OpenAI IPO pressure builds as sales growth slows to 18%

OpenAI IPO prospects face fresh scrutiny after quarterly sales growth cooled to 18% and losses deepened, while Anthropic widened its lead.

OpenAI said quarter-on-quarter sales growth slowed to 18%, a weaker pace than earlier growth phases implied by the company’s fundraising narrative. The company also reported deeper losses, adding a harder cost question to an offering expected to rank among the largest public listings.

18% growth changes the IPO pitch

The figures shift attention from demand for artificial intelligence tools to the cost of supplying them. For OpenAI, the issue is not whether revenue is growing; it is whether that growth can support the computing, talent and product costs behind frontier AI models.

Anthropic’s reported annualized revenue topped $65 billion, more than seven times its level at the end of last year and more than 50% above OpenAI’s comparable figure. That comparison gives investors a direct rival against which to test OpenAI’s growth rate, pricing power and path to profitability.

Annualized revenue is a run-rate measure, not audited full-year sales. It can move quickly when subscription demand, enterprise contracts or usage-based billing changes, which makes the figure useful for sizing momentum but less definitive than reported annual results.

Anthropic’s $65 billion run rate

The divergence matters as both companies move toward public markets with business models still being tested at scale. OpenAI enters that process with a larger public profile, but Anthropic’s reported run rate gives it a cleaner growth comparison at a moment when losses at OpenAI are drawing closer attention.

The companies are also competing in a market where model quality no longer moves in one direction. Some Chinese competitors are described as offering models only marginally behind leading US systems while charging at a wide discount, putting pressure on the premium tier of the market.

An unnamed expert quoted in the source material summarized the shift this way: “The US labs have cut the middle and are defending the top.” The phrase points to a market structure in which expensive frontier models compete for the highest-value customers while cheaper models compress the pricing room below them.

Chinese discounts test AI pricing

For OpenAI, the near-term effect is a narrower IPO argument. Investors can still pay for growth, but slowing sales expansion and deeper losses make the quality of that growth more important than the headline size of the company’s user base.

For the wider AI sector, price competition changes the economics of model development. If lower-cost models keep improving, buyers may use them for routine tasks and reserve premium US systems for work where accuracy, safety controls or enterprise support justify a higher price.

The global macro channel runs through capital spending and investor risk appetite rather than a single consumer price measure. AI labs, cloud providers and chip suppliers are tied together by large computing requirements, so a weaker pricing outlook for model providers can change how quickly infrastructure spending is financed.

Three paths for public markets

If OpenAI stabilizes growth above the latest 18% quarter-on-quarter pace and narrows losses before listing, the IPO case would rest on scale and operating discipline. That path would support continued AI infrastructure spending, help OpenAI defend premium pricing and give the wider sector more room to fund expensive model development.

If growth slows further while losses keep widening, public investors may push for a lower valuation or a delayed offering. That would tighten the company’s financing options, cool risk appetite around other AI listings and make cloud and chip demand more dependent on a smaller set of well-funded customers.

If Chinese discounts keep spreading across enterprise software and developer tools, the pressure would fall first on pricing rather than usage. OpenAI would need to prove that its top-tier products deserve a premium, the sector would face thinner margins, and the macro effect would be a more selective AI investment cycle rather than a broad pullback.

The main open question is whether customers keep paying premium prices as cheaper models close part of the performance gap. The answer will shape OpenAI’s IPO timing, Anthropic’s valuation case and the amount of capital the AI industry can keep drawing into frontier computing.