v1.0 · published 9 August 2026 · CC BY 4.0
Download the Monitor (PDF) · SFQ Monitor home · doi:10.5281/zenodo.21860768
Single-flux-quantum logic — RSFQ, ERSFQ/eSFQ, RQL, AQFP and DSFQ, treated as one field — is a genuine outlier. It holds the outright digital speed record, a 770 GHz toggle flip-flop that no technology has beaten, and it switches at roughly 2×10⁻¹⁹ J, about 1,000× below CMOS before the cryogenic cooling tax and a conditional 10–30× after it.
It is also, stated plainly: not dense — some 10³–10⁵× behind CMOS; memory-poor, with a 4 K RAM record of 64 kb unbeaten since 2013; and tool-poor, with no commercial EDA flow.
Why this warrants its own monitor
SFQ sits on four critical paths at once.
It is an enabler and a multiplier for superconducting quantum computing. Room-temperature control tops out at roughly a thousand lines into a cryostat; at 10⁵–10⁶ qubits, control electronics has to move inside it. SFQ is the only logic family that has demonstrated qubit-control waveform synthesis at millikelvin — up to 99.9% single-qubit fidelity in 2026 — and demonstrated digital demultiplexing that breaks one-line-per-qubit scaling. The Monitor’s verdict is deliberately bounded: physics-favored, not physics-mandated.
It is the bridge from quantum computing to HPC. Control and readout are where a quantum processor meets the classical machine around it. SFQ operates on the cold side of that boundary, which makes it a candidate substrate for the QC→HPC interface rather than an accessory bolted onto it.
It is one of the plausible futures of HPC itself. Wall-plug energy per unit of work is now the binding constraint on large-scale computing, and SFQ’s projected system figures sit well below current datacenter accelerators even after the cooling tax is paid. Those are vendor projections rather than silicon — which is precisely why this Monitor tiers them instead of asserting them.
It is very nearly the only route to computing at hundreds of gigahertz. The outright digital speed record in any technology is an SFQ flip-flop, and Josephson-junction physics supports switching approaching a terahertz. Complex SFQ circuits run at tens of gigahertz today; no other digital family is credibly in that conversation.
What the Monitor contains
- a subsystem scoreboard where every load-bearing figure carries a confidence tier (T1 demonstrated-primary through T4 modeled-estimate) and a per-subsystem readiness level, never blended;
- a 2026 ecosystem census — seven disclosed fabrication lines, three disclosed SFQ-first venture companies totalling ~$110M, and the $1.8B IonQ–SkyWater acquisition that closed on 31 July 2026;
- market scenarios to 2035 — bear $0.5–1.5B, base $3–8B, bull $20–35B, scenario-weighted mean ≈$6.5B.
Derived figures recompute from stated inputs in a companion workbook, claims that failed primary-source tracing were excluded, and corrections ship under a logged version number.
Cite as
Neeman, R., “SFQ Technology Monitor 2026,” v1.0, Qodeh, 2026. doi:10.5281/zenodo.21860768 (all versions: 10.5281/zenodo.21860767).
Factual corrections are welcome: raveh.neeman@qodeh.com