THE QUANTUM INTELLIGENCER

IonQ, a trapped-ion quantum platform developer, has published an architectural resource estimate for executing Shor's algorithm against the secp256k1 curve, the elliptic curve that underpins Bitcoin and much of the wider blockchain infrastructure. The study, reported by the Quantum Computing Report, projects that a fault-tolerant device of 19,397 physical qubits could solve the 256-bit discrete logarithm problem in 25.7 days per attempt, with a heuristic single-run success probability of 63.3 percent. This is a fully compiled design estimate, not a working machine; no such device is claimed to exist. The arithmetic optimizations reduce the logical requirements to 1,457 logical qubits and 39 million Toffoli gates, and the authors derive success-probability bounds holding with confidence at least 1 - 2^-128. IonQ places the relevant fault-tolerant hardware on its roadmap for the 2028 timeframe and urges enterprise migration ahead of that date, observing that compromised ECDSA signatures present forward-looking exploitation risks rather than retroactive decryption threats. The company further notes that post-quantum signature standards, including ML-DSA and SLH-DSA, remain mathematically unaffected. As a publication, the study obliges no institution; it is an estimate issued by a vendor, not a standard, a regulation, or a deadline set by any authority. Between the publication and the obligation lies a gap that no quantity of arithmetic can close. The resource estimate answers a genuine question, namely what a machine would cost to build and operate. It does not answer the institutional question: who has agreed to build that machine, under whose authority it would run, or on whose books the cost of readiness now sits. IonQ's language is the language of persuasion. The company urges migration, notes risks, highlights vulnerabilities; those are the verbs of a party seeking converts, not of a party imposing terms. Urging binds no one who has not consented to be bound. The 2028 timeframe is a statement about the vendor's own hardware targets, not a migration deadline accepted by the enterprises whose certificates are at issue. A standard, a regulation, or a directive names an authority and the parties it governs. This document names neither. It is a roadmap with a single signatory, and that signatory is the author. The publication may inform a decision; it cannot make one. Whatever obligation arises from these figures will have to be assumed elsewhere, by some institution willing to put its own name beside the date. The figures that follow are set down because they are the material out of which any eventual obligation will be built. The Quantum Computing Report, under date of September 8, 2026, records that the optimised arithmetic lowers the earlier Toffoli baseline of 58 million gates, at a cost of 273 million T-states. Where sequential Toffoli gates would have required more than four hundred days, the design introduces a two-level magic state factory producing CCZ states directly in qLDPC memory, with depth-one CCZ state injection, non-overlapping cat-state measurement channels, and a pipelined logical CliNR protocol for Clifford frame clearing; the acceleration is thirty-onefold, bringing each Toffoli gate to 29.5 milliseconds. Four CCZ factories consume 1,276 physical qubits and maintain an average production time under 28 milliseconds. The compiler leaves routing overhead, Clifford operations, and hardware decoherence in the model rather than abstracting them away; it accounts for two-dimensional ion shuttling, physical qubit transport, syndrome extraction schedules, and a Swap-Loss model using Leakage and Loss Reduction Units that eliminate the need for dedicated beacon qubits. The memory blocks, designated Q102, encode 22 logical qubits into 102 data qubits, and integrated routing confines total routing overhead to under five percent of the end-to-end execution runtime. These are the quantities a purchasing office would have to price, and a standards body would have to test, before the estimate passed from announcement to instruction. The migration deadline has no signatory. The only date in the field is IonQ's own 2028 hardware target, and a vendor's roadmap is not an obligation accepted by the institutions whose certificates are at risk. No standards body has published a date by which secp256k1 signatures must be retired. No regulator has fixed an end to ECDSA use in code signing or root-of-trust hierarchies. No enterprise has declared when its own certificates will move to ML-DSA or SLH-DSA. Each of those decisions would carry authority and bind someone; none has been made. The resource estimate can be priced, tested, and argued over, but it does not by itself create a duty. Somewhere an institution with the power to set a date must eventually put its name beside one. Until then the deadline is not late; it simply does not exist. —Elias Hartwell Dispatch from The Prepared E0