mix. The milestone wave heading into earnings reduces the execution risk a touch; it doesn't change what to watch for.
Quantinuum's pre-IPO pipeline strengthened in two places.
Quantinuum's collaboration with BP (the energy major, formerly British Petroleum) moved from feasibility pilot to production scaling. The work uses quantum computing to improve seismic imaging: the technique oil and gas companies use to map underground reservoirs before drilling. Oil and gas exploration has clear economic value, and a paying customer like BP using the technology at production scale makes Quantinuum's reported ~$12.7 billion IPO valuation more defensible.
Separately, Quantinuum was named as a recipient of a $100 million CHIPS R&D Letter of Intent from the Department of Commerce, the same level as five other non-foundry quantum recipients. The interesting part of the LOI isn't the funding itself; Quantinuum's cash position doesn't depend on it. It's the signal: the same federal CHIPS quantum program that committed $1 billion to IBM's superconducting-qubit foundry is also funding Quantinuum's trapped-ion infrastructure. The program is architecture-agnostic, not superconducting-only. That's a useful read for how to think about future CHIPS announcements.
For HON (Honeywell, Quantinuum's parent), both items strengthen the case for what the Quantinuum stake is worth heading into IPO pricing.
IONQ uses the same trapped-ion technology as Quantinuum, so Quantinuum's progress matters for IONQ too. The read here is mixed. Each Quantinuum customer win further validates trapped-ion architecture, which is good for IONQ by association. But Quantinuum's IPO will also give IONQ a direct public competitor for the first time. Investors will start valuing the two trapped-ion names against each other, which could move IONQ's stock either way depending on where Quantinuum prices. My position on IONQ doesn't change here; what matters is where Quantinuum prices.
IBM put a number on its quantum commitment: $10 billion over five years.
IBM disclosed a $10 billion commitment to its quantum business over the next five years, on top of last week's $1 billion CHIPS Act foundry LOI. This is IBM saying how much of its own capital it plans to spend, not a structural move like spinning the quantum business out as a separate company. The way the asset sits inside IBM today doesn't change.
But the number is worth noting. The combined R&D budgets of all the public quantum pure-plays don't add up to $2 billion a year. IBM is now planning to spend more on quantum than the entire public pure-play universe combined. And it's doing this with a major milestone coming at the end of 2026: a verified demonstration of quantum computers outperforming classical ones on a useful task.
Put those two facts together and the read is straightforward: IBM is keeping the spin-out option open. The capital scales the quantum business up to standalone size. The milestone, if it hits, would peak investor interest in the asset. Together they're the two conditions IBM would want in place before spinning a business off. This isn't the spin-out itself; it's IBM giving itself the option to do one.
Other items worth flagging.
NIST advanced nine PQC digital signature candidates to Round 3. A procedural step in the standardization of post-quantum cryptography (the cryptographic primitives that will replace today's standards once quantum computers can break them). More signature schemes in the standard set means more migration work for implementation-layer vendors. Constructive context for MCHP and NET; doesn't accelerate any existing deadline.
Pending in the news, not yet ratified. EDGAR filings from Rigetti (unregistered equity sale plus an 8-K) and Xanadu (routine secondary registration) came through this week. I'm still working through the details. The dilution question for the pure-plays is a standing watch item, and any meaningful equity issuance into the post-CHIPS-Act rally would be worth paying attention to. Will flag in next week's digest if anything material surfaces.
What to watch next week.
Three things would change the read.
First, whether any of the May 21 CHIPS Act recipients moves from Letter of Intent to a final award. LOIs have historically converted at high rates, but conversion is a separate confirmation step, and the first conversion will set the pattern for the rest.
Second, IONQ or INFQ pre-announcements ahead of Q2 earnings season. The sector is well into the window where the next print starts to anchor where the stocks trade.
Third, any concrete signal on Quantinuum's IPO pricing. F-1 amendments and roadshow indicators are the earliest reads we'll get on where the trapped-ion peer comes public.
Quiet weeks aren't empty weeks. What moved this week was the architecture race and the pre-IPO setup. Those are the pieces that decide the second half of the year.
Not investment advice. For informational purposes only.
Position disclosure: The author may hold positions in companies referenced.