GINC Critical Technology
Capability group

Artificial Intelligence

The capacity to build, apply and govern artificial intelligence: the models, their application across the economy and the state, and the safety, security and evaluation capacity that makes them usable.

Domain
Critical Technology · Hard Power
Identifier
factbase.org/cpg/5086TTJV27
Capabilities
3

Bounds. Holds AI technology as national capacity. The compute it trains on sits with Semiconductors & Advanced Computing, the data-centre estate with Strategic Infrastructure, the skills base with Human Capital, diffusion through the economy at this group's deploy layer with economic outcomes in the economic domains, and military employment with National Security, where AI registers as a means. Regulation and international rules belong to the governance domain, with the evaluation capacity that informs them held here.

Capabilities

Machine Learning & Foundation ModelsAH3D4HX51FThe capacity to train, tune and field machine learning at the frontier, foundation and general-purpose models, and the model families across language, vision and science built on them.Research680k papers +268%China 21%Patents506.7k families +420%China 75%Applied AI & Advanced AnalyticsF4GA2HBF1EThe capacity to apply AI and advanced analytics across the economy and the state, turning models and data into decisions, products and services at scale.Research45.5k papers +89%India 22%Patents916k families +207%China 79%AI Safety, Security & EvaluationT266WRYB20The capacity to evaluate, secure and align AI systems: testing and evaluation, robustness against attack and misuse, and the standing institutions that assess frontier models.Research32.2k papers +750%China 30%Patents2.3k families +167%China 45%

The capabilities compared

Papers and patent families, 2020–2024, with growth on the five years before and the leading nation in each.

CapabilityPapersGrowthResearch leaderFamiliesGrowthPatent leader
Machine Learning & Foundation Models680k+268%China 21%506.7k+420%China 75%
Applied AI & Advanced Analytics45.5k+89%India 22%916k+207%China 79%
AI Safety, Security & Evaluation32.2k+750%China 30%2.3k+167%China 45%

Research across the group

Papers by year for each capability, and the nations' shares of the group's papers, 2020–2024. A capability that shares a v0.0 topic with another counts it in both.

  1. 1China15.4%116,847 papers
  2. 2United States10.3%78,194 papers
  3. 3India9.9%75,023 papers
  4. 4United Kingdom3.0%22,674 papers
  5. 5Germany2.2%16,300 papers
  6. 6Canada1.7%13,233 papers
  7. 7Indonesia1.6%12,177 papers
  8. 8Australia1.6%11,841 papers
  9. 9South Korea1.5%11,306 papers
  10. 10Italy1.4%10,676 papers

Patents across the group

Patent families by priority year for each capability, and the origin nations' shares of the group's families, 2020–2024. A family in two capabilities' classes counts in both.

  1. 1China77.4%1,103,488 families
  2. 2United States7.6%107,795 families
  3. 3South Korea7.0%99,713 families
  4. 4Japan3.1%43,936 families
  5. 5Germany1.2%16,961 families
  6. 6Taiwan0.5%7,808 families
  7. 7France0.4%5,116 families
  8. 8Canada0.3%4,498 families
  9. 9India0.3%3,929 families
  10. 10United Kingdom0.3%3,885 families

Companies

107 companies in Artificial Intelligence, across 8 regions. The leading 12:

  1. 1MicrosoftRuns Azure AI Foundry, Microsoft 365 Copilot, Fabric and Power BI, the largest enterprise channel for deploying AI and analytics, and is a top international AI patent filer.Software & services
  2. 2OpenAIDeveloper of the GPT and o-series frontier models and ChatGPT, the most widely used general-purpose foundation model product.Research
  3. 3Alphabet (Google DeepMind)Google DeepMind builds the Gemini frontier model family and scientific models such as AlphaFold, trained on Google's own TPU infrastructure.Research
  4. 4AnthropicFrontier lab built around alignment and interpretability research; it runs a Responsible Scaling Policy with capability-threshold evaluations and open-sourced evaluation tooling.Research
  5. 5Amazon (AWS)AWS offers SageMaker, Bedrock and Redshift, the most widely used cloud platform for building and running applied AI and analytics workloads.Software & services
  6. 6Meta PlatformsReleases the open-weight Llama model family and runs FAIR and Meta Superintelligence Labs, plus PyTorch, the dominant ML framework.Research
  7. 7Palantir TechnologiesIts Foundry, Gotham and AIP platforms put data integration and AI decision-making into government, defence and industry operations.Software & services
  8. 8IBMSells the watsonx AI and data platform and its consulting arm runs large enterprise AI and analytics deployments.Integration
  9. 9NVIDIASupplies most of the GPUs and the CUDA software stack used to train foundation models worldwide, and publishes its own Nemotron models.Equipment & materials
  10. 10Scale AIIts SEAL research lab runs leaderboards and red-teaming and evaluation services for frontier labs and US government (Meta took a 49% stake in 2025).Software & services
  11. 11AccentureThe largest IT services and consulting firm by revenue, with tens of thousands of data and AI practitioners deploying AI for enterprises and governments.Integration
  12. 12DeepSeekDeveloper of the open-weight DeepSeek-V3 and R1 models, which reached near-frontier performance at low training cost; backed by the hedge fund High-Flyer.Research
All 107 in the Companies table →