Draft, not yet released. In stewardship review. Not indexed until the first approved release.

Risks and capability

Four HPC risk objects open the portfolio. Order carries no rank. Each is read across driver, exposure, vulnerability, capacity, cascade and consequence.

5 min read0 figures1 table18 notesNR-XTECH-HPC-GLOBAL-2026-BASELINE

Contents
  1. ·Edition home
  2. ·Executive summary
  3. 1HPC is power, cooling and policy before it is speed
  4. 2Key findings
  5. 3Systems and sectors at baseline
  6. 4Risks and capability
    1. 4.1 Risk objects
    2. 4.2 Capability ladder
  7. 5Power and siting limits meet fast-growing demand
  8. 6Readings from the other desks
  9. 7The 2027 agenda
  10. 8Systems in depth
  11. AMethodology
  12. BData appendix
  13. CReferences and record
  14. ·Complete edition, for print

4.1Risk objects

Each risk is read across the chain from hazard to consequence, with its status and evidence-quality level.

  1. InferredEQL 3RSK-XT07-ENE-001

    Power and siting limits for large systems

    Driver
    tens of megawatts per system and growing data centre load112
    Exposure
    grid connections; host regions; research sites
    Vulnerability
    single-site dependence; long lead times for power
    Capacity
    efficient designs and heat reuse39
    Cascade
    ENE → INF
    Horizon
    1 to 5 years

  2. InferredEQL 3RSK-XT07-GOV-002

    Fragmented controls and hidden capacity

    Driver
    licence rules and guidance; systems that do not submit to lists1330
    Exposure
    research collaboration; supply chains
    Vulnerability
    uneven access to chips and skills
    Capacity
    public lists, standards forums and national plans1151623
    Cascade
    GOV → INF, FIN
    Horizon
    1 to 5 years

  3. Expert inputEQL 2RSK-XT07-INF-003

    Benchmarks read alone

    Driver
    different benchmarks give different leaders45
    Exposure
    planning and public debate
    Vulnerability
    single-number comparisons
    Capacity
    benchmark tables read together5
    Cascade
    INF → FIN
    Horizon
    1 to 5 years

  4. InferredEQL 2RSK-XT07-FIN-004

    Dependence on a few public programmes

    Driver
    large public budgets for a few systems171143
    Exposure
    national laboratories; software teams
    Vulnerability
    programme end dates; budget cycles
    Capacity
    continuing funding lines and software stacks17
    Cascade
    FIN → INF
    Horizon
    5 to 10 years

4.2Capability ladder

Existence is not readiness. Each of the fifteen capability functions is placed on the ladder by the strongest public evidence for HPC at the global level.

Table 4.1Capabilities on the ladder of evidence: listed, exists, demonstrated, tested
FunctionRungLevelStrongest public evidence at baseline
Prevention and mitigationLISTEDno common facility energy and water series found
Monitoring and observationEXISTSpublic lists of systems and efficiency13
Early warningLISTEDno consolidated watch on power and siting limits
Preparedness and planningEXISTSnational plans and roadmaps142326
Redundancy and reservesLISTEDspare capacity across sites is not documented
ResponseLISTEDno tested response to a long outage of a large system
Continuity of critical servicesLISTEDcontinuity plans of large systems not bound
RecoveryLISTEDnot applicable at baseline; no event to recover from
AdaptationEXISTSwarm-water cooling and heat reuse at one system9
Social protectionLISTEDnot applicable at baseline
Finance and insuranceEXISTSpublic programme funding171143
Institutional coordinationEXISTSEuroHPC, DOE programmes and standards forums11261516
Technical infrastructureEXISTSexascale systems on three continents1
WorkforceLISTEDskills gaps not bound; no figure used
Research capacityADEQUATEpublic systems, lists and shared-access programmes12122
A capability listed in policy is not a tested one; each rung is the strongest public evidence at baseline.

Notes

Numbers are the edition’s reference numbers, so a note reads the same in every chapter. The full list, with every source and its licence, is in Annex C. Annex C, references

  1. 1TOP500 (ed.), TOP500 list, June 2026, 67th edition (list page; top systems, Rmax, Rpeak, cores and power). TOP500 project, Hamburg. Observation year 2026. Accessed 2026-11-03.back
  2. 12US Department of Energy (ed.), DOE releases new report evaluating increase in electricity demand from data centres (news item, 20 December 2024; 176 TWh in 2023, 4.4% of US electricity; 6.7% to 12% projected for 2028). US Department of Energy, Washington. Observation year 2023. Accessed 2026-11-03.back
  3. 3TOP500 (ed.), Green500 list, June 2026 (list page; top five by gigaflops per watt). TOP500 project, Hamburg. Observation year 2026. Accessed 2026-11-03.back
  4. 9EuroHPC Joint Undertaking (ed.), JUPITER launching Europe’s exascale era (news item, 5 September 2025; about 24,000 GH200 superchips; warm-water cooling; funding shares; power and budget not stated). EuroHPC JU, Luxembourg. Observation year 2025. Accessed 2026-11-03.back
  5. 13US Department of Commerce Bureau of Industry and Security (ed.), BIS homepage (notice of 16 January 2025 on advanced computing integrated circuits; extension of 7 April 2026 to 31 December 2026; guidance of May 2026 on Country Group D:5 and Macau; full texts not opened). BIS, Washington. Observation year 2026. Accessed 2026-11-03.back
  6. 30UNVERIFIED, not opened: HPCwire (ed.), Surprise: Chinese LineShine takes number 1 on TOP500 (23 June 2026), and press coverage that China had stopped submitting to the list in 2023, taken from search-result summaries; the HPCwire page returned an access error. Observation year 2026. Accessed 2026-11-03.back
  7. 15OpenMP Architecture Review Board (ed.), OpenMP specifications (index page; version 6.0 of November 2024; technical report 15 as a public comment draft of 6.1, July 2026). OpenMP ARB, Urbana. Observation year 2026. Accessed 2026-11-03.back
  8. 16MPI Forum (ed.), MPI documents (index page; MPI 5.0 voted through 5 June 2025; MPI 4.1 of 2 November 2023). Message Passing Interface Forum. Observation year 2025. Accessed 2026-11-03.back
  9. 23UK Government (ed.), AI Opportunities Action Plan (page; expand the AI Research Resource by at least 20 times by 2030; non-AI HPC and exascale flagged as unresolved; no budget figure). Department for Science, Innovation and Technology, London. Observation year 2025. Accessed 2026-11-03.back
  10. 4TOP500 (ed.), LineShine Debuts at No. 1 as the TOP500 Enters a New Global Exascale Era (news release, 23 June 2026; LineShine, HPCG, HPL-MxP, Green500 values). TOP500 project, Hamburg. Observation year 2026. Accessed 2026-11-03.back
  11. 5TOP500 (ed.), Highlights, June 2026 (list totals, accelerator and processor shares, HPCG and HPL-MxP tables; the page gives some figures that differ from the list page). TOP500 project, Hamburg. Observation year 2026. Accessed 2026-11-03.back
  12. 17Exascale Computing Project (ed.), project site (2016 to 2024; USD 1.8 billion; about 2,800 collaborators; concluded). US Department of Energy, Oak Ridge. Observation year 2024. Accessed 2026-11-03.back
  13. 11EuroHPC Joint Undertaking (ed.), Discover EuroHPC JU (about page; budget of at least EUR 8.2 billion for 2021 to 2027; Council Regulations 2021/1173, 2024/1732 and 2026/150). EuroHPC JU, Luxembourg. Observation year 2026. Accessed 2026-11-03.back
  14. 43US Congress (ed.), CHIPS and Science Act, Public Law 117-167, Division A, section 102 (statute text, enacted 9 August 2022; CHIPS for America Fund appropriations for fiscal years 2022 to 2026 sum to USD 50 billion, of which USD 39 billion for incentives and USD 11 billion for research). Government Publishing Office, Washington. Observation year 2022. Accessed 2026-11-03.back
  15. 14RIKEN (ed.), FugakuNEXT project with Fujitsu and NVIDIA (news item, 22 August 2025; MONAKA-X CPU, NVIDIA GPU; more than 5 times hardware performance; up to 100 times application performance in about 40 MW; more than 600 exaflops FP8 sparse; operation around 2030). RIKEN, Wako. Observation year 2025. Accessed 2026-11-03.back
  16. 26POST-CLOSE, context only: US Department of Energy (ed.), Genesis Mission (programme page; goal of 100,000 science and technology professionals over a decade; announcements dated August to October 2026; no systems named). US Department of Energy, Washington. Observation year 2026. Accessed 2026-11-03.back
  17. 21Swiss National Supercomputing Centre (ed.), Alps (system page; 2,688 Grace-Hopper nodes; 434.9 petaflops; inaugurated September 2024). CSCS, Lugano. Observation year 2024. Accessed 2026-11-03.back
  18. 22National Artificial Intelligence Research Resource Pilot (ed.), NAIRR pilot site (more than 950 projects; 14 federal agencies; 26 non-governmental partners; Operations Center award of 1 September 2026). US National Science Foundation, Alexandria. Observation year 2026. Accessed 2026-11-03.back

Cite this chapter

Valadbaygi, S. (Ed.). (2026). Chapter 4: Risks and capability. In Exponential Technology Report 2026: HPC and Advanced Computing (Baseline Edition 2026, Report No. NR-XTECH-HPC-GLOBAL-2026-BASELINE) [Draft under review]. The Global Centre for Risk and Innovation (GCRI).

Every citation form of the whole report, BibTeX and RIS included, is on the edition home. Cite this report

ReportsContribute

Send evidence or a correction on HPC and Advanced Computing.

Reports is the Consortium’s quarterly publication, written by six desks. This page sits with the Global Centre for Risk and Innovation (GCRI) Systems Evidence desk. What you know about this subject can become evidence, a case or a correction in the next edition, credited to you and published under CC BY 4.0 where it is your own work. Sent to the wrong desk, it is forwarded, not refused.

GCRI Systems Evidence deskExponential Technology

Never send personal data, sensitive infrastructure detail or anything you have no right to share. A contribution is reviewed before use; sending it confers no authorship or editorial rights, and sponsorship never buys a finding.

CC BY 4.0 Free to read. Text and figures: CC BY 4.0. Data: CC0 1.0. Authors keep copyright. Published by GCRI Canada.

The Global Centre for Risk and Innovation

Science · evidence · measurement

Not for profit

Our articles, maps and twins are free to read and free to republish. Readers and partners keep it that way.

Partner with us

Who funds us

GCRI is a not-for-profit institution. To ask who funds a piece of work, use the funding questions form.

Independence

Editors and the expert reviewers named on each article decide what we publish. Funders and partners do not review or approve articles before publication.

Corrections policy

Licence

Articles: CC BY 4.0. Figures, maps, twins and data: CC BY 4.0. Other uses: ask through the permissions form.

Privacy choices