.svg){kind=icon}
Insider Brief
- A DOE advisory panel proposed a 2026–2028 quantum grand-challenge program to demonstrate independently validated scientific uses of fault-tolerant quantum computing before considering a national user facility.
- The roadmap identifies near-term applications in chemistry, catalysis, materials, fusion, particle and nuclear physics, and sensing, with progress measured by scientific results rather than qubit counts alone.
- The proposed facility would combine cloud access with laboratory-based systems connected to supercomputers, AI, experiments and scientific users while preserving competition among multiple quantum hardware approaches.
For a committee advising the U.S. government on fault-tolerant quantum computing, the message appears to be money talks, but qubits still have to walk.
A proposed national quantum computing user facility would depend on quantum systems first demonstrating validated scientific value, technical maturity and demand from researchers, according to a Department of Energy advisory panel’s new roadmap.
The proposal, outlined in a report from the DOE Office of Science’s Scientific Computing Advisory Committee (SCAC) Quantum Subcommittee , calls for a three-phase effort. The first phase would establish Quantum Grand Challenges from 2026 through 2028. A second phase could create a DOE Quantum Computing User Facility if the early demonstrations show sufficient scientific utility, technical maturity and user demand. The third phase would integrate quantum computing with DOE supercomputers, artificial intelligence systems, scientific instruments, networks and experiments.
The report recommends that DOE begin planning for the facility while the grand-challenge work proceeds, so the department can make a decision when the relevant technical and scientific evidence is available.
“The Committee found broad consensus that progress should be measured by the ability to solve compelling scientific problems, not by hardware metrics alone,”the committee writes in the report. “Scientific grand challenges should define the requirements for algorithms, software, hardware, AI, and systems engineering through continuous co-design, rather than adapting applications to existing technologies. This science-driven approach will accelerate the development of meaningful applications, guide technology development, and build the broad user communities needed for future quantum computing capabilities.”
The committee was charged with developing a roadmap toward scientifically useful, fault-tolerant quantum computing by 2028 and a longer-term vision for a DOE Quantum Computing User Facility. Its members include representatives from national laboratories, academia, medicine and industry, and its process included stakeholder interviews, written input and a public town hall.
Scientific Utility Standard
The report proposes a set of illustrative milestones for systems capable of scientific demonstrations in 2028. These include about 50 to 100 logical qubits, 10,000 to 100,000 hard logical operations, an end-to-end scientific calculation returned within 24 hours and independent validation against experiments, classical-computing limits or predictive scientific value.
Logical qubits are groups of physical qubits whose errors are detected and corrected as a calculation proceeds. The panel’s focus on logical qubits and logical operations reflects the difference between demonstrating a device’s component performance and operating a system long enough to complete a useful calculation.
The report does not treat a larger logical-qubit count alone as proof that quantum computing has become useful. It calls for fault-tolerant systems that can run reliably, repeatedly and at enough throughput to support scientific campaigns. It also identifies a range of enabling requirements, including real-time error decoding, control electronics, cryogenics, packaging, calibration, software, compilers, verification, automated operations, manufacturability and uptime.
For its 2030-plus user-facility capability, the roadmap envisions roughly 1,000 to 10,000 logical qubits, billions to tens of billions of hard logical operations, production-level reliability and results that extend beyond routine classical reach.
The committee’s near-term scientific milestones are narrower. In chemistry and biology, it proposes chemically accurate predictions for selected protein-ligand or enzyme active sites. The earlier steps would include calculations involving molecular fragments, a small number of active sites and photoreactive molecules.
For chemical manufacturing and catalysis, the roadmap calls for simulations of strongly correlated bonds and small catalytic centers, followed by calculations of reaction steps involving transition metals and metalloenzymes. A 2028 demonstration could resolve a key step in an industrial catalyst with chemically meaningful accuracy.
The roadmap also identifies correlated materials, fusion, nuclear and particle physics, and sensing. Examples include a quantum-validated model or prediction for a correlated material; a validated prediction relevant to fusion design; a nuclear response calculation for neutrino experiments; and first phenomenological inputs to particle-physics event generators. In sensing, the panel proposes a demonstration beyond the coherence limit with a published sensitivity gain on a scientific measurement.
Those demonstrations would not establish that quantum computers can solve every problem in those fields. The committee describes them as benchmarks that could validate a system’s correctness and reliability, then establish whether it can produce predictive scientific results that extend beyond current computational methods.
Facility Will Depend on Results
The report places the proposed Quantum Computing User Facility after the grand-challenge phase rather than treating it as an automatic hardware acquisition.
The facility would be considered if quantum systems demonstrate scientific value, technical readiness and a sustainable user program. The committee recommends that DOE assess the timing, scale, operating model, refresh strategy, projected hardware development and prospective demand through a dedicated planning activity involving laboratories, universities, industry, the National Quantum Information Science Research Centers and prospective users.
The panel leaves open several fundamental questions including whether the DOE could eventually operate a single leadership-class facility, a distributed network of specialized resources or a combination of both. It could deploy an initial facility once useful systems become available or wait for larger-capability machines. The report does not resolve those questions because the answers depend on the progress of hardware and scientific applications.
It also calls for future cost estimates through a formal, community-informed DOE process.
The committee describes the facility as a national scientific resource rather than a replacement for commercial quantum services. Its user model would resemble DOE’s other facilities, with peer-reviewed access, resident scientific expertise, training, benchmarks and collaborative software development.
The ultimate purpose of a DOE Quantum Computing User Facility is not simply to maximize quantum hardware performance, but to maximize scientific discovery,” the committee writes. “The Committee envisions the facility as a scientific instrument enabling transformative advances across chemistry, biology, materials science, condensed matter physics, particle and nuclear physics, fusion energy sciences, optimization, and other emerging disciplines.”
The proposed facility could support researchers from DOE laboratories, universities, industry and other federal agencies. The report identifies chemistry, biology, materials science, condensed-matter physics, particle and nuclear physics, fusion, optimization and emerging fields as potential user communities.
Cloud Access and Co-Located Systems
Cloud-based quantum access would remain part of the proposed model and the report suggests that cloud services can provide broad availability, rapid experimentation and access to commercially developed hardware.
It recommends combining that access with systems located at DOE national laboratories. Co-located machines could be integrated with leadership-class high-performance computing, AI infrastructure, data systems and experimental facilities. The report also cites cryogenic infrastructure, quantum interconnects, networking, systems engineering and embedded multidisciplinary teams as capabilities that could be developed more readily in that setting.
A scientific workflow could combine an experimental measurement, a classical simulation, AI-based analysis and a quantum calculation. The report therefore treats quantum computing as a component of a wider scientific system, not as a stand-alone processor intended to displace conventional supercomputers.
The panel also says that scientists require more access than a typical remote cloud user may receive. Researchers need sufficient access to hardware architectures, control interfaces, diagnostics and system-performance information to optimize applications and validate results, according to the report. The committee acknowledges that companies must protect intellectual property, but recommends partnership frameworks that provide scientific openness without requiring companies to disclose proprietary technology.
Suggested mechanisms include embedded teams, with laboratory researchers inside hardware companies and industry employees at national laboratories; joint appointments; shared pre-competitive prototypes; and DOE-industry co-investments tied to scientific outcomes.
Platform Competition Remains Open
The report recommends against prematurely choosing one quantum hardware approach for the national effort. It identifies superconducting circuits and cavities, trapped ions, neutral atoms, photonics, spin qubits and emerging technologies as candidates for a portfolio of complementary systems.
The same approach applies to system design. The report leaves room for modular and hybrid systems, quantum-HPC integration, qudits, distributed computing and quantum networking. Different architectures may advance at different rates and may be suited to different scientific problems, according to the committee.
The report also records disagreement among stakeholders. Some participants believed scientifically useful fault-tolerant systems could be demonstrated by 2028, while others expected large-scale fault tolerance to require more time. Participants also differed over the balance between application development and foundational hardware work, the degree to which DOE should host frontier hardware and the eventual structure of a user facility.
The proposed sequence gives DOE a way to evaluate those questions through scientific demonstrations rather than vendor roadmaps alone. The national facility would follow proof that quantum systems can produce validated results for scientific users, not merely the arrival of a machine with a larger qubit count.
Underlying the report is the idea that now is the ideal time to begin the project: “The Committee believes that the rapid maturation of multiple quantum computing platforms makes this the appropriate time to broaden the national effort by engaging a much larger scientific community in defining and solving ambitious scientific grand challenges on quantum computers. Continued advances in hardware, software, algorithms, systems engineering, and AI should proceed through continuous co-design with domain scientists, ensuring that scientific applications define technology requirements while technological advances continually expand scientific opportunity.”