DARPA moves Atom Computing, Diraq, IonQ and IBM into Stage C of its Quantum Benchmarking Initiative

[BrokenMitchell27]

DARPA has picked four companies to go through to Stage C of its Quantum Benchmarking Initiative, according to Interesting Engineering. Atom Computing, Diraq, IonQ and IBM are the names, and each one uses a different kind of qubit. Stage C is the final part of the programme, where a government team tests each fault tolerant design through physical testing and system measurements. The question QBI is trying to answer is whether a utility scale quantum computer, one whose value is greater than the cost of building and running it, can exist by 2033

The spread of approaches is the interesting part. Atom Computing uses neutral atoms held in optical tweezers and already has more than 1,200 qubits. Diraq uses silicon spin qubits and will validate a 300 mm chip that has already come out of a foundry, with a target of over two million physical qubits on one die by 2031. IonQ uses trapped ions and recently launched Superion, a 256 qubit system that fits in a rack, while IBM sticks with its superconducting transmon qubits

Microsoft and PsiQuantum were already at the equivalent stage through the earlier US2QC pilot, so this widens the pool rather than replacing anyone. QBI managing director Micah Stoutimore was candid about the uncertainty. He said nobody yet knows which team or which technical approach will get there, but added that DARPA increasingly expects someone to build a utility scale machine by 2033. That is a stronger statement than you usually hear from a government programme

The company releases fill in some of the money. IonQ says its Stage C agreement is worth up to $300 million and runs to 2029, and Atom Computing says it also has potential funding of up to $300 million, with Microsoft providing error correction support. IBM's Jay Gambetta repeated the company's line that IBM expects to deliver the first fault tolerant system in 2029. Those are bold dates, and Stage C is exactly where they get tested

I like that DARPA is keeping several horses in the race instead of betting on one. Silicon spin qubits in particular get less attention than ions or atoms, so seeing Diraq here is a nice surprise. The flip side is that six serious contenders all aiming for roughly the same window cannot all be right about their timelines. Somebody is going to end up explaining a delay

Which of the four do you think has the best shot? And do you buy the 2033 target, or is it still too optimistic?

Four quantum firms chosen to build utility-scale computers for DARPAinterestingengineering.com