Hardware · Industry · Quantum Computing
Four more companies enter the final stage of DARPA’s Quantum Benchmarking Initiative
Atom Computing, Diraq, IBM and IonQ join Microsoft and PsiQuantum: DARPA will test their hardware to see if any can be useful and cost-effective by 2033.
Key points
- Atom Computing, Diraq, IBM and IonQ enter the final stage of DARPA's QBI, alongside Microsoft and PsiQuantum: six different technologies.
- In this stage, DARPA will test the hardware to see whether any of it can reach utility scale by 2033.
- The QBI managing director increasingly expects someone to achieve it by 2033, but does not yet know who.
DARPA, the research agency of the US Department of War, announced on 7 October that Atom Computing, Diraq, IBM and IonQ are moving to Stage C, the last one, of the Quantum Benchmarking Initiative (QBI). The programme aims to find out whether any quantum computing approach can reach “utility scale” by 2033, which DARPA defines as the point at which the value of what it computes exceeds its cost.
The four join Microsoft and PsiQuantum, which reached this stage from US2QC, the earlier pilot programme. That makes six different technologies:
| Company | Headquarters | Technology, according to DARPA |
|---|---|---|
| Atom Computing | Boulder (Colorado) | Scalable arrays of neutral atoms |
| Diraq | Sydney (Australia) | Spin qubits in CMOS silicon |
| IBM | Yorktown Heights (New York) | Modular superconducting processors |
| IonQ | College Park (Maryland) | Trapped ions |
From plans to proof
In Stage B, which lasted a year, DARPA examined each company’s research and development plans, technical risks and the prototypes needed to demonstrate progress. In Stage C, its evaluation team will work with them to check whether the systems they propose can be built as designed and perform as promised. “Stage C is where we begin moving from plans to proof,” said Micah Stoutimore, the QBI managing director.
DARPA stresses that it is not a competition and does not aim to pick winners. Since its launch in mid-2024, the QBI has evaluated more than 20 companies; the most recent to join, in Stage A in May 2026, was Quandela, and the agency expects to promote more companies to Stage C soon.
The companies have given some details of their own. IonQ says its Stage C agreement has a potential value of up to 300 million dollars, although funding beyond the initial allocation depends on future budgets, and that the stage runs until 2029. IBM says it expects to deliver its first fault-tolerant system in 2029.
Note · what this announcement does not say
Moving to Stage C does not certify that any of these machines will work: it is the start of verification. DARPA gives neither the length of the stage nor the total budget, and the IonQ and IBM figures come from the companies themselves. “Utility scale” does not mean the ability to break today’s cryptography either: they are different thresholds.
What changes for decision-makers
The figure that matters is the evaluator’s own expectation. “The quantum computing field is moving quickly, and we increasingly expect that someone will build a utility-scale quantum computer by 2033,” said Stoutimore, adding that what is not yet known is which team or which technical approach will get there. It comes from a programme whose job is to separate promises from facts, not from a manufacturer. For those who invest or plan, the QBI will be for the next few years one of the few independent sources on whether roadmaps are being met. None of the six companies in the final stage is European.