The Department of Energy is asking industry to deliver a fault tolerant quantum machine by 2028, with 150 to 250 logical qubits (error corrected bundles of physical qubits), an error rate of 10^ 8 — roughly one logical mistake per hundred million operations — and circuits of…
The Department of Energy wants a scientifically useful, fault-tolerant quantum computer by 2028, and has written the exact error rate the machine must hit: 10^-8, roughly one logical mistake per hundred million operations. The target, the deadline, and the competition that follows are real. Whether the field can actually deliver is the open question.
That bar matters because quantum computers fail in a specific way. Physical qubits, the actual hardware components, are noisy. A logical qubit is a bundle of many physical qubits error-corrected together so the useful computation survives. The agency's June 24 announcement of an initiative it is calling Quantum Genesis sets the bar at 150 to 250 of these logical qubits executing circuits of roughly 10^5 hard operations, with that 10^-8 logical error rate. The agency itself describes 10^-8 as "ultra-low," and the program is asking industry to deliver the combination via a Request for Information rather than via a contract, a structure that is its own comment on how far the public field is from the target.
The program is the newest piece of a US industrial-policy bet on quantum information science, or QIS. It sits inside the broader Genesis Mission, a multi-agency effort that began life focused on artificial intelligence and supercomputing and is now being asked to integrate advanced quantum processors with the country's leading supercomputers. Quantum Genesis is the hardware-and-applications arm. Earlier executive orders, including one titled "Ushering the Next Frontier of Quantum Innovation," have already directed the federal government to accelerate American leadership in QIS. The new initiative turns that directive into a measurable test.
How the test is structured matters. The Department of Energy has set up three pieces. The first is the DOE Q Competition, a milestone-driven hardware competition aimed at the 150-to-250-logical-qubit, 10^-8-error bar by 2028. The second is targeted application research and development, a co-design effort funded for breakthroughs in chemistry, materials science, plasma physics, and high-energy physics, fields where a fault-tolerant machine would be genuinely useful, not a stunt. The third is a new National Quantum Supercomputing User Facility, a hub that will give US researchers and industry open access to multiple quantum hardware modalities. A separate effort called the Quantum Computer for Application Development and Discovery Science, or QC-ADDS, is meant to feed the same pipeline.
In May 2026 the agency issued a formal Request for Information, or RFI, asking companies what they can actually deliver against the 2028 target. That is an information-gathering step, not a contract. Specific dollar figures, award sizes, and named awardees are not in the current announcement, and any claim about who is going to win would be unverified. The Department of Energy says it will work closely with DARPA, which is already running a Quantum Benchmarking Initiative, and the two agencies have a memorandum of understanding coordinating quantum computing research, development, and engineering.
The calendar is staged, not rhetorical. The agency will finalize the exact technical specifications in September 2026. In December 2026 the Department of Energy and the Office of Management and Budget will set up the private-sector partnership framework, the terms for how manufacturing and delivery pipelines actually get built. Then in 2028 the bar is supposed to be hit or missed.
The skepticism is fair and lives in the numbers. The 10^-8 logical error rate is a demanding target, and the agency's own "ultra-low" framing acknowledges that. The 2028 timeline is aggressive for a program of this kind, even with a public-private framework and an active RFI. The agency's framing, that it is asking industry what it can deliver rather than signing for delivery, is the right one to read carefully: the program is requesting a measurement, not announcing a result.
The next thing to watch is narrow. When the September 2026 technical specifications drop, the program will have a fixed bar. When the December 2026 partnership framework drops, industry will know the commercial shape of what is being asked. From there the question is not whether the Department of Energy wants a fault-tolerant quantum computer. It clearly does. The real question is whether any of the companies now answering the RFI can clear the 10^-8 bar on 150 to 250 logical qubits before 2028 ends.