GCRI · Data and Compute
Chips and semiconductors
National computing capacity rests on processors and accelerators produced in a concentrated global value chain. Access to advanced chips shapes which institutions can run large simulations or train AI, and export controls, shortages and price swings affect public research directly. This page sets out how chip supply affects computing capacity, how dependence is measured, and what evidence serves research and digital agencies planning capacity.
Chips and semiconductors in seven parts.
- 01The questionChips and semiconductors: how the risk develops
- 02Across the domainsChips are an input to every domain's digital control
- 03IndicatorsWhat is measured, and how well
- 04SourcesWhere the evidence comes from
- 05DecisionsWho decides, and what the evidence serves
- 06Core questionsWhat supply chain experts ask
- 07MechanismsWhy the chain is fragile
Chips and semiconductors: how the risk develops.
Critical compute hardware and supply chains. The dossier below sets out how the risk develops, what is measured, where the data come from, how it is modelled and who decides.
Accelerators set capability
Modern HPC and AI rely on GPUs and other accelerators from few designers, manufactured at few advanced fabs. Supply constraints directly limit capacity.
Export controls restrict access
Controls on advanced chips and manufacturing tools limit which countries and institutions can buy leading hardware.
Prices follow demand surges
Demand for AI accelerators can outstrip supply, raising prices and lead times for public research buyers competing with large firms.
Hardware generations turn over
Each generation brings efficiency gains, so older hardware becomes costly to run per unit of work. Procurement timing matters.
Compute facility console.
The screen a team would use to follow this subject: its measures, states and decisions. Names are examples and values are illustrative.
Compute facility console
Chips and semiconductors · HPC facility · Month 3
- Utilisation, % of nodes94Active
- PUE1.18Normal
- WUE1.4l/kWhWatch
- Queue wait, p507.5h
Energy by use
- IT load
- Cooling
- Other
Allocations
| Climate ensemble | 398 | 854 | Near quota |
| Flood modelling | 426 | 487 | Within |
| Genomics pipeline | 118 | 254 | Near quota |
Chips are an input to every domain's digital control.
Supply studies are designed for the Strategic Systems series. The European Chips Act (2023) is one regional response to fabrication concentration.
- Lead time records.
- Concentration analyses with sources.
How chips and semiconductors is modelled, measured and tested.
The model, its assumptions, its data and its uncertainty, in the form a scientist or engineer would check. Structures follow published methods. The values drawn are illustrative, not findings.
Datasets and provenance
The data classes behind chips and semiconductors and what is recorded about each before it is used.
| Data class | Provenance recorded | Resolution stated | Licence checked |
|---|---|---|---|
| Satellite imagery, optical and radar | sensor, processing level, acquisition date | pixel size and revisit | per scene |
| Reanalysis and climate model output | model, experiment, member, version | grid and time step | per release |
| In-situ sensors and gauges | station, instrument, calibration record | point and sampling rate | per network |
| Population and asset exposure | method, vintage, input sources | cell size and unit | per release |
| Boundaries and reference geography | producer, edition, date | scale | per edition |
- Lineage
- W3C PROV record from source to figure
- Principles
- FAIR: findable, accessible, interoperable, reusable
- Gaps
- shown as missing, never filled silently
What is measured, and how well.
Each indicator carries its unit, its cadence, the kind of source behind it and the evidence label a value would carry. No values are shown: a programme supplies them under its own authority.
| Indicator | Unit | Cadence | Source type | Label |
|---|---|---|---|---|
| Lead time for accelerator procurement | weeks | per procurement | Procurement records | OBSERVED |
| Share of national compute on imported accelerators | percent | per assessment | Inventory of systems | ESTIMATED |
| Accelerator price per unit of performance | currency per teraflop | per procurement | Procurement and market data | ESTIMATED |
| Export control measures affecting computing | count | per change | Government notices | Published statistics |
| Hardware age of national systems | years since installation | annual | System inventories | OBSERVED |
| Diversity of processor architectures | count of architectures in national systems | per assessment | System inventories | OBSERVED |
Where the evidence comes from.
Named public datasets and institutional records, each with its licence class. Sources by agreement are used only under their terms and never republished.
| Source | What it gives | Licence class |
|---|---|---|
| TOP500 list architecture and accelerator data | Hardware used in leading systems | Public, TOP500 terms |
| UN Comtrade semiconductor trade (HS 8542) | Trade in integrated circuits | Free access with limits, UN terms |
| Government export control notices | Rules on advanced chips and tools | Public records, licence per government |
| OECD studies on semiconductor value chains | Concentration and policy analysis | Public, OECD terms |
| Public procurement notices for computing systems | Prices, lead times and suppliers | Public records, licence per portal |
Who decides, and what the evidence serves.
Data and Compute prepares evidence. The decisions below are taken by the authorities named, under their own rules, and never by the platform.
| Who decides | The question | Evidence this page serves |
|---|---|---|
| Science and digital ministries | How to secure hardware supply for national computing | Dependency profiles; export control evidence |
| HPC and AI centres | When and what hardware to buy | Procurement and cost analyses |
| Research consortia | Whether to pool procurement across institutions | Cost and lead time evidence |
Outputs this topic produces
- National compute hardware dependency profiles
- Procurement timing and cost analyses
- Architecture diversity reviews
- Evidence notes on export control effects on research
What supply chain experts ask.
- Which stages of the chain have few or single suppliers, and how long would it take to replace them?
- How exposed are fabrication clusters to earthquakes, water shortages, power disruption and geopolitical tension?
- How do export controls and industrial policies change access to advanced hardware?
- How long can research and public computing run on existing hardware if new supply stops?
- Which older or less specialised chips are essential to infrastructure and could become scarce as production shifts to newer products?
Why the chain is fragile.
Leading-edge fabrication plants cost enormous sums and take years to build, so only a few firms can operate them. They depend on highly specialised lithography and process equipment from an equally small set of suppliers, and on ultra-pure materials, gases and water. Fabrication itself needs very reliable power and large volumes of water.
Because each stage is concentrated, the chain behaves like a series of single points of failure. Demand shocks, such as rapid growth in AI accelerators, can create shortages that last until new capacity is built. Policy measures to diversify supply take years to change the physical picture, and they may also fragment markets.
How supply risk is analysed.
- 01
Supply chain mapping
Tracing suppliers stage by stage. Limits: deeper tiers are often opaque and commercially sensitive.
- 02
Concentration indicators
Measures of market share and geographic concentration. Limits: describe exposure, not likelihood of disruption.
- 03
Hazard exposure analysis
Overlaying fabrication locations with natural hazard and water stress data. Limits: site resilience measures are not public.
- 04
Materials and physics
Understanding process constraints through physics and quantum science. Limits: technology roadmaps change quickly.
How it is studied, and what the methods cannot say.
- 01High-performance, cloud and edge computingWorkload characterisation · Porting codes to new architectures · Hybrid cloud and HPC workflowsApplied here
- 02Physics and quantum scienceEnergy limits of computation · Thermal management physics · Assessing new computing paradigmsApplied here
- 03Optimisation and operations researchJob scheduling policies · Carbon-aware workload shifting · Capacity planningApplied here
- 04Energy generation, storage and grid technologyCooling technology comparison · On-site energy options · Heat reuse feasibilityApplied here
Where this page's references lead.
The bodies and documents this page is read against. Status and editions change; each link goes to the issuing body.
Questions and answers.
Why do chips matter for national research?
What is an accelerator?
Do export controls affect public research?
Can a country become self-sufficient in advanced chips?
Why does fabrication need so much water?
Does the platform assess individual companies?
Contribute evidence from this page.
Reports is the Consortium's quarterly publication. 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.
- EmailGateway GW-GCRISend it from this pageEvidence, a correction, a case, a method, a review, a translation or profile input. It arrives tagged with this page, so the editor who owns it receives it.
- ZenodoOpen depositDeposit it and submit to the communityPublish your own record on Zenodo with your ORCID iD, then submit it to the Reports community there. Curators review it; the DOI stays yours.
- ReportsGCRISee every way to contributeThe contribution types, how each is reviewed, what credit you receive, the licence and what is never accepted.
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.
Bring a method, a dataset or a question.
Information and enquiries are open today. An enquiry receives a reference and names the next step; applications, payments, compute allocations and registrations open only when their team, terms and review capacity exist.
Legal documents.
Maps, data and figures on this site are shown for context and general information. They are not advice, forecasts or warnings; the limits of the information are set out in the Disclaimers and Data Notices.