| Entity | Atom Computing; Diraq; IBM; IonQ |
|---|---|
| Counterparty | DARPA (Quantum Benchmarking Initiative) |
| Date | 7 Oct 2026 |
| Region | U.S. & Allies |
On 7 October 2026, DARPA advanced Atom Computing, Diraq, IBM, and IonQ to Stage C, the final stage of the Quantum Benchmarking Initiative. Stage C is where QBI’s government evaluation team works alongside the organizations to test whether their proposed utility-scale systems can be constructed as designed and operated as intended.
The four join Microsoft and PsiQuantum, already in Stage C from the Underexplored Systems for Utility-Scale Quantum Computing (US2QC) precursor. QBI’s stated test is whether any approach can reach utility-scale operation — computational value exceeding its cost — by 2033.
Who this affects
- For exporters: Not specified in the record.
- For investors: The release states QBI is not a competition and is not intended to pick winners and losers; Atom Computing, Diraq, IBM, and IonQ enter Stage C alongside Microsoft and PsiQuantum.
- For counsel/compliance: Stage C follows a yearlong Stage B review of research and development plans, technical risks, mitigation strategies, and prototypes. No penalty, order, or case number is in the record. The next documented step is government-side scrutiny of hardware and systems against the performers’ models and roadmaps.
- For policymakers: QBI, launched in mid-2024, is framed as an assessment of whether different commercial approaches offer credible paths to utility-scale operation by 2033. Managing Director Micah Stoutimore said DARPA increasingly expects that someone will build a utility-scale quantum computer by 2033, and that which team or technical approach will get there is not yet known.
The read
The release keeps four qubit approaches in the final stage at once: scalable arrays of neutral atoms (Atom Computing, Boulder), silicon CMOS spin qubits (Diraq, Sydney), modular superconducting processors (IBM, Yorktown Heights), and trapped-ion systems (IonQ, College Park), plus Microsoft and PsiQuantum from US2QC. The documented shift is from Stage B plan scrutiny to Stage C testing of constructability and intended operation. The release states the initiative does not select winners, and that further Stage A selections and Stage C promotions are anticipated. The read beyond that pattern: [needs analyst input]
What was the action?
DARPA selected Atom Computing, Diraq, IBM, and IonQ for Stage C of QBI after the yearlong Stage B examination of their development plans, risks, mitigations, and prototypes. They join Microsoft and PsiQuantum, which reached Stage C through US2QC, a pilot precursor to QBI.
What’s the penalty / scope, and how is it set?
No penalty is stated. Stage C places QBI’s government evaluation team alongside the organizations to test whether the proposed utility-scale systems can be constructed as designed and operated as intended. Stoutimore described Stage C as moving from plans to proof, with the team scrutinizing hardware and systems against the models and roadmaps. Utility-scale is defined in the release as computational value exceeding cost, with 2033 as the year QBI is using for that question.
Who else is exposed?
Microsoft and PsiQuantum are already in Stage C via US2QC. Since mid-2024, QBI has evaluated over 20 companies. Quandela joined Stage A in May 2026. The release says a solicitation issued earlier this year invited new Stage A entrants, that additional selections from those proposals are anticipated, and that further promotions from Stages A and B — including more companies into Stage C — are expected. Approaches named for this advancement are neutral-atom arrays, silicon CMOS spin qubits, modular superconducting processors, and trapped ions.
What happens next
The release anticipates additional Stage A selections from this year’s solicitation and further promotions, including more companies advancing to Stage C. In Stage C, the government team tests whether the systems perform as the models and roadmaps say they should. The open question stated in the release is which team or teams, and which approach or approaches, reach utility-scale operation by 2033.