The CREATE AI Act is a U.S. federal bill, re-entered by Sen. Todd Young (R-IN) during the week of April 27, 2026, that would codify a National Science Foundation initiative providing research compute, data, and tooling for U.S. researchers and small businesses. It is associated with the National AI Research Resource (NAIRR) track, which proposes public-sector compute and data access as a counterweight to frontier-lab compute concentration. (Source: insideaipolicy.com)
Status and timeline
Young re-introduced the bill during the week of April 27, 2026; a previous version had been pending in earlier Congresses. The same week, Sen. Mark Warner (D-VA) introduced an AI jobs-data measure, the data-infrastructure component of a broader Senate AI policy push and connected to the data side of AI labor disruption. (Source: insideaipolicy.com)
The re-entry coincided with other federal AI activity that same week: Sen. Bernie Sanders convened a meeting of U.S. and Chinese AI scientists; the Center for Democracy & Technology (CDT) released a report on compute thresholds; and the House Homeland Security Committee and the House Select Committee on the Chinese Communist Party opened a joint investigation of Chinese AI models, including DeepSeek and Alibaba. (Source: insideaipolicy.com)
On June 25, 2026, the House Science Committee passed ten bipartisan AI bills spanning cybersecurity, research, and workforce preparation, including measures to codify the CREATE AI Act and the National AI Research Resource at the National Science Foundation, and to establish a Center for AI Security and Innovation at NIST. The committee had announced the markup of the ten measures days earlier (Source: insideaipolicy.com); their passage moved the NAIRR public-compute track from re-introduced bill to a bill reported out of committee (Source: insideaipolicy.com).
Scope and provisions
The bill would codify a National Science Foundation initiative to provide AI research compute, data, and tooling to U.S. researchers and small businesses. It sits on the public-AI-resources track associated with the National AI Research Resource (NAIRR), under which public-sector compute and data access are proposed as a counterweight to the concentration of compute among frontier labs. If enacted, NIST is a potential implementation partner. (Source: insideaipolicy.com)
The bill has been positioned against the widening gap between private and publicly accessible research compute. OpenAI reported a 10 GW compute milestone on April 30, 2026, putting frontier-lab compute well above publicly accessible research compute. Hyperscaler 2026 capital expenditure was reported at roughly $725 billion, up 77% from $410 billion in 2025. Nvidia B300 servers were reported selling for roughly $1 million each on the China black market, reflecting a widening international compute-access gap. (Source: insideaipolicy.com)
Relationships
- proposed-by: Sen. Todd Young (R-IN)
- related: Compute Governance, AI Divides (Literacy / Occupational / Ethico-Philosophical), AI as Normal Technology, Stanford HAI (NAIRR advocacy), National Institute of Standards and Technology (NIST)