Research
Everything the Rankings publish.
Ranking reports, methodology handbooks, frameworks and briefings — each published with the evidence base and the data behind it.
Research and publications Under development
The evidence behind the results.
Global AI Rankings 2026
The 2026 edition covers 195 governments across six equally weighted pillars, computed from published pillar data under a stated composite rule. Absent evidence is recorded as absent rather than estimated, and year-on-year movement is shown only where the source publishes it.
Also published
Methodology and Indicator Handbook: rules, rounding and reproduction
Framework In preparationAssessing oversight capacity rather than oversight law
Briefing In preparationMethodology revisions for the 2026 edition
Data In preparationPillar dataset: six pillars for every assessed government
Inside the 2026 report Under development
What the edition actually found.
A summarized preview of the full document. Every figure below is computed from the same register the report is built on.
195
Governments assessed — complete register, no sampling
43.6
Mean composite score, against a median of 41.1
36
Point mean spread across a government’s own six pillars
177
Record policy capacity above zero — 91% of the register
The leaders are outliers, not a leading group
The composite mean is 43.6 against a maximum of 87.7. The top decile begins more than thirty points above the median government. Read from the top the field looks close; read from the middle it does not.
Governments are uneven within themselves
The average government varies 36.2 points across its own six pillars — wider than the standard deviation of the whole register. A single composite conceals more variation than it reveals, which is why the report leads with pillar profiles rather than positions.
One weakness is nearly universal at the top
Of the leading 25 governments, 20 record development and diffusion as their weakest pillar. It is the lowest-scoring pillar register-wide. Capacity to govern AI is running ahead of capacity to build and spread it, and the pattern holds regardless of rank.
The strategy-to-delivery gap runs both ways
97 of 195 governments score higher on policy capacity than on delivery. The remaining 98 are the reverse case — deploying without a published policy position. The familiar claim that governments write strategies they do not implement describes about half the register, and the other half carries the accountability risk it does not.
What is in the document
Executive summary
Scope and method
Distribution and ledger
Pillar profiles and regions
The central finding
Coverage and limitations
Global AI Rankings is under development. Scores are synthesized from publicly available data and presented as aggregated estimates produced by a stated method. They are not audited findings, certifications, ratings or statements of fact about any government, institution, company or individual, and they are subject to revision without notice.
Discuss the 2026 findings with the research team.
Methodology questions, indicator-level data requests, briefings for policy and executive audiences, and media inquiries are handled directly by the assessment team.