OpenAI is considering delaying its initial public offering until 2027 as chief executive Sam Altman remains unwilling to compromise on the artificial intelligence company’s targeted valuation of as much as $1 trillion, according to a report.

The ChatGPT maker’s advisers have presented executives with two options: wait until next year and pursue a $1 trillion valuation, or lower its valuation expectations to list sooner, the New York Times reported, citing three people involved in the deliberations.

Altman reportedly told advisers that any reduction in the trillion-dollar valuation target was a “non-starter”.

OpenAI has already confidentially filed for an initial public offering in the United States. Chief Financial Officer Sarah Friar has told some associates that the company is aiming for a 2027 listing, Reuters previously reported.

A delayed debut would give OpenAI more time to strengthen its financial position and convince investors that its rapid revenue growth can justify one of the largest valuations ever sought by a newly listed company.

However, it would also extend the wait for investors seeking exposure to one of the biggest beneficiaries of the generative AI boom. OpenAI has emerged as a central player in the sector since the launch of ChatGPT, but the company also faces high spending requirements as it builds computing infrastructure and develops increasingly powerful models.

The possible IPO delay comes as OpenAI faces greater scrutiny from the US government over the release of its newest AI systems.

The administration of US President Donald Trump has asked the company to stagger the rollout of its latest model because of security concerns, a person familiar with the matter told Reuters.

Altman told employees that OpenAI would release GPT-5.6 through a limited preview for selected partners, with the government “approving access customer by customer during this preview period,” according to The Information.

The phased launch was requested by the White House’s Office of the National Cyber Director and the Office of Science and Technology Policy, the report said.

The approach points to growing government involvement in how frontier AI models are released, particularly as their capabilities in areas such as software development, cybersecurity and scientific research improve.