Case studies

Case Study / Stanford HAI

Stanford AI Index 2026, turned into a navigable working brief.

This case study reorganizes the 2026 AI Index Report into a practical reading path: one overview page, one page per chapter, and compact data tables that keep the most decision-relevant numbers visible.

The Stanford HAI project

Stanford HAI frames the AI Index as an independent measurement effort: it tracks, collates, distills and visualizes broadly sourced AI data so policymakers, researchers, executives and the public can reason about a fast-moving field with better evidence.

As of June 17, 2026, the official Stanford HAI report overview lists nine chapters. This case study follows the full chapter lineup rather than an eight-section approximation.

How to use it
  1. Treat the overview as an executive brief and each chapter page as a drill-down.
  2. Use the tables to compare numeric signals across time, actors and sectors.
  3. Follow the official chapter page and PDF when you need the full visual evidence and figure context.

Key figures

Cross-report snapshot

These signals capture the report-wide pattern: capabilities are accelerating, adoption is spreading, investment is concentrating and governance still lags the deployment curve.

>90%Notable AI models produced by industry in 2025Stanford's overview highlights that frontier model production is now overwhelmingly industrial.
88%Organizations reporting AI use in 2025Adoption is no longer an experiment-only story: AI is already present in most surveyed organizations.
362Documented AI incidents in 2025Capability growth is outrunning responsible AI measurement and deployment controls.
$172BEstimated annual U.S. consumer value from generative AI by early 2026The economic footprint is spreading into everyday use, not only enterprise experimentation.

Analysis

What the report is really saying

Stanford's overview page compresses the report into a few cross-cutting signals. These cards restate that logic in a faster working format.

Capability is still accelerating

Stanford's overview states that AI is not plateauing: benchmark saturation is faster, agent benchmarks are moving, and frontier models already exceed several human baselines.

Competition is now geopolitical and infrastructural

The report repeatedly links model performance to compute, data centers, chips, supply chains and national strategy, not only to model architecture.

Governance is becoming a production bottleneck

Responsible AI reporting, policy clarity and institutional trust are advancing more slowly than capability, adoption and capital formation.

AI is now a sector story, not only a model story

Science, medicine, education and public opinion chapters show that AI's real impact is now measured inside disciplines, workflows and social systems.

Report structure

Chapter map

Each chapter below gets its own page with a summary lens, curated metrics, implications and direct links back to the official Stanford source materials.

01Models, compute, infrastructure

Research and Development

How model production, compute capacity, disclosure, open source and infrastructure concentration are reshaping the AI supply side.

Open chapter
02Benchmarks, convergence, jagged intelligence

Technical Performance

How fast benchmarks are saturating, how narrow the frontier leaderboard has become and where model competence still breaks in surprising ways.

Open chapter
03Incidents, transparency, trade-offs

Responsible AI

Where safety, fairness, transparency and governance are improving, and where the measurement and disclosure gap is actually widening.

Open chapter
04Investment, adoption, labor

Economy

A chapter on AI as capital formation, enterprise adoption and labor-market change rather than only as technical progress.

Open chapter
05Research workflows, discipline-specific AI

Science

A new chapter in the 2026 report that tracks how AI is moving from generic tooling into scientific infrastructure and end-to-end research tasks.

Open chapter
06Biology models, clinical workflow, regulation

Medicine

How AI is changing biological modeling, clinical documentation, diagnostics, search behavior and medical-device regulation at the same time.

Open chapter
07Students, policy lag, skills pipeline

Education

A look at AI adoption in schools and universities, the mismatch between usage and policy, and the changing pipeline of AI skills.

Open chapter
08Sovereignty, public spending, institutions

Policy and Governance

The policy chapter tracks national strategy, public investment, data sovereignty and the institutional shape of AI competition.

Open chapter
09Trust, anxiety, labor expectations

Public Opinion

How people, experts and institutions diverge on AI benefits, risks, jobs, companionship and regulation.

Open chapter

Sources

Primary sources

All content here is based on Stanford HAI's official AI Index overview, chapter pages, chapter PDFs and linked public data folder.

Stanford HAI overview page Official chapter lineup, top takeaways and report framing.
Full report PDF Full visual evidence, figures, notes and citations from Stanford HAI.
Public data folder The report's linked public data workspace for deeper analysis.