# Does the U.S. Need a National Quantum Facility — or Proof First?

The Department of Energy's Scientific Computing Advisory Committee (SCAC) Quantum Subcommittee has drawn a hard line: a proposed national quantum computing user facility will not be built on hardware benchmarks alone. According to the panel's new roadmap, released September 19, 2026, access to federal facility-class infrastructure depends on quantum systems first hitting specific, independently validated scientific milestones — including roughly **50 to 100 [logical qubits](https://quantumintel.tech/glossary/logical-qubit)**, between 10,000 and 100,000 hard logical operations, and an end-to-end scientific calculation returned within 24 hours.

That framing — scientific utility before shovels in the ground — is a significant policy signal for hardware vendors and national labs alike. The SCAC subcommittee's three-phase roadmap structures the path as: grand-challenge demonstrations from 2026 through 2028, a conditional decision on a DOE Quantum Computing User Facility, and eventual integration with DOE supercomputers, AI systems, scientific instruments, and experimental networks. None of phase two is guaranteed. The committee was explicit: sufficient scientific utility, technical maturity, and user demand must all be present before a facility decision is made.

For the quantum industry, this is both an opportunity and a gate.

---

## What the SCAC Roadmap Actually Proposes

The report originates from the DOE Office of Science's Scientific Computing Advisory Committee (SCAC) Quantum Subcommittee, whose membership spans national laboratories, academia, medicine, and industry. The committee's mandate was to develop a roadmap toward scientifically useful, [fault-tolerant quantum computing](https://quantumintel.tech/glossary/fault-tolerant-quantum-computing) by 2028 and a longer-term vision for a DOE user facility.

Its process included stakeholder interviews, written input, and a public town hall — signaling that the panel was not operating in a vacuum. The resulting three-phase structure is:

**Phase 1 (2026–2028): Quantum Grand Challenges**
Systems must demonstrate independently validated scientific use cases. The illustrative 2028 milestones are specific:
- Approximately 50 to 100 logical qubits
- Between 10,000 and 100,000 hard logical operations
- An end-to-end scientific calculation completed and returned within 24 hours
- Independent validation against experiments, classical-computing limits, or predictive scientific value

**Phase 2: DOE Quantum Computing User Facility (conditional)**
The facility would combine cloud access with laboratory-based systems connected to supercomputers, AI systems, scientific instruments, and experimental networks. It is explicitly contingent on Phase 1 outcomes. The committee recommends DOE begin planning now so a decision can be made promptly when evidence is available — but planning is not commitment.

**Phase 3 (2030+): Full Integration**
The longer-term vision targets 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 Scientific Utility Standard and Why It Matters

The committee's core finding deserves careful reading: *"Progress should be measured by the ability to solve compelling scientific problems, not by hardware metrics alone."*

This is a deliberate departure from the qubit-count arms race that has dominated vendor marketing for the better part of a decade. The subcommittee explicitly calls for scientific grand challenges to define requirements for algorithms, software, hardware, AI, and systems engineering through "continuous co-design" — rather than adapting applications to existing hardware capabilities.

That co-design language is important. It positions DOE-funded science as the specification engine for hardware development, not as an afterthought. For vendors pitching national lab contracts, it means that demonstrating high physical qubit counts or impressive [gate fidelity](https://quantumintel.tech/glossary/gate-fidelity) numbers in isolation will not clear the bar. What clears the bar is a complete calculation — one that produces a result a domain scientist can evaluate against experiments or classical limits.

The report also enumerates the enabling stack required alongside logical qubit counts: real-time error decoding, control electronics, cryogenics, packaging, calibration, software, compilers, verification, automated operations, manufacturability, and uptime. This is a systems-engineering checklist, not a physics checklist — and that distinction matters for investors evaluating which layer of the quantum stack DOE is most likely to reward.

---

## Near-Term Scientific Milestones by Domain

The subcommittee's proposed grand-challenge targets are domain-specific, which is unusual for a government advisory document and suggests serious engagement with scientific communities. Key examples from the report:

**Chemistry and Biology:** Chemically accurate predictions for selected protein-ligand or enzyme active sites by 2028, with earlier steps including molecular fragment calculations and photoreactive molecule simulations.

**Chemical Manufacturing and Catalysis:** Simulations of strongly correlated bonds and small catalytic centers; a 2028 demonstration resolving a key industrial catalyst step with chemically meaningful accuracy.

**Correlated Materials:** A quantum-validated model or prediction for a correlated material.

**Fusion Energy:** A validated prediction relevant to fusion design.

**Nuclear and Particle Physics:** A nuclear response calculation for neutrino experiments; first phenomenological inputs to particle-physics event generators.

**Sensing:** A demonstration beyond the [coherence time](https://quantumintel.tech/glossary/coherence-time) limit with a published sensitivity gain on a scientific measurement.

Each of these is a verifiable claim — not a benchmark run, but a published scientific result with an independent validation pathway. That is a substantially higher bar than most current NISQ demonstrations.

---

## Industry Implications: A Gate, Not Just a Goal

The "hardware-agnostic" framing in the report — the proposed facility would preserve competition among multiple quantum hardware approaches — suggests DOE is not planning to pick a single modality. Transmon, trapped ion, neutral atom, photonic, and topological approaches all remain in contention. But the 2026–2028 grand-challenge phase will produce data that inevitably advantages some platforms over others.

Vendors and national labs that can credibly execute a full scientific calculation — with independent validation — before 2028 will have a structural advantage in shaping the facility's eventual design. Those that cannot will find themselves outside the specification process.

For venture-backed hardware companies, this policy document functions as a forcing function. The 50–100 logical qubit milestone with tens of thousands of hard logical operations is not a soft target; it is the stated threshold for federal facility consideration. Companies that have publicly claimed near-term logical qubit roadmaps now have a government-defined benchmark against which those claims will be evaluated.

The conditional structure also protects DOE from committing prematurely. The committee's recommendation that DOE begin planning while grand-challenge work proceeds is politically astute: it keeps the facility option open without authorizing spending before scientific evidence exists. For budget hawks and program managers alike, that is a defensible posture.

---

## Key Takeaways

- The DOE SCAC Quantum Subcommittee has proposed a three-phase roadmap: Quantum Grand Challenges (2026–2028), a conditional national user facility, and full integration with DOE supercomputers and instruments.
- The 2028 scientific utility bar includes approximately 50–100 logical qubits, 10,000–100,000 hard logical operations, and end-to-end calculations validated independently within 24 hours.
- The long-term (2030+) facility target scales to roughly 1,000–10,000 logical qubits and billions to tens of billions of hard logical operations with production reliability.
- Scientific results — not qubit counts — are the mandated progress metric; the committee explicitly rejected hardware-metric-only benchmarking.
- The facility would remain hardware-platform-agnostic, preserving competition across modalities.
- Near-term grand challenges span chemistry, catalysis, correlated materials, fusion, nuclear and particle physics, and sensing — each requiring independent experimental or classical validation.
- DOE is instructed to begin facility planning now so a rapid decision can be made once Phase 1 evidence is in hand.

---

## Frequently Asked Questions

**What is the DOE SCAC Quantum Subcommittee proposing?**
The subcommittee, part of the DOE Office of Science's Scientific Computing Advisory Committee, has proposed a three-phase roadmap for fault-tolerant quantum computing. Phase 1 runs 2026–2028 and requires independently validated scientific demonstrations. Phase 2 is a conditional national quantum computing user facility. Phase 3 is full integration with DOE supercomputers, AI, and experimental networks.

**What are the specific 2028 quantum milestones the DOE panel set?**
The report specifies illustrative 2028 milestones of approximately 50 to 100 logical qubits, between 10,000 and 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.

**Why is DOE requiring scientific validation rather than hardware benchmarks?**
The committee found broad consensus that the ability to solve compelling scientific problems — not hardware metrics alone — should define progress. This reflects the view that qubit counts and gate fidelity figures do not guarantee that a system can complete a useful scientific calculation reliably and repeatedly.

**What does the long-term DOE quantum facility vision look like?**
The 2030-plus vision targets 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 computing reach, integrated with DOE supercomputers, AI systems, scientific instruments, and networks.

**Which quantum hardware platforms will be eligible for the DOE facility?**
The report explicitly states the proposed facility would preserve competition among multiple quantum hardware approaches, meaning no single modality — transmon, trapped ion, neutral atom, photonic, or otherwise — is pre-selected. The 2026–2028 grand-challenge demonstrations will produce the evidence that informs any eventual hardware decisions.