Belgium
Cognitive Power
Total national capacity, AI capital, compute, models & robotics. Strongest in semiconductor capability, economic mass, researchers (total).
Cognitive Coefficient
How deep infrastructure runs per person, rewards intensity, like GDP-per-capita or the HDI.
Belgium's agentic capacity
How many capable, autonomous agents this society can field, how intelligent they are, and what they are permitted to do, across citizens, business and government. An automated army of agentic problem-solvers may be the single clearest near-term lever of cognitive advantage.
CC projection to 2040, Baseline
All 8 scenarios →Monte-Carlo band (P10-P90) around the median path. Dashed = illustrative historical reconstruction.
Pillar profile
Belgium, cognitive standing in full
Belgium is the world's #23 economy by GDP ($726bn) and ranks #28/50 on absolute Cognitive Power and #20/50 on per-capita intensity. Its standout strength is AI Readiness (60.4, Strong); its trajectory is Maintaining, holding position but not compounding an advantage. It carries no single binding weakness across the headline indices.
Economic & scale context
The economic base AI capacity is built on, and how its scale interacts with the cognitive indices.
Index-by-index read
For each index: what it means when high (the value) versus low (the risk), and Belgium's own standing.
| Index | Value | Standing | What a high vs low value means, and where Belgium sits |
|---|---|---|---|
| CPI Cognitive Power Index | 33.9 | Moderate · #28/50 | Mid-pack. High would mean great-power scale: the country can shape the frontier, attract capital and talent, and set standards others follow; low would mean limited absolute weight: a price-taker that depends on capacity built elsewhere. ▲ high: great-power scale: the country can shape the frontier, attract capital and talent, and set standards others follow · ▼ low: limited absolute weight: a price-taker that depends on capacity built elsewhere |
| CC Cognitive Coefficient | 60.8 | Moderate · #20/50 | Mid-pack. High would mean deep capability per person: broad productivity gains and a well-equipped workforce; low would mean thin per-person depth: gains stay concentrated and most of the workforce is under-equipped. ▲ high: deep capability per person: broad productivity gains and a well-equipped workforce · ▼ low: thin per-person depth: gains stay concentrated and most of the workforce is under-equipped |
| ADI Agentic Deployment Index | 62.6 | Moderate · #21/50 | Mid-pack. High would mean an automated-agent workforce multiplier is already in place across society; low would mean agents are under-deployed, the clearest near-term productivity lever is being left on the table. ▲ high: an automated-agent workforce multiplier is already in place across society · ▼ low: agents are under-deployed, the clearest near-term productivity lever is being left on the table |
| CDI Cognitive Development Index | 0.605 | Moderate · #18/50 | Mid-pack. High would mean broad, balanced development with no single dimension dragging it down; low would mean imbalance or underdevelopment, at least one dimension is a binding weakness. ▲ high: broad, balanced development with no single dimension dragging it down · ▼ low: imbalance or underdevelopment, at least one dimension is a binding weakness |
| AIR AI Readiness | 60.4 | Strong · #14/50 | High here, institutions and infrastructure are ready to convert capability into real deployment. ▲ high: institutions and infrastructure are ready to convert capability into real deployment · ▼ low: governance, data or infrastructure gaps will slow how fast AI can actually be used |
| AIX Augmentation Index | 45.8 | Moderate · #21/50 | Mid-pack. High would mean AI is actively in use across the workforce, so capability is being turned into value; low would mean access without adoption, the tools exist but are under-used, so the value is unrealised. ▲ high: AI is actively in use across the workforce, so capability is being turned into value · ▼ low: access without adoption, the tools exist but are under-used, so the value is unrealised |
| CRI Cognitive Resilience Index | 52.9 | Strong · #14/50 | High here, strategic autonomy: a broad home-grown base that is hard to cut off or coerce. ▲ high: strategic autonomy: a broad home-grown base that is hard to cut off or coerce · ▼ low: dependency risk: heavy reliance on frontier infrastructure others control, a strategic vulnerability |
| CM Cognitive Momentum | 46.9 | Moderate · #20/50 | Mid-pack. High would mean investment and pipeline favour sustained improvement, today's position is likely to compound; low would mean a weak pipeline, the current position may stall or erode without new investment. ▲ high: investment and pipeline favour sustained improvement, today's position is likely to compound · ▼ low: a weak pipeline, the current position may stall or erode without new investment |
| CV Cognitive Velocity | 3.4% | Moderate · #20/50 | Mid-pack. High would mean improving quickly relative to a moving frontier; low would mean slow improvement, at risk of drifting backwards in relative terms as the frontier advances. ▲ high: improving quickly relative to a moving frontier · ▼ low: slow improvement, at risk of drifting backwards in relative terms as the frontier advances |
Trajectory & growth potential
Trajectory & growth potential
MaintainingBelgium's trajectory is Maintaining, holding position but not compounding an advantage. Risk: the frontier keeps moving, so standing still means slow relative decline. A stable base, but it needs fresh momentum to avoid drifting back.
Strengths to build on
- AI Readiness (60.4, Strong), institutions and infrastructure are ready to convert capability into real deployment.
- Cognitive Resilience Index (52.9, Strong), strategic autonomy: a broad home-grown base that is hard to cut off or coerce.
Risk factors
- Relative trajectory, Maintaining: Risk: the frontier keeps moving, so standing still means slow relative decline. A stable base, but it needs fresh momentum to avoid drifting back.
- Readiness-utilization gap, AIR 60.4 but AIX 45.8: the institutions are ready but AI is under-used, so the capacity is not yet converting to value.
Recommended priorities
- Prioritise the weakest pillar (AI Adoption) to lift the geometric-mean development index (CDI).
- Convert access into utilization: workforce AI-skilling and agentic-workflow adoption.
- Broaden the base (education + data governance) to raise resilience and reduce concentration risk.
Pillars
Nearest peers
- Austria59.8
- Australia61.9
- Norway62.2
- Ireland62.2
- United Arab Emirates57.8
Provenance & what changed · 39 indicator cells · year, source, measured vs estimated
Rank movement
- no rank movement since the previous snapshot
Applied research changes
- no applied research changes name this economy in the window
| Indicator | Value | Year | Source | Kind |
|---|---|---|---|---|
| agent_access_pct | 62.0 | 2025 | curated (agentic-deployment estimate) | estimated |
| agent_autonomy_idx | 68.0 | 2025 | curated (agentic-deployment estimate) | estimated |
| agent_capability_idx | 72.0 | 2025 | curated (agentic-deployment estimate) | estimated |
| ai_capital_b | 6.0 | 2024 | Stanford AI Index / state & infra capex (curated) | estimated |
| ai_index_score | 34 | 2024 | Tortoise Global AI Index 2024 | estimated |
| ai_private_investment_b | 1.5 | 2024 | Stanford AI Index 2025 (cumulative) | estimated |
| ai_pubs_share | 0.9 | 2023 | OECD / Elsevier | estimated |
| ai_regulation_idx | 76 | 2024 | EU AI Act/curated | estimated |
| ai_skill_penetration | 67 | 2024 | LinkedIn/Stanford AI Index 2025 | estimated |
| ai_strategy_year | 2022 | 2022 | OECD AI Policy Observatory | estimated |
| ai_talent_idx | 51 | 2024 | Tortoise Global AI Index 2024 | estimated |
| cloud_regions | 2 | 2024 | AWS/Azure/GCP region maps 2024 | measured |
| data_governance_idx | 86 | 2024 | OECD/GDPR/curated | estimated |
| datacenter_mw | 400 | 2024 | curated estimate | estimated |
| digital_skills_idx | 73 | 2024 | WEF / ITU-style composite | estimated |
| enterprise_agent_adoption | 50.0 | 2025 | curated (agentic-deployment estimate) | estimated |
| fixed_broadband_per100 | 41 | 2023 | ITU 2023 | measured |
| gdp_pc_ppp | 74676.0 | 2025 | World Bank WDI (PPP) | measured |
| gdp_total_b | 725.5 | 2025 | World Bank WDI | measured |
| gii_score | 57.5 | 2024 | WIPO Global Innovation Index 2024 | measured |
| gov_agent_adoption | 54.0 | 2025 | curated (agentic-deployment estimate) | estimated |
| gov_ai_readiness | 76 | 2024 | Oxford Insights Gov AI Readiness 2024 | measured |
| human_capital_index | 0.76 | 2020 | World Bank HCI 2020 | measured |
| internet_pct | 94 | 2023 | ITU 2023 | measured |
| mean_years_schooling | 12.4 | 2022 | UNDP HDR 2023/24 | measured |
| mobile_broadband_per100 | 100 | 2023 | ITU 2023 | estimated |
| notable_models | 0 | 2024 | Stanford AI Index 2025 | estimated |
| patents_per_million | 90 | 2023 | WIPO | estimated |
| population_m | 11.9 | 2025 | World Bank WDI | measured |
| rd_gdp_pct | 3.43 | 2021 | OECD / Eurostat | measured |
| researchers_per_million | 5710 | 2021 | UNESCO UIS / OECD | measured |
| robot_density | 220.0 | 2024 | IFR World Robotics | estimated |
| robot_installs_k | 1.3 | 2024 | IFR World Robotics | estimated |
| robot_stock_k | 11.0 | 2024 | IFR World Robotics | estimated |
| sci_pubs_per_million | 2100 | 2022 | World Bank WDI | estimated |
| semiconductor_idx | 72 | 2024 | curated estimate (imec) | estimated |
| stem_grads_share | 21 | 2020 | OECD/Eurostat | measured |
| tertiary_enrollment_pct | 80 | 2021 | UNESCO UIS | measured |
| top500_systems | 3 | 2024 | Top500 2024 | estimated |