{"resourceId":"ctol-jupiter-simulation-no-logical-error-rate-90-percent-claim-20261019","versions":[{"version":"external-fad5153583a080104d24d0526363880b2edd4d7721f6e70fffddd419a38394df","resource":{"id":"ctol-jupiter-simulation-no-logical-error-rate-90-percent-claim-20261019","title":"CTOL analysis: NVIDIA's JUPITER run used about 35% of the machine, reports no logical error rate, and its 90% saving traces to a 2024 Google estimate in a different unit","organization":"CTOL Digital Solutions (user-submitted news analysis)","sector":"Independent technical commentary","geography":"Germany / Europe","publishedAt":"2026-10-19","publicationDate":"2026-10-19","eventDate":"2026-10-19","sourceName":"CTOL Digital Solutions","sourceLabel":"Independent analysis article (user-submitted)","sourceUrl":"https://www.ctol.digital/news/nvidia-jupiter-quantum-simulation-memory-arithmetic-90-percent-claim/","evidenceClass":"independent-reporting","outcomeClass":"cautionary","topics":["infrastructure","governance-procurement"],"finding":"CTOL argues the JUPITER result is a memory-management achievement, not a measurement of quantum error rates. Using NVIDIA's CUDA-Q technical blog, it notes the 104-qubit protocol was stored as a 45-bit encoded register (about 35.2 trillion amplitudes, 512 TiB, 64 GiB per GPU in double precision) on 8,192 of JUPITER's 23,536 Booster GPUs, about 35% of the machine; by its arithmetic one more bit would need 16,384 GPUs and two more would exceed the machine. It says the blog reports no logical error rate or fidelity, so the run cannot show whether coherent errors and leakage erode the saving. The 90% figure traces to a 2024 Google paper (Gidney, Shutty, Jones) measured in qubit rounds (space-time volume) under simpler uniform depolarizing noise, not physical-qubit counts. It credits the run as a useful reference simulator and notes cross-simulator agreement validated the software.","sledRelevance":"Useful to research institutions and public funders as a model of how to read vendor HPC and quantum claims: check the unit, the missing metric and the share of the machine consumed. The arithmetic is checkable against public figures, but the article is user-submitted, so treat it as informed commentary.","evidence":"Independent commentary built on NVIDIA's CUDA-Q blog and Jülich's published node counts; GPU projections for 46 and 47 bits are the author's arithmetic assuming double precision and ignoring communication overhead. Published under the site's news-submission rules; no author is named on the page.","architectureImplications":"State-vector simulation memory doubles per encoded bit, so scale is bounded by machine memory; approximate methods such as tensor networks are the route beyond it.","governanceImplications":"Supports requiring error metrics and peer-reviewed methods before public bodies cite vendor quantum results in funding or policy decisions.","securityPrivacyImplications":"Not applicable to this source.","caveats":"User-submitted article, author unnamed; projections are calculated, not measured. It disputes NVIDIA's framing but does not claim the run is wrong.","streamIds":["nvidia"],"roles":{"sales":"Sales framing: this analysis is a reminder that research buyers and public funders will test vendor headline numbers, and that a missing metric can undercut an otherwise impressive result. Interpretation: account teams should present NVIDIA's quantum-simulation work as a capability demonstration and avoid repeating the 90% figure as a validated result. The article is user-submitted commentary, so cite it as a caution, not a definitive rebuttal.","engineering":"Engineering framing: the useful technical points are the memory arithmetic, where each added encoded bit doubles GPU memory, and the missing logical error rate. Pre-sales engineers advising research centres should describe state-vector simulation limits honestly and point to tensor-network approximation as the path beyond machine memory. The projections assume double precision and ignore communication overhead, so treat them as bounds rather than measurements.","delivery":"Delivery framing: research-computing teams can turn this into routine review practice, asking for outcome metrics and share-of-machine figures in every large allocation proposal and in post-run reports. That helps committees compare projects fairly. Reviewers may need short training on error-correction terms to apply it consistently. Acceptance criteria below are proposals for that practice, not results reported by the source."},"roleDetails":{"sales":{"customerProblem":"Funders and research leaders must judge vendor HPC and quantum claims before committing resources.","stakeholders":["Research office","HPC allocation committee","Grant officers"],"discoveryQuestions":["What metric would show the scientific result, not just scale?","What share of the machine does the claim require?","Is there a peer-reviewed paper?"],"valueHypothesis":"Interpretation: disciplined claim review avoids over-commitment to unproven approaches.","potentialEngagement":"Claims-review checklist for research-computing allocation and funding decisions.","unsupportedClaims":["That the NVIDIA run is invalid","That the 90% saving is disproven"]},"engineering":{"fit":"Technical due-diligence reference for quantum-simulation claims.","architectureIntegration":"Distributed state-vector simulation bounded by GPU memory; tensor networks for approximation.","prerequisites":["Access to methods and metrics"],"constraints":["User-submitted source","Calculated projections"],"security":"Not applicable to this source.","proofOfValue":["Request the logical error rate for the simulated state","Check precision and memory assumptions"]},"delivery":{"work":"Embed metric and resource-share questions in allocation review.","dependencies":["Allocation policy"],"ownership":"HPC allocation committee.","skillsAdoption":"Reviewers need basic quantum-error-correction literacy.","governanceCheckpoints":["Metric stated in proposal","Post-run results report"],"acceptanceCriteria":["Proposed: every large-allocation proposal states its outcome metric","Proposed: share of machine reported"],"risks":["Reviewer capacity limits","Vendor resistance to disclosure"]}},"retrievedAt":"2026-10-20T02:54:54Z","enrichedAt":"2026-10-20T03:03:02Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Not addressed by the source.","procurementImplications":"Large-allocation justifications should state the share of a public machine consumed and the scientific metric delivered.","operatingModelImplications":"Allocation committees can ask for the outcome metric (here, logical error rate) as a condition of large runs.","sourceVerification":{"openedUrl":"https://www.ctol.digital/news/nvidia-jupiter-quantum-simulation-memory-arithmetic-90-percent-claim/","referenceExcerpt":"Yet the blog reports no logical error rate or fidelity for the simulated state.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}