Entrants get $250K upfront to deliver a demo within a year, which sounds feasible given quantum computing's track record
The US Department of Energy hopes to spur creation of something no one has yet realized: a fault-tolerant, scientifically relevant quantum computer, and to do it within a couple of years with relatively little federal funding upfront.
The DoE announced its Quantum Genesis Q Competition, an initiative that it said in a request for applications on the program’s opportunity site aims to produce a first-generation fault tolerant, scientifically relevant quantum computer (SRQC) by 2028.
If that sounds like a lofty goal in and of itself, that goal only becomes loftier when you realize what sort of research grants the DoE intends to award. The department anticipates up to $215 million in funding for the program, subject to congressional appropriations, with the money tied to milestones rather than handed out upfront.
Up to ten awardees will be selected, with each eligible for up to $1.5 million during the first phase of the competition. All but $250,000 of that will be held back until awardees deliver a validated prototype about a year into the contest, meaning those whose verification and validation plans are approved by DoE will have just $250,000 in milestone funding upfront before delivering the prototype.
Those who succeed in delivering a validated prototype will be eligible to move into the next phase, where participants can share a $100 million incentive pool in their bid to build a first-generation SRQC by fall 2028. DoE’s illustrative benchmark for the prototype calls for at least 10 logical qubits (quantum bits) capable of performing 10³ hard operations.
For the first-generation SRQC, DoE’s illustrative benchmark rises to at least 100 logical qubits capable of performing 10⁵ hard operations with fault tolerance.
The DoE is asking for a lot, and it’s not giving the private entities it’s hoping to entice much to accomplish its requests.
For those thinking that quantum systems with bigger raw qubit counts have already been built, that’s correct. Caltech researchers last year demonstrated an array of 6,100 qubits. Those are physical qubits, however, while DoE’s targets are measured in logical ones.
Logical qubits are a solution to the tendency for quantum computers to be error prone because qubits are fragile. Light, sound, temperature changes, and even the activity of other qubits can all cause them to break down, introducing errors into quantum computers.
Logical qubits are supposed to solve those problems by grouping a bunch of physical qubits into a single system that can correct for errors. As it stands now, researchers have demonstrated fault-tolerant techniques and systems using logical qubits, but no one has yet built a fault-tolerant machine at the scale and scientific capability DoE is targeting.
QuEra and its research collaborators have demonstrated an architecture using up to 96 logical qubits – well beyond DoE’s illustrative 10-logical-qubit prototype target and just shy of the 100 logical qubits envisioned for its first-generation SRQC. QuEra also plans to introduce its Libra system by 2028 with more than 256 logical qubits and a logical error rate of 10⁻⁶. On paper, that would put Libra in the same broad performance territory as DOE’s first-generation targets, although DoE’s "hard operations" metric and QuEra’s logical error rate aren’t directly interchangeable. With that plan already in the works, QuEra would appear to be a plausible entrant, although it hasn’t said it will toss its hat into the ring.
Either way, the DoE is asking a lot from participants in the Quantum Genesis Q contest, and whether or not the whole thing results in a working machine is anyone’s guess. Commercially viable quantum computing is one of those things that always seems a few years out, and DoE setting another deadline doesn't mean the industry will meet it. ®