{"resourceId":"nyu-torch-hpe-nvidia-h200-onprem-supercomputer-20261019","versions":[{"version":"external-511f55fefb3c62970fe3ff7d2ae9225ef29144a77b5b2d0e16b84db800ee5dd8","resource":{"id":"nyu-torch-hpe-nvidia-h200-onprem-supercomputer-20261019","title":"NYU deploys Torch, an on-premises HPE Cray XD670 cluster with 500+ NVIDIA H200 GPUs: 10.79 petaflops LINPACK, fivefold its predecessor, No. 40 on Green500 (HPE-authored account)","organization":"New York University; HPE; NVIDIA","sector":"Research university (research computing)","geography":"United States (New York)","publishedAt":"19 October 2026 (HPCwire, byline \"by HPE\")","publicationDate":"2026-10-19","eventDate":null,"sourceName":"HPCwire (HPE-authored content)","sourceLabel":"\"Torch Lit: NYU Debuts Supercomputer to Power Academic Discovery\", HPCwire, by HPE","sourceUrl":"https://www.hpcwire.com/2026/10/19/torch-lit-nyu-debuts-supercomputer-to-power-academic-discovery/","evidenceClass":"vendor-claim","outcomeClass":"emerging","topics":["infrastructure","operating-model","accessibility-workforce"],"finding":"An HPE-authored HPCwire article (19 October 2026) says NYU has deployed Torch, an HPE Cray XD670 cluster with more than 500 NVIDIA H200 GPUs, measured at 10.79 petaflops on LINPACK, about five times its predecessor Greene (roughly 4 quadrillion calculations per second on 332 NVIDIA GPUs). Torch is liquid-cooled and ranked No. 40 on the Green500. The article presents on-premises capacity as an alternative to national-lab allocations and to public cloud, which it says brings recurring costs, egress fees and privacy constraints. It names AI and robotics, climate modelling and biomedical work as target uses, says students will be trained on the system, and says Torch will support the Global AI Frontier initiative. Claims of shorter time to results are not measured.","sledRelevance":"A large private research university chose owned, on-premises GPU capacity rather than cloud or national allocations, and justifies it on cost, data and queue grounds. That is a live decision for many research VPs and CIOs, and this case is a reference point, though it is the vendor's account.","evidence":"Opened the HPCwire article (byline \"by HPE\", dated 19 October 2026) and read the full text including specifications, predecessor comparison and stated uses. Did not find a separate NYU release in this run.","architectureImplications":"On-premises H200 cluster with direct liquid cooling; implies campus facility power and cooling upgrades, local scheduler and storage, and in-house operations rather than a hosted service.","governanceImplications":"Owned capacity keeps sensitive data on campus and sets local allocation policy, but the institution carries lifecycle, refresh and utilisation risk.","securityPrivacyImplications":"The article cites privacy constraints of public cloud as a reason to stay on-premises; no specific data classes or controls are described.","caveats":"Vendor-authored content with promotional framing. Benchmarks (LINPACK, Green500 rank) are standard but say nothing about research outcomes. No cost, utilisation, staffing, PUE value or allocation policy is given. The cloud cost comparison is asserted, not measured.","streamIds":["research"],"roles":{"sales":"Lead with the build-versus-rent decision. NYU chose an owned H200 cluster and frames cloud as costly and constraining for research data. Research VPs and CIOs facing the same choice need a neutral comparison, not a vendor story. A bounded engagement models total cost, utilisation and data constraints for owned, cloud and national-allocation options. Interpretation: this is HPE-authored, with no cost or utilisation data, so it cannot support claims that on-premises is cheaper or faster.","engineering":"Treat Torch as a reference design, not a template. It is an HPE Cray XD670 build with 500-plus H200 GPUs and direct liquid cooling, which implies facility power and cooling work and a local scheduler and storage tier. Pre-sales should check facility readiness, scheduler and container standards, and how workloads would burst to cloud or national systems. Interpretation; the article gives no network, storage or software stack detail.","delivery":"Frame delivery as running a shared campus GPU service. Work would cover allocation and fair-share policy, onboarding and training, utilisation reporting, and refresh planning. Central research computing owns it with facilities and the research office, and should plan for teaching use alongside research jobs. Interpretation: acceptance should rest on measured utilisation, wait times and user reach, because the source reports no operational metrics."},"roleDetails":{"sales":{"customerProblem":"Research institutions must decide whether to own GPU capacity, rent cloud capacity or rely on shared national allocations.","stakeholders":["Vice president for research","Chief information officer","Research computing director","Facilities and sustainability office"],"discoveryQuestions":["What utilisation and queue times does the current cluster achieve?","Which research data cannot leave campus?","What facility power and cooling headroom exists for liquid-cooled GPUs?"],"valueHypothesis":"A neutral total-cost comparison may help leadership choose and justify a compute strategy before committing capital.","potentialEngagement":"A four- to six-week research compute options study comparing owned, cloud and allocation-based capacity on cost, data constraints and utilisation.","unsupportedClaims":["On-premises is cheaper than cloud","Fivefold benchmark gain means fivefold faster research","Energy savings in absolute terms"]},"engineering":{"fit":"Fits universities with facility capacity and sustained GPU demand; poor fit for low or bursty demand.","architectureIntegration":"Integrated vendor cluster with direct liquid cooling, local scheduler and storage, plus defined paths to cloud or national systems for overflow.","prerequisites":["Facility power and liquid-cooling readiness","Demand forecast by department","Scheduler and container standards"],"constraints":["Capital and refresh cost carried by the institution","Facility upgrades lengthen deployment","Single-vendor integration"],"security":"Keep restricted research data on owned systems with local access controls; document which data classes may burst to cloud.","proofOfValue":["Measure queue time and utilisation before and after deployment","Benchmark three representative research workloads, not only LINPACK","Compare five-year total cost with a cloud scenario using real usage data"]},"delivery":{"work":"Stand up allocation and fair-share policy, onboarding, training, utilisation reporting and refresh planning for a shared GPU service.","dependencies":["Facilities completion","Staffing for operations and user support","Faculty governance of allocation rules"],"ownership":"Central research computing with facilities and the research office.","skillsAdoption":"Workshops and course integration for students; consulting support for research groups moving from smaller systems.","governanceCheckpoints":["Allocation policy approval","Semiannual utilisation review","Refresh decision ahead of hardware end of life"],"acceptanceCriteria":["Proposal: median queue wait below an agreed target within six months","Proposal: sustained GPU utilisation above an agreed threshold","Proposal: number of active research groups and courses using the system reported each term"],"risks":["Underutilisation of costly hardware","Demand outgrowing capacity before refresh","Operations staffing gaps"]}},"retrievedAt":"2026-10-19T21:58:30-05:00","enrichedAt":"2026-10-19T21:58:30-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Plans to train graduate and undergraduate students in data science, physics, engineering and digital humanities on the cluster; staffing for operations is not described.","procurementImplications":"Single-vendor integrated purchase (HPE with NVIDIA GPUs). Institutions comparing options should model total cost of ownership, facility work and refresh cycles against cloud and shared allocations.","operatingModelImplications":"Central research computing runs the system as a shared campus service with teaching use; capacity planning and fair-share policies become ongoing work.","sourceVerification":{"openedUrl":"https://www.hpcwire.com/2026/10/19/torch-lit-nyu-debuts-supercomputer-to-power-academic-discovery/","referenceExcerpt":"By bringing this capacity on-premises, NYU has established a localized computing environment to expand access to hardware for faculty members, researchers, and students.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}