Is the US tightening grip on AI? Trump administration slows GPT-5.6 rollout

The US President Donald Trump administration has reportedly asked OpenAI to stagger the rollout of GPT-5.6, signalling a broader push to tighten oversight of advanced AI systems amid growing national security concerns and the global race for artificial intelligence.

The United States government has reportedly asked OpenAI to stagger the public rollout of its upcoming GPT-5.6 artificial intelligence model, marking another sign of Washington's increasing involvement in regulating access to frontier AI technologies.

According to reports, the request was made over national security concerns, with OpenAI expected to release GPT-5.6 through a limited preview rather than an immediate public launch. The model would initially be available only to selected partners, with government approval reportedly required before additional organisations gain access.

The reported change comes as OpenAI prepares its next major upgrade to ChatGPT. While the company had been expected to launch GPT-5.6 in the coming weeks, reports suggest it is also considering postponing the wider public release until next year.

Limited rollout under government oversight

Reports indicate that the Office of the National Cyber Director and the Office of Science and Technology Policy requested the phased rollout. OpenAI CEO Sam Altman is said to have informed employees that the company would begin with a limited preview programme before expanding access.

The identities of the initial enterprise customers have not been disclosed, and it remains unclear how the US government intends to approve future users during the preview period.

OpenAI IPO plans also under discussion

The reported rollout changes come as OpenAI weighs the timing of its planned initial public offering (IPO).

The company confidentially filed for a US IPO earlier this month and is reportedly targeting a valuation of up to $1 trillion. However, reports suggest advisers have discussed delaying the listing until 2027 to better support that valuation, while an earlier listing could require a lower valuation. CEO Sam Altman has reportedly opposed lowering the company's valuation target.

Anthropic restrictions highlight broader policy shift

The reported request to OpenAI follows recent US restrictions on rival AI company Anthropic.

Earlier this month, Anthropic suspended access to its advanced Mythos and Fable AI models for foreign users after receiving export control directives from the US government. The restrictions reportedly affected even some of the company's non-US employees.

The move reflected growing concerns within Washington that the world's most capable AI systems could have significant national security implications if widely distributed without safeguards.

US and India discussing advanced AI access

At the same time, the United States and India are reportedly holding discussions on the future availability of advanced AI models, including Anthropic's restricted systems.

US officials have described the talks as sensitive, saying both countries are working towards a gradual and secure approach to sharing frontier AI technologies while balancing innovation with national security.

Bigger picture

If the reports are accurate, the request to stagger GPT-5.6's release would represent another step in the US government's evolving strategy to manage the development and international distribution of advanced artificial intelligence.

Rather than focusing only on export controls for semiconductors, policymakers now appear to be paying closer attention to the deployment of cutting-edge AI models themselves, as governments seek to balance technological leadership with security concerns.