All indicators
    Technological autonomy

    Number of foundation models per capita

    Published foundation models from the region relative to population, a proxy for frontier AI capability.

    LiveHuggingFace Model HubUpdated continuous

    Where Europe stands now

    Scientific & technological outcomes

    Frontier AI models

    Advanced AI models released by a region's labs relative to population. A direct, current read on who is building the frontier.

    EU
    3.39
    US
    10.47
    China
    0.44
    D9+
    0.56
    Europe trails the US by 3.1×HuggingFace · live

    Number of foundation models per capita

    Models per million inhabitants

    0.003.006.009.0012.00USAEUD9+China

    Counted live from each region's frontier labs and shown per million inhabitants.

    The story

    Foundation models are the large AI systems that power technologies such as ChatGPT, Claude and Gemini. They form the backbone of today's most advanced AI applications and increasingly underpin innovation across industries, from healthcare and manufacturing to finance and scientific research.

    This indicator measures the number of foundation models developed relative to population size. Rather than tracking AI adoption, it measures a country's capacity to create frontier AI technologies. A higher number of foundation models suggests a stronger ecosystem of AI research, computing infrastructure and commercial innovation. Europe has a growing AI ecosystem, but the United States continues to produce a significantly larger share of the world's leading models.

    Why it matters

    Foundation models are becoming a general-purpose technology with the potential to transform almost every sector of the economy. Countries that develop these models are more likely to attract investment, talent and high-growth AI companies, while also shaping the standards and platforms that others build upon.

    Building foundation models requires a unique combination of world-class researchers, large-scale computing infrastructure, access to capital and ambitious technology companies. As AI becomes a key driver of productivity and competitiveness, the ability to develop foundation models is increasingly becoming a measure of technological leadership.

    Policy context

    Artificial intelligence has become a strategic priority for Europe's long-term competitiveness. Recent EU initiatives increasingly emphasise strengthening Europe's AI ecosystem by expanding access to computing infrastructure, supporting frontier research and accelerating the commercialisation of AI technologies.

    Developing foundation models depends on far more than scientific excellence alone. It requires an innovation ecosystem where researchers, startups and established companies can access the capital, computing resources and markets needed to build globally competitive AI businesses. Strengthening these conditions will play an important role in determining Europe's position in the next wave of technological development.

    How this number is calculated

    SourceHuggingFace Model Hub API
    DefinitionPublished models from a curated list of the region's frontier labs, per million inhabitants (World Bank population).
    EU aggregationSum of models across each region's known frontier-lab organisations, divided by the region's population. EU includes its D9+ members' labs; the D9+ line is the same labs counted again as a subset, the convention used on every chart here.
    ComparisonEurope, the United States and China, each on the same definition and the latest available year.
    UpdatesPulled automatically from the official source and refreshed every 24 hours. No manual edits.
    CaveatModel origin is tagged from the publishing organisation; counts capacity, not quality.
    NotesLive count of published models from a curated list of each region's frontier organisations on the HuggingFace Hub, normalised per million inhabitants (World Bank population). Counts capacity, not quality; closed models never published to the Hub are not counted.

    Source & cadence

    UnitModels per million inhabitants
    US sourceHuggingFace Model Hub
    FrequencyContinuous

    Related policy insights

    No insight references this indicator yet; new analyses appear here automatically.

    Browse insights