OpenAI is putting more weight on delaying the timing of its initial public offering to 2027.

On June 26 (local time), major foreign media outlets including Cryptopolitan reported that OpenAI executives are considering a 2027 listing instead of the late-2026 IPO previously discussed.

The key issue is valuation. Inside OpenAI, discussions have included waiting until 2027 to aim for a $1 trillion valuation, and listing first at a lower valuation in late 2026. But Sam Altman (샘 알트먼), OpenAI's CEO, did not back away from the $1 trillion target, and Sarah Friar (세라 프라이어), the CFO, reportedly has supported a delay to 2027 internally.

OpenAI's desire to push back its IPO also reflects its funding structure and disclosure burdens. OpenAI has compute infrastructure commitments totaling $600 billion through 2030, and it is also burning large amounts of cash. Readiness to shoulder public-company level reporting obligations is also cited as a variable.

Market conditions are also having an impact. OpenAI's most recent private investment round valued the company at about $850 billion. To be valued at $1 trillion in an IPO would require additional investment enthusiasm or revenue growth. But recent SpaceX share performance has increased the burden around listing highly valued technology companies.

SpaceX listed at $135 per share in a $75 billion offering and at one point rose above $225, taking its market value above $2 trillion, but it did not sustain the rise. As of June 25, the stock was trading around $153, down about 25 to 30 percent from its peak. Against this backdrop, market participants have begun re-examining IPO market demand for listings by large AI companies.

Banks view the company that lists first between OpenAI and Anthropic as setting the benchmark for a new AI IPO market. But OpenAI has effectively closed off the option of listing first at a lower valuation. After Altman told executives he would not go below the $1 trillion target, the remaining choice narrowed to continuing private fundraising while waiting for conditions in the IPO market to improve.

Separately from the IPO issue, OpenAI has also faced constraints on how it releases its next-generation model. The White House requested on June 25 that GPT-5.6 be provided first only to a small number of government-approved partners before it is widely distributed. The Office of the National Cyber Director and the Office of Science and Technology Policy were involved in the request.

Altman told employees in an internal Q&A on June 24 that the government would approve access by customer during the preview period. In a separate memo, he said he hoped to broaden availability within a few days after the initial preview. A feature is that instead of a blanket ban on overseas access, the approach involves approving corporate customers one by one.

This measure also connects with Anthropic's approach to controlling its models. The White House had previously restricted access to Anthropic's Claude Mythos and Fable 5 through a Commerce Department export control order, and Anthropic had to disable the two models entirely. By contrast, a more relaxed approval system was applied to GPT-5.6. In connection with this, Commerce Secretary Howard Lutnick (하워드 러트닉) reportedly discussed the matter directly with Altman and sought to verify whether testing and approvals by relevant government departments had taken place.

In this situation, Anthropic is increasing the likelihood it enters the IPO market before OpenAI. Anthropic submitted a confidential S-1 to the U.S. Securities and Exchange Commission on June 1 and is targeting a Nasdaq listing in October 2026. Its target valuation is $965 billion. Goldman Sachs, JPMorgan and Morgan Stanley are serving as lead underwriters.

Some political variables have also eased. U.S. President Donald Trump said on June 19 that he no longer views Anthropic as a national security threat. If OpenAI indeed tilts toward 2027, the likelihood increases that Anthropic lists first and secures the benchmark for the AI IPO market.

Anthropic's recent competitive indicators are also drawing market attention. Over the past 30 days, Polymarket reflected a 94.8 percent probability of being the top model, and 2028 profit-and-loss projections anticipated reaching break-even. By contrast, OpenAI was projected to post a $74 billion loss in 2028. As the IPO timing, valuation and model-release regulations intersect, competition between the two companies appears to be spreading into capital markets and the policy arena at the same time.