AI Safety Tracker

AI in the United States

10 instruments 77 datacentre sites 700 notable models

33.0% of people of working age in the United States were using AI tools in 2026-Q2, by Microsoft's estimate, behind 20 of 147 economies. 10 instruments relevant to AI in US law and policy are indexed, 6 of them binding. Epoch AI lists 700 notable models with a developer in the United States, and 77 AI datacentre sites are recorded there in Epoch AI's hub.

  • 33.0% of the working-age population using AI tools, 2026-Q2 AI Diffusion Report: AI User Share Microsoft AI for Good Lab Licence MIT Updated quarterly Retrieved , today Derived from anonymised Microsoft telemetry, then adjusted for OS/device market share, internet penetration and a modelled mobile-to-desktop ratio. It therefore measures Microsoft-visible AI use, scaled, not observed total AI use. Systematically less reliable where Microsoft platform share is low, and it cannot see AI embedded in products users do not recognise as AI. Cite Microsoft AI for Good Lab, "AI Diffusion Report: AI User Share". https://github.com/microsoft/ai-diffusion-report. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.
  • 6 of 10 indexed instruments that are binding law AI Law and Policy Index This site Licence MIT Updated quarterly Retrieved , 5 days ago Data JSON CSV Our own reading of primary instruments, with a link to each. Not legal advice, not a compliance checklist and not exhaustive. Instruments that never mention AI are included where they apply to it, which is most of the Australian list. Statuses and commencement dates go stale; every record carries its review date. Cite AI Safety Tracker, "AI Law and Policy Index". https://aisafetytracker.org/policy/. Retrieved 23 October 2026.
  • 6 of 10 governance readiness points AI Governance Readiness Index This site Licence MIT Updated quarterly Retrieved , 5 days ago Data JSON Our own editorial coding from primary legal sources, not an objective measurement. Every score cites a specific instrument and article. "Readiness" means the machinery exists to govern AI. It is emphatically not a judgement that the resulting rules are good, proportionate, or well enforced. Cite AI Safety Tracker, "AI Governance Readiness Index". https://aisafetytracker.org/methodology/governance-readiness/. Retrieved 23 October 2026.
  • 33.1 GW of IT power planned at 77 sites in Epoch AI's hub, 12 GW of it operating AI Data Centers Epoch AI Licence CC-BY-4.0 Updated several times a week Retrieved , today The largest AI sites Epoch has found, mostly in the United States. Epoch estimated in September 2026 that they hold 44% of the world's deployed AI compute, with a 90% interval of 23% to 81%. Power, compute and cost are Epoch's estimates from satellite imagery, permits and company statements; four in five IT power estimates fall within a factor of 1.4 of the true value. The owner is the owner of the chips, not necessarily of the building. Capacity after the date of retrieval is Epoch's projection. Coordinates are read from the hub's map page, which publishes them; the downloads do not. Cite Epoch AI, "AI Data Centers". https://epoch.ai/data/ai-data-centers. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

Law and policy

10 instruments that apply to AI in the United States, grouped by how binding each is. 6 are binding law, and 2 of those name AI or deepfakes in their titles. AI Law and Policy Index This site Licence MIT Updated quarterly Retrieved , 5 days ago Data JSON CSV Our own reading of primary instruments, with a link to each. Not legal advice, not a compliance checklist and not exhaustive. Instruments that never mention AI are included where they apply to it, which is most of the Australian list. Statuses and commencement dates go stale; every record carries its review date. Cite AI Safety Tracker, "AI Law and Policy Index". https://aisafetytracker.org/policy/. Retrieved 23 October 2026.

Binding instrument Kind Year
Colorado SB26-189, Automated Decision-Making Technology Act 2026
California SB 53, Transparency in Frontier Artificial Intelligence Act Act 2025
New York RAISE Act, Responsible AI Safety and Education Act Act 2025
TAKE IT DOWN Act Act 2025
Advanced computing and semiconductor export controls Rule under an Act 2022
Federal Trade Commission Act, section 5 Act 1914
AI law and policy in the United States

Data: the AI Law and Policy Index, read from each primary instrument. Not legal advice.

Use of AI tools

Share of the working-age population using AI tools, by reporting period, and the economies ranked either side of the United States in 2026-Q2. Microsoft's estimate from its own telemetry, scaled to each population. AI Diffusion Report: AI User Share Microsoft AI for Good Lab Licence MIT Updated quarterly Retrieved , today Derived from anonymised Microsoft telemetry, then adjusted for OS/device market share, internet penetration and a modelled mobile-to-desktop ratio. It therefore measures Microsoft-visible AI use, scaled, not observed total AI use. Systematically less reliable where Microsoft platform share is low, and it cannot see AI embedded in products users do not recognise as AI. Cite Microsoft AI for Good Lab, "AI Diffusion Report: AI User Share". https://github.com/microsoft/ai-diffusion-report. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

  • 26.3% 2025-H1
  • 28.3% 2025-H2
  • 31.3% 2026-Q1
  • 33.0% 2026-Q2
Economy 2026-Q2 Rank
Canada 38.5% 16
Sweden 37.3% 17
Austria 35.4% 18
Taiwan 33.8% 19
Hungary 33.3% 20
United States 33.0% 21
Denmark 32.4% 22
Germany 32.3% 23
Poland 32.3% 23
Italy 31.4% 25
Saudi Arabia 31.4% 25

Ranked within 147 economies. Bars are scaled to the highest share, United Arab Emirates at 73.3%. The estimate measures AI use visible to Microsoft and is least reliable where its platform share is low.

AI adoption

Data: Microsoft AI for Good Lab, AI Diffusion Report: AI User Share

Governance readiness

6 of 10 points on a five-dimension rubric recording how much machinery exists to govern AI, not how strict it is. 4 of 31 nationally coded jurisdictions score higher and 4 score the same. AI Governance Readiness Index This site Licence MIT Updated quarterly Retrieved , 5 days ago Data JSON Our own editorial coding from primary legal sources, not an objective measurement. Every score cites a specific instrument and article. "Readiness" means the machinery exists to govern AI. It is emphatically not a judgement that the resulting rules are good, proportionate, or well enforced. Cite AI Safety Tracker, "AI Governance Readiness Index". https://aisafetytracker.org/methodology/governance-readiness/. Retrieved 23 October 2026.

Dimension Score What the score records
Binding law 1 of 3 Sectoral or partial rules only.
Oversight body 1 of 2 A coordinating or advisory body.
Evaluation capacity 2 of 2 An operating safety or evaluation institute.
Transparency duty 1 of 2 A general data-protection route only.
Strategy 1 of 1 A published national AI strategy.

No federal horizontal AI statute. Executive Order 14110 was rescinded in January 2025 and replaced by Executive Order 14179, which directs policy without creating duties on developers; Executive Order 14365 of December 2025 directs federal challenges to state AI laws. Binding obligations sit in sectoral rules and in state law: Colorado's SB 26-189 on automated decisions, which replaced its 2024 AI Act in May 2026, and frontier-model statutes in California, SB 53, in force from January 2026, and New York, the RAISE Act, from January 2027. The Center for AI Standards and Innovation at NIST, formerly the US AI Safety Institute, runs model evaluations. Cited for the score: Executive Order 14179 (2025), Executive Order 14365 (2025), NIST AI Risk Management Framework (2023), Colorado SB 26-189 (2026), California SB 53 (2025), New York RAISE Act (S6953-B) (2025).

Governance readiness rubric

Data: the AI Governance Readiness Index. Against AI use in every jurisdiction: exposure against governance readiness

Compute thresholds

The 4 training-compute thresholds in US law and policy, the day each took its final form, and the models with a published compute figure above each that day and now. Models with a developer in the United States are counted separately. Compute Threshold Index This site Licence MIT Updated quarterly Retrieved , 5 days ago Data JSON CSV Training-compute thresholds written into law, executive action or published policy, with the text that sets each one. Whether a given model crosses a threshold here is arithmetic on Epoch's compute estimates, most of which are estimates rather than disclosures, and is not a legal finding about any model or developer. Cite AI Safety Tracker, "Compute Threshold Index". https://aisafetytracker.org/capability/#thresholds. Retrieved 23 October 2026. Notable AI Models Epoch AI Licence CC-BY-4.0 Updated weekly Retrieved , 5 days ago Inclusion is editorial ("notability criteria"), not exhaustive. Compute figures are frequently estimates; check the per-row confidence and notes columns before treating any value as measured. Cite Epoch AI, "Notable AI Models". https://epoch.ai/data/ai-models. Retrieved 23 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

Where Instrument Threshold Status Dated Above that day First crossed Above now Developer here
California SB 53, Transparency in Frontier Artificial Intelligence Act 1e26 in force 2025-09-29 signed 3 Grok 3, 2025-02-17 4 4
Federal Executive Order 14110, section 4.2 1e26 revoked 2023-10-30 signed 0 Grok 3, 2025-02-17 4 4
California SB 1047, Safe and Secure Innovation for Frontier AI Models Act 1e26 vetoed 2024-09-29 vetoed 0 Grok 3, 2025-02-17 4 4
New York RAISE Act, Responsible AI Safety and Education Act 1e26 enacted 2025-12-19 signed 3 Grok 3, 2025-02-17 4 4
What each threshold requires 4

  • SB 53, Transparency in Frontier Artificial Intelligence Act

    Applies to Frontier developers training models above the threshold

    Consequence Duty to publish a frontier AI framework and transparency report, and to report critical safety incidents to the state.

    Counts the initial training run plus later fine-tuning and material modification, so it is a cumulative figure rather than a single-run one. Larger obligations attach above an annual revenue threshold as well.

  • Executive Order 14110, section 4.2

    Applies to Developers of dual-use foundation models

    Consequence Reporting of training runs, results of red-team testing and security measures.

    The first compute threshold in any official instrument, revoked in January 2025. Included because it is still widely quoted as current, and because California later adopted the same number.

  • SB 1047, Safe and Secure Innovation for Frontier AI Models Act

    Applies to Developers of covered models above the threshold and above $100m training cost

    Consequence Would have required a safety protocol, third-party audit and a shutdown capability.

    Vetoed in September 2024. Its successor, SB 53, kept the compute threshold and dropped the audit and shutdown requirements.

  • RAISE Act, Responsible AI Safety and Education Act applies from 2027-01-01

    Applies to Frontier developers training models above the threshold, with the heavier duties on those with over $500 million in annual revenue

    Consequence Duty to publish safety and security protocols and to report safety incidents to the state.

    Signed in December 2025 and amended in March 2026 to track California's SB 53. It applies from 1 January 2027, so it is counted here as enacted rather than in force.

Whether a model sits above a line is arithmetic on Epoch's compute figures, most of which are estimates rather than disclosures. Nothing here is a legal finding about any model or developer.

Every compute threshold

Data: the Compute Threshold Index, and Epoch AI's compute estimates

AI datacentres and compute

77 AI datacentre sites in Epoch AI's hub in the United States, by the IT power planned at the last stage of each site's timeline, with the power operating now. AI Data Centers Epoch AI Licence CC-BY-4.0 Updated several times a week Retrieved , today The largest AI sites Epoch has found, mostly in the United States. Epoch estimated in September 2026 that they hold 44% of the world's deployed AI compute, with a 90% interval of 23% to 81%. Power, compute and cost are Epoch's estimates from satellite imagery, permits and company statements; four in five IT power estimates fall within a factor of 1.4 of the true value. The owner is the owner of the chips, not necessarily of the building. Capacity after the date of retrieval is Epoch's projection. Coordinates are read from the hub's map page, which publishes them; the downloads do not. Cite Epoch AI, "AI Data Centers". https://epoch.ai/data/ai-data-centers. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

Site Planned IT power Now
Microsoft Fairwater Wisconsin Microsoft, expanding 2.3 GW 369 MW
Anthropic-Amazon New Carlisle Amazon, expanding 1.9 GW 1 GW
OpenAI Stargate New Mexico Oracle, under construction 1.8 GW none
Meta Hyperion Meta, under construction 1.6 GW none
Colossus 2 SpaceXAI, expanding 1.5 GW 946 MW
OpenAI Stargate Shackelford Oracle, under construction 1.4 GW none
QTS Cedar Rapids under construction 1.1 GW none
Meta Prometheus Meta, expanding 1 GW 471 MW
Goodnight Google, under construction 1 GW none
OpenAI Stargate Michigan Oracle, under construction 988 MW none
Show the remaining 67
OpenAI Stargate Wisconsin Oracle, under construction 902 MW none
Google Fort Wayne Google, under construction 863 MW none
OpenAI Stargate Milam Softbank, under construction 857 MW none
OpenAI Stargate Abilene Oracle, expanding 843 MW 421 MW
Amazon Ridgeland Amazon, expanding 840 MW 228 MW
Amazon Madison Mega Site Amazon, expanding 683 MW 284 MW
Crusoe Abilene Expansion Microsoft, under construction 672 MW none
Microsoft Fairwater Atlanta Microsoft, operating 636 MW 636 MW
Coreweave Helios CoreWeave, expanding 528 MW 132 MW
Meta Bowling Green Meta, under construction 422 MW none
CoreWeave Ellendale ND CoreWeave, expanding 400 MW 68 MW
Google Omaha Google, expanding 395 MW 237 MW
Google Columbus Google, expanding 386 MW 303 MW
Anthropic Lake Mariner AI XPV Platform, expanding 378 MW 42 MW
Google Pryor (North) Google, operating 368 MW 368 MW
Meta Montgomery Meta, expanding 358 MW 153 MW
STACK Infrastructure NVA02 expanding 348 MW 85 MW
Colossus 1 SpaceXAI, operating 340 MW 340 MW
AWS New Albany Amazon, expanding 338 MW 213 MW
Google New Albany Google, operating 333 MW 333 MW
QTS Eagle Mountain Meta, under construction 304 MW none
Google Storey County Google, expanding 300 MW 161 MW
Google Lincoln Google, expanding 283 MW 141 MW
CoreWeave Denton TX CoreWeave, expanding 282 MW 262 MW
Google Cedar Rapids Google, under construction 261 MW none
QTS Richmond 3 expanding 240 MW 144 MW
Google Bristow Google, operating 238 MW 238 MW
QTS Richmond 1 operating 238 MW 238 MW
Google Council Bluffs (East) Google, operating 237 MW 237 MW
Google Papillion Google, operating 237 MW 237 MW
Stream Phoenix expanding 220 MW 64 MW
Meta Los Lunas Meta, expanding 218 MW 118 MW
OpenAI Stargate Lordstown Softbank, under construction 214 MW none
Anthropic Barber Lake Cipher Mining, under construction 207 MW none
Meta Cheyenne Meta, expanding 207 MW 152 MW
Microsoft Goodyear Microsoft, operating 202 MW 202 MW
Microsoft Project Osmium Microsoft, operating 190 MW 190 MW
Google Mesa Google, operating 183 MW 183 MW
Meta-QTS Hillsboro 2 Meta, operating 180 MW 180 MW
QTS Richmond 2 operating 180 MW 180 MW
Meta Jeffersonville Meta, operating 178 MW 178 MW
Meta Rosemount Meta, operating 178 MW 178 MW
Meta Temple Meta, operating 178 MW 178 MW
AWS Berwick Amazon, expanding 169 MW 108 MW
Google Red Oak Google, expanding 154 MW 77 MW
Google The Dalles Google, operating 154 MW 154 MW
CoreWeave Muskogee OK CoreWeave, under construction 153 MW none
Meta Kuna Meta, operating 152 MW 152 MW
Meta Huntsville Meta, operating 146 MW 146 MW
Google Lancaster Google, operating 137 MW 137 MW
Meta Eagle Mountain Meta, operating 133 MW 133 MW
Meta Aiken Meta, expanding 129 MW 69 MW
Vantage TX1 expanding 128 MW 32 MW
CoreWeave Chester VA CoreWeave, expanding 113 MW 82 MW
Google Kansas City East Google, operating 105 MW 105 MW
Google Midlothian Google, operating 103 MW 103 MW
CoreWeave Lancaster Greenfield site CoreWeave, under construction 92 MW none
Meta Gallatin Meta, operating 87 MW 87 MW
Meta Sarpy Meta, operating 80 MW 80 MW
Google Arcola Google, operating 77 MW 77 MW
Microsoft SAT40 Microsoft, operating 75 MW 75 MW
CoreWeave Marble NC CoreWeave, operating 65 MW 65 MW
Core42 Lake Mariner Core42, operating 58 MW 58 MW
Microsoft-Nebius New Jersey Nebius, operating 50 MW 50 MW
Microsoft SAT14 Microsoft, operating 47 MW 47 MW
CoreWeave Dalton 1 & 2 CoreWeave, operating 28 MW 28 MW
xAI QTS Atlanta SpaceXAI, operating 17.3 MW 17.3 MW
  • 12 GW IT power operating at those sites now AI Data Centers Epoch AI Licence CC-BY-4.0 Updated several times a week Retrieved , today The largest AI sites Epoch has found, mostly in the United States. Epoch estimated in September 2026 that they hold 44% of the world's deployed AI compute, with a 90% interval of 23% to 81%. Power, compute and cost are Epoch's estimates from satellite imagery, permits and company statements; four in five IT power estimates fall within a factor of 1.4 of the true value. The owner is the owner of the chips, not necessarily of the building. Capacity after the date of retrieval is Epoch's projection. Coordinates are read from the hub's map page, which publishes them; the downloads do not. Cite Epoch AI, "AI Data Centers". https://epoch.ai/data/ai-data-centers. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.
  • 12,789,000 H100 equivalents of compute at those sites AI Data Centers Epoch AI Licence CC-BY-4.0 Updated several times a week Retrieved , today The largest AI sites Epoch has found, mostly in the United States. Epoch estimated in September 2026 that they hold 44% of the world's deployed AI compute, with a 90% interval of 23% to 81%. Power, compute and cost are Epoch's estimates from satellite imagery, permits and company statements; four in five IT power estimates fall within a factor of 1.4 of the true value. The owner is the owner of the chips, not necessarily of the building. Capacity after the date of retrieval is Epoch's projection. Coordinates are read from the hub's map page, which publishes them; the downloads do not. Cite Epoch AI, "AI Data Centers". https://epoch.ai/data/ai-data-centers. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.
  • 119 GPU clusters in Epoch AI's cluster records, 2 GW of known power GPU Clusters Epoch AI Licence CC-BY-4.0 Updated monthly Retrieved , today Compiled from public reporting; each row carries a certainty flag ranging from confirmed to speculative. Chinese clusters are systematically less well documented than US ones, so country totals understate opacity as much as they measure capacity. Cite Epoch AI, "GPU Clusters". https://epoch.ai/data/gpu-clusters. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

Hub power and compute are Epoch's estimates, and planned power is its projection from permits and announcements. Epoch's cluster records are a separate compilation, not reconciled with the hub, and count power only where it is known.

AI datacentres

Data: Epoch AI

Notable models

700 models in Epoch AI's notable models list with an organisation in the United States among their developers, 189 of them from academic groups and 117 published since January 2025. The largest training compute recorded for any of them is 9.4e26 FLOP, for GPT-6 Astra. Notable AI Models Epoch AI Licence CC-BY-4.0 Updated weekly Retrieved , today Inclusion is editorial ("notability criteria"), not exhaustive. Compute figures are frequently estimates; check the per-row confidence and notes columns before treating any value as measured. Cite Epoch AI, "Notable AI Models". https://epoch.ai/data/ai-models. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

Year Models
2026 53
2025 64
2024 57
2023 86
2022 68
2021 54
2020 37
2019 56
2018 27
2017 29
Show the remaining 43
2016 36
2015 19
2014 16
2013 13
2012 6
2011 5
2010 4
2009 7
2008 2
2007 4
2006 2
2005 1
2004 1
2003 2
2002 1
2001 3
1999 1
1998 1
1997 1
1996 2
1995 2
1994 1
1993 2
1992 3
1990 2
1989 5
1988 1
1987 3
1986 4
1983 1
1982 1
1973 1
1966 1
1965 1
1963 1
1962 2
1960 2
1959 3
1957 1
1955 2
1954 1
1952 1
1950 1
Organisation Models
Google 82
OpenAI 80
Google DeepMind 50
Meta AI 44
Stanford University 35
Google Brain 31
Nvidia 31
Google Research 30
Carnegie Mellon University (CMU) 29
Microsoft 27
Show the remaining 118
Anthropic 22
Massachusetts Institute of Technology (MIT) 21
University of California (UC) Berkeley 20
University of Washington 18
Microsoft Research 17
New York University (NYU) 16
IBM 14
Facebook 13
AT&T 12
Allen Institute for AI 11
Princeton University 11
Columbia University 9
xAI 9
Apple 8
Johns Hopkins University 8
Salesforce Research 8
University of California San Diego 8
University of Illinois Urbana-Champaign (UIUC) 8
Bell Laboratories 7
Amazon 6
Facebook AI 6
Harvard University 5
Hugging Face 5
California Institute of Technology 4
Cornell University 4
Intel Labs 4
Runway 4
Salesforce 4
Toyota Technological Institute at Chicago 4
University of Texas at Austin 4
Brown University 3
Character.ai 3
Harvard Medical School 3
NEC Laboratories 3
Texas A&M 3
University of California Los Angeles (UCLA) 3
University of Michigan 3
University of Southern California 3
University of Wisconsin Madison 3
UT Austin 3
ASAPP 2
Cerebras Systems 2
Cornell Aeronautical Laboratory 2
Cursor 2
Duke University 2
EleutherAI 2
Georgia Institute of Technology 2
Google Cloud 2
Inception Labs 2
Inflection AI 2
MIT-IBM Watson AI Lab 2
Northeastern University 2
Oak Ridge National Laboratory 2
Physical Intelligence 2
Rutgers University 2
Thinking Machines 2
Twitter 2
University at Buffalo 2
University of Chicago 2
University of Pennsylvania 2
Writer 2
Adobe 1
Bloomberg 1
Capital One 1
Chai discovery 1
Cognition 1
Collaborations Pharmaceuticals 1
Compaq CRL 1
Cortex AI 1
Courant Institute of Mathematical Sciences 1
CZ Biohub Network 1
Databricks 1
Deep Trading 1
DeepScale 1
EvolutionaryScale 1
Generalist 1
Google AI 1
Grammarly 1
Institute for Advanced Study 1
Jane Street 1
Magic 1
MetaMind Inc 1
Michigan State University 1
Midjourney 1
Mitsubishi Electric Research Labs 1
National Cancer Institute 1
Netflix 1
Oregon State University 1
Perplexity 1
Prime Intellect 1
Purdue University 1
Reed College 1
Reka AI 1
Roblox 1
Saama 1
Sandia Corporation 1
Science Applications International Corporation / SAIC 1
ServiceNow 1
Suno 1
TensorTour 1
Toyota Research Institute 1
Uber AI 1
UC Irvine 1
University of Iowa 1
University of Kansas 1
University of Maine 1
University of Maryland 1
University of Massachusetts Amherst 1
University of Massachusetts Lowell 1
University of Minnesota 1
University of North Carolina 1
University of Notre Dame 1
University of Pittsburgh 1
University of Texas at Houston 1
University of Vermont 1
Wellesley College 1
Williams College 1
Yahoo Research 1

Notability is Epoch AI's editorial judgement, not a census. Organisations are named as Epoch names them, so one company can appear under several names. 32 models are counted in the total and by year but not by organisation, because Epoch's lists of organisations and countries for them do not line up.

Training compute and cost

Data: Epoch AI, Notable AI Models

Frontier developers

The developers in the Frontier Safety Framework Index with notable models Epoch AI credits to an organisation in the United States. 7 are based there. A developer is based where most of its credited models were built. Notable AI Models Epoch AI Licence CC-BY-4.0 Updated weekly Retrieved , 10 days ago Inclusion is editorial ("notability criteria"), not exhaustive. Compute figures are frequently estimates; check the per-row confidence and notes columns before treating any value as measured. Cite Epoch AI, "Notable AI Models". https://epoch.ai/data/ai-models. Retrieved 18 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/. Release Safety Disclosure Index This site Licence MIT Updated monthly Retrieved , 10 days ago Data JSON CSV Our own reading of the system card, model card or safety report each developer published for each release Epoch AI lists from 2025, joined to Epoch's list. Every cell is what a developer says about its own model under its own framework: whether a card came out with the release, whether it reports dangerous-capability results, which government institutes it names as pre-release testers, and whether it says a risk threshold was reached. None of it is verified. A release Epoch adds after the last review is shown as awaiting review. Cite AI Safety Tracker, "Release Safety Disclosure Index". https://aisafetytracker.org/alignment/#releases. Retrieved 18 October 2026. Notable AI Models Epoch AI Licence CC-BY-4.0 Updated weekly Retrieved , 10 days ago Inclusion is editorial ("notability criteria"), not exhaustive. Compute figures are frequently estimates; check the per-row confidence and notes columns before treating any value as measured. Cite Epoch AI, "Notable AI Models". https://epoch.ai/data/ai-models. Retrieved 18 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

Developer Models credited here Of all its models Based in Card on the day
Google DeepMind 176 217 here 6 of 8
OpenAI 80 80 here 22 of 30
Meta 63 92 here 3 of 5
Microsoft 42 47 here 1 of 1
Anthropic 22 22 here 15 of 15
xAI 9 9 here 1 of 6
Amazon 6 6 here -

Models are credited where Epoch's organisations and countries pair up, so a model whose two lists do not line up is not credited anywhere here. Card on the day counts releases since January 2025 in the Release Safety Disclosure Index, each the developer's own statement.

Frontier AI developers compared

Data: Epoch AI, Notable AI Models, and the Release Safety Disclosure Index

Official publications

18 publications on AI from US government feeds in the news feed's 120-day window, the latest 6 below, from 6 publishers in all. Headline and a short extract, each linking to the publisher. Federal Register API US Office of the Federal Register Licence public-domain Updated daily Retrieved , today Full-text keyword matching on "artificial intelligence" returns documents that merely mention AI in passing. Requires topical filtering before display. Cite US Office of the Federal Register, "Federal Register API". https://www.federalregister.gov/developers/documentation/api/v1. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/. NIST news US National Institute of Standards and Technology Licence public-domain Updated daily Retrieved , today Covers all of NIST's remit; AI is a small fraction. Requires topical filtering. Cite US National Institute of Standards and Technology, "NIST news". https://www.nist.gov/caisi. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

  • Government publication Notice United States
    Notice of Public Meetings of the Michigan Advisory Committee to the U.S. Commission on Civil Rights

    Notice is hereby given, pursuant to the provisions of the rules and regulations of the U.S. Commission on Civil Rights (Commission) and the Federal Advisory Committee Act, that the Michigan...

    Civil Rights Commission Federal Register API US Office of the Federal Register Licence public-domain Updated daily Retrieved , today Full-text keyword matching on "artificial intelligence" returns documents that merely mention AI in passing. Requires topical filtering before display. Cite US Office of the Federal Register, "Federal Register API". https://www.federalregister.gov/developers/documentation/api/v1. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

    • AI Policy Federal Advisory Committee Act, regulations
  • Government publication Notice United States
    Notice of Order Regarding Exclusion and Removal of DeepSeek Products From Federal Civilian Executive Branch Information Technology Systems and Procurements

    The Department of Homeland Security is publishing notice of the Secretary of Homeland Security's issuance of an order (1) excluding information technology products and services developed by...

    Homeland Security Department Federal Register API US Office of the Federal Register Licence public-domain Updated daily Retrieved , today Full-text keyword matching on "artificial intelligence" returns documents that merely mention AI in passing. Requires topical filtering before display. Cite US Office of the Federal Register, "Federal Register API". https://www.federalregister.gov/developers/documentation/api/v1. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

  • Government publication Notice United States
    Division of Program Coordination, Planning, and Strategic Initiatives, Office of the Director Notice of Proposed Reorganization

    The Division of Program Coordination, Planning, and Strategic Initiatives (DPCPSI) in the Office of the Director (OD), National Institutes of Health (NIH) is seeking public comment regarding its...

    Health and Human Services Department Federal Register API US Office of the Federal Register Licence public-domain Updated daily Retrieved , today Full-text keyword matching on "artificial intelligence" returns documents that merely mention AI in passing. Requires topical filtering before display. Cite US Office of the Federal Register, "Federal Register API". https://www.federalregister.gov/developers/documentation/api/v1. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

  • Government publication Notice United States
    Request for Information (RFI) Regarding the Digitization and Modernization of the National Technical Reports Library (NTRL) and the Leveraging of Scientific, Technical and Engineering Information (STEI) Stored in the NTRL

    The National Technical Information Service (NTIS) is seeking general information, feedback, suggestions, and experiential and technical insights from stakeholders to inform NTIS's goal of fully...

    Commerce Department Federal Register API US Office of the Federal Register Licence public-domain Updated daily Retrieved , today Full-text keyword matching on "artificial intelligence" returns documents that merely mention AI in passing. Requires topical filtering before display. Cite US Office of the Federal Register, "Federal Register API". https://www.federalregister.gov/developers/documentation/api/v1. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

  • Government publication Notice United States
    Solicitation of Nominations for Membership on the NOAA Science Advisory Board (SAB)

    The NOAA Science Advisory Board (SAB) is the only Federal Advisory Committee with responsibility to advise the Under Secretary of Commerce for Oceans and Atmosphere on long- and short-range...

    Commerce Department Federal Register API US Office of the Federal Register Licence public-domain Updated daily Retrieved , today Full-text keyword matching on "artificial intelligence" returns documents that merely mention AI in passing. Requires topical filtering before display. Cite US Office of the Federal Register, "Federal Register API". https://www.federalregister.gov/developers/documentation/api/v1. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

  • Government publication Notice United States
    Notice of Public Meeting of the Maryland Advisory Committee to the U.S. Commission on Civil Rights

    Notice is hereby given, pursuant to the provisions of the rules and regulations of the U.S. Commission on Civil Rights (Commission) and the Federal Advisory Committee Act, that the Maryland...

    Civil Rights Commission Federal Register API US Office of the Federal Register Licence public-domain Updated daily Retrieved , today Full-text keyword matching on "artificial intelligence" returns documents that merely mention AI in passing. Requires topical filtering before display. Cite US Office of the Federal Register, "Federal Register API". https://www.federalregister.gov/developers/documentation/api/v1. Retrieved 28 October 2026 via AI Safety Tracker, https://aisafetytracker.org/united-states/.

    • AI Policy Federal Advisory Committee Act, regulations
AI news

Data: government feeds, filed by keyword rule

Not covered

  • Depth of the law index. It holds 10 instruments for the United States and 28 for Australia, which was indexed first and in more depth, so the counts are not a comparison.
  • Enterprise use of AI. The enterprise survey on AI adoption is Eurostat's and covers Europe.
  • AI datacentres outside Epoch AI's hub. The AI Datacentre Index, which records sites from developers' statements, covers Australia only.
  • State law beyond California, Colorado and New York. The law index holds those three states' AI statutes and no other state's.
  • Electricity and water at individual US datacentres. LBNL's national estimates are on AI datacentres; no indexed source publishes figures per site.

Related

  • US law Every instrument indexed for the United States, with its primary source and what it binds.
  • Compute thresholds The US thresholds beside every other, on a timeline against the models.
  • Developers The frontier developers credited here, side by side with their disclosures.