OpenAI's Altman to brief US officials on next-gen model, rules weeks away

OpenAI Chief Executive Officer Sam Altman plans to brief Trump administration officials and lawmakers in Washington next week on the company's next-generation AI model series, as US regulators race to complete a safety review framework within weeks.

"The model series will bring some very interesting capabilities, especially in work scenarios and scaling work," Chris Lehane, OpenAI's global public affairs head, said.

The briefing comes as the Trump administration evaluates creating an independent regulatory body to audit AI models, with Treasury Secretary Scott Bessent involved in drafting proposals. The framework, expected within weeks, would give AI developers clearer compliance expectations after months of ad-hoc government interventions that have left the industry guessing.

The outcome carries billions in implications for the US technology sector. A light-touch framework could accelerate AI adoption and boost mega-cap valuations, while stringent restrictions would impose compliance costs and potentially slow product releases. The uncertainty has already weighed on tech stocks, which sold off after Chinese startup Moonshot AI's Kimi K3 model demonstrated competitive performance against US frontier models, raising questions about America's lead in AI development.

The regulatory push follows a turbulent period for AI oversight. Chris Fall stepped down as director of the Center for AI Standards and Innovation after roughly three months, the third leadership change at the agency in less than a year. His predecessor, Collin Burns, was pushed out after four days over his ties to Anthropic. The agency, formerly known as the US AI Safety Institute, was renamed under the Trump administration to emphasize standards and innovation over safety enforcement. NIST Director Dr. Arvind Raman will serve as acting director in the interim.

The leadership churn has coincided with an internal struggle over who gets to regulate AI. In May, the Commerce Department quietly deleted a webpage detailing a security-testing agreement with Microsoft, Google, and xAI. A June executive order later restored a lighter version, asking companies to submit new models for review 30 days before release on a voluntary basis rather than mandatory testing.

OpenAI Navigates Federal Gridlock With State-Level Strategy

With federal legislation stalled, OpenAI is pursuing an alternative path. Lehane said the company is pushing states to adopt mirroring AI standards — a strategy he called "reverse federalism" — to create de facto national rules without congressional action. OpenAI is supporting a Massachusetts bill requiring major AI developers to identify catastrophic risks before deployment. Anthropic has also backed the proposal.

The company's Washington engagement comes as it faces its own regulatory scrutiny. OpenAI was required to make significant adjustments to its latest GPT-5.6 model before release under government request, according to reports. Anthropic was forced to temporarily take down its Fable 5 and Mythos 5 models due to US export controls last month, highlighting the unpredictable nature of current enforcement.

For investors, the regulatory trajectory is the key variable. Microsoft, a major OpenAI investor, faces direct exposure to any restrictions on the startup's model releases. Google and Amazon, racing their own AI offerings, would benefit from rules that constrain OpenAI more than their in-house efforts. Nvidia, whose H100 GPUs power most frontier model training, could see demand shift if regulation slows the pace of new model launches.

OpenAI's new model series, expected to demonstrate significant gains in reasoning and task automation, will test whether the government's voluntary review framework can keep pace with the technology. If the agency cannot vet models quickly enough, the gap between what companies ship and what regulators can check will only widen.

This article is for informational purposes only and does not constitute investment advice.