Barcelona Supercomputing Centre

SAIL rating system

The most comprehensive system for evaluating sovereign AI strategies and systems.

Barcelona Supercomputing Centre | SAIL-certified AI infrastructure facility

Featured essay

Decoding sovereign AI: same word, different interests

When policymakers say sovereign AI, they rarely mean the same thing. This essay maps the five interests behind the slogan — security, industrial policy, enterprise control, cultural identity, and middle-power alliance — and explains when the technical stack can help.

Read the essay →

SAIL-certified sovereign AI systems are better systems

SAIL (Sovereign AI Leadership) is a rating and evaluation system that provides a structured assessment of a country's or agency's AI sovereignty across technical, legal, and governance dimensions. It offers a point-based system that signals resilience, control, and strategic leadership.

SAIL applies to national strategies, ministries, agencies, and public programs that deploy or govern AI systems. Whether assessing sovereign capacity in applications, data, models, or infrastructure, SAIL helps governments identify strengths and dependencies, optimize capacity, and benchmark progress over time.

How SAIL works

SAIL evaluates sovereignty across seven layers — from applications and orchestration to compute and legal governance — through credit categories tailored for each assessment track. An entity must first satisfy baseline prerequisites, then accrue points toward a certification level: Certified, Silver, Gold, or Platinum.

Platinum certification icon
Platinum
80+ points
Gold certification icon
Gold
60–79 points
Silver certification icon
Silver
50–59 points
Certified certification icon
Certified
40–49 points

Read the full SAIL specification →

SAIL assessment process

1

Select track and scope

Determine whether the assessment applies to a national strategy, ministry, agency, or specific program, and review relevant prerequisites.

2

Define the assessment

Provide baseline documentation and define the scope of AI systems, governance processes, and dependencies.

3

Build the scorecard

Choose credits aligned with your strategy and map responsibilities for evidence gathering and documentation.

4

Implement and document

Execute technical, organizational, and legal actions needed to meet prerequisites and pursue credit categories.

5

Complete peer review

Submit evidence for independent review, scoring, and certification level determination.

SAIL tools and resources

Specification

Complete methodology: seven layers, credits, flags, and control-maturity levels.

Models

Sovereign AI models tagged by the interests they serve.

Countries

Interest profiles and illustrative stack-control assessments.

Interest decoder

Reference guide to the five interests behind sovereign AI claims.

Cases

Annotated examples: Mistral, OpenAI for Countries, allied public models, and more.

Featured essay

Why decision-makers care about sovereign AI — and when the spec applies.

Getting started

  • Review prerequisites — Ensure your entity meets the minimum requirements (e.g. legal authority, inventory of AI use).
  • Select the relevant track — National, subnational, or programmatic sovereignty.
  • Explore the credit library — Identify which credits are achievable and which require further capacity building.

About SAIL

SAIL is developed and maintained by the Public AI Network as part of its work on AI as public infrastructure.