Anthropic, OpenAI Face AI Model Launch Halts under Federal Standards

- US intensifies AI oversight as Anthropic and OpenAI prepare IPOs - Executive orders restrict model releases amid unresolved state–federal conflicts On 2026-07-02, Axios reported that the Trump administration is implementing voluntary federal standards for advanced AI model releases. The administration is using a series of executive orders issued from December 2025 to June 2026 to centralize oversight and challenge conflicting state laws, and this framework allows federal agencies to suspend or delay launches by major companies including Anthropic and OpenAI. However, it stops short of establishing a new federal regulator or fully preempting state-level rules, thereby fueling uncertainty as both companies gear up for IPOs. According to Axios, federal oversight began with a December 2025 executive order instructing the Commerce Department to review state AI laws for possible preemption and form an AI Litigation Task Force targeting restrictive states. In addition, a June 2026 order created a classified benchmarking system led by the National Institute of Standards and Technology (NIST) and established a government review window of up to 30 days for “frontier” AI models, while existing agencies such as NIST and the Cybersecurity and Infrastructure Security Agency (CISA), rather than a dedicated federal body, continue to manage oversight. Axios reported that these executive actions have given the federal government broad discretionary power to halt or restrict new model releases. As a result, Anthropic was required to restrict or discontinue its Claude Fable 5 and Mythos 5 models, while OpenAI delayed GPT-5.6 launches to select government-approved partners. Meanwhile, recent moves have lifted export controls on Anthropic’s models, and negotiations are ongoing for broader voluntary standards. Both firms still face unpredictable federal interventions, which complicate their IPO plans and amplify industry-wide regulatory uncertainty. According to Axios on 2026-07-02, the administration aims to replace conflicting state regulations with a national standard. However, Axios also confirmed that executive orders alone cannot fully override state authority, and only Congress can comprehensively preempt state AI laws. Therefore, several states continue to pursue their own AI requirements, and new federal measures add further complexity for leading firms. Reporting from Axios, Forbes, Politico, and The Washington Post indicates that regulatory ambiguity remains for Anthropic, OpenAI, and other AI leaders. Voluntary federal standards, discretionary model reviews, and the unresolved state–federal split continue to shape the landscape as the administration works to consolidate national AI oversight.