OpenAI doesn’t expect to be profitable until at least 2030 as AI costs surge

As OpenAI and Anthropic move closer to their planned initial public offerings, more details about the finances of both artificial intelligence giants are starting to emerge. It was no secret these companies were bleeding cash, but seeing the actual numbers is still striking.

Neither company has made its filings official. Both are in the process of recruiting investors and have recently closed funding rounds, which meant opening their books. The Wall Street Journal got a peek. According to internal estimates, OpenAI will not turn a profit until 2030, while Anthropic expects slight positive results this year, followed by another year of losses before staying in the green in 2028 and 2029.

Spending on AI training will be staggering. In 2028, OpenAI projects spending $121 billion on computing power for its AI research. The estimate for 2029 is slightly higher, before AI model training costs dip back below $100 million in 2030. (This year, for perspective, the company expects to spend just over $25 billion on AI model training.)

Anthropic’s totals are smaller but still climb steadily, surpassing $30 billion in 2029.

These losses come despite an expected surge in revenue at both companies. OpenAI’s revenue is projected to nearly double annually, reaching roughly $275 billion in 2030. Anthropic expects to approach $150 billion in 2029.

Anthropic’s projections include sales through cloud partners, something OpenAI does not emphasize to the same extent. As a result, Anthropic expects most of its income to come from enterprise customers.

That channel is also key for OpenAI, but the company is betting heavily on consumer usage. It is still unclear how that revenue will break down between paid memberships, advertising, or other streams. (OpenAI projects roughly $150 billion in consumer revenue in 2030.)

For now, OpenAI is effectively subsidizing free users as it pushes adoption. While not stated explicitly, the strategy is clear: build habit now, convert later. Loyal users are easier to turn into paying customers over time.

Usage trends suggest the approach is gaining traction. Data from SimilarWeb shows AI platform visits rose 28.6% between January 2025 and January 2026. Users are also spending more time on these platforms, while referral traffic to external sites, which supply much of the underlying data, has plateaued.

OpenAI’s lead remains significant. As of September 2025, ChatGPT accounted for roughly 79% of global generative AI web traffic, according to SimilarWeb. Google’s Gemini grew 157% between April and September 2025, reaching 1.1 billion monthly visits. Perplexity reached 170 million monthly visits, while Anthropic’s Claude saw 157 million.

Corporate demand is rising alongside consumer usage. A December study from advisory firm Teneo found that 68% of CEOs plan to increase AI spending in 2026. So far, however, fewer than half of corporate AI projects have generated returns that exceed their costs.

The race for users is driving the current arms race in AI, and the rapid cash burn that comes with it. It also helps explain why companies like OpenAI and Anthropic are pushing exchanges to adjust their rules. Nasdaq, it seems, is on board, saying it would allow newly listed companies (like OpenAI and SpaceX) to join its Nasdaq 100 index, which would make it easier for them to sell shares.