OpenAI’s $20 Billion Revenue Problem Mitchell Duran Sun, October 11, 2026 at 8:02 AM EDT 3 min read OPAI.PVT NVDA -0.52% ANTH.PVT QQQ +0.49% Explore stocks on Coinbase Trading disclosure Trading disclosure The above button links to Coinbase. Yahoo Finance is not a broker-dealer or investment adviser and does not offer securities or cryptocurrencies for sale or facilitate trading. Coinbase pays us for certain activity generated through this link. Prices displayed are informational.
Our analysts just identified a stock with the potential to be the next Nvidia. Tell us how you invest and we'll show you why it's our #1 pick. Tap here.
Open-weight labs pitching AI sovereignty are climbing the usage charts. Closed-model giants like OpenAI and Anthropic are handing them the opening with mounting compute losses, shaky financials, bad press, and growing public resentment toward the sector and its billionaires. Much lower costs don't hurt either.
As retail users and enterprises build out their AI stacks, they may settle on a mix of open and closed models. Or the race could turn zero-sum, with every gain for one side coming straight out of the other.
The biggest tell that all isn't well in AI land: OpenAI, which is in talks to raise at a $1.4 trillion valuation, told investors its annualized revenue was approaching $50 billion, per the Financial Times.
That's about $20 billion below the figure investors had pieced together from its own disclosures. OpenAI never denied the higher number while it circulated, and it declined to comment to the FT. Tech stocks didn't wait. Oracle fell 5.5% and Nvidia 2.9% after the story broke.
This is the same company that expects to burn $278 billion in cash between 2026 and 2030. The expense side is worse. OpenAI expects to spend $856 billion on compute and AI infrastructure through 2030, all while promising (sorry, projecting) that revenue will climb from $36 billion this year to $350 billion in 2030. That's a tenfold jump in four years, adding up to $840 billion in cumulative revenue.
One stock. Nvidia-level potential. 30M+ investors trust Moby to find it first. Get the pick. Tap here.
OpenAI told the world it alone would change it. Making money while doing so is clearly a problem. So whose fault is it? Open-weight models? Partly.
Chinese labs are taking a growing share of token volume. In the last week of September, DeepSeek alone processed 36.8 trillion tokens on OpenRouter, according to a Social Capital chart. That's more than OpenAI, Google, Anthropic, and xAI combined, at 33 trillion. Chinese models first overtook American ones on the platform in February.
OpenAI is fighting back on volume. Its weekly OpenRouter tokens hit 19.5 trillion by late September. It paid for that growth with price cuts. OpenAI slashed GPT-5.6 Luna by 80% and Terra by 20% in July, then halved GPT-6 prices in September. OpenAI says efficiency gains covered the cuts. Its $278 billion projected burn says otherwise.
Worse, Ramp's September AI Index found that the top 1% of AI spenders cut spending per employee by nearly 10% in August. Much of that drop comes from Ramp revising July's figure upward. Still, the direction isn't what AI bulls want to see. Ramp economist Ara Kharazian doesn't blame open models. He blames a price war among the model makers and a shift to their cheaper models. That matters most for OpenAI, whose enterprise business now brings in more revenue than its consumer side, CFO Sarah Friar told investors at a closed-door meeting on August 14.
OpenAI now expects to hit $70 billion in annualized revenue by year-end, per Bloomberg, the number investors thought it had already reached. Meanwhile, it has pushed its IPO past 2026, citing AI safety. It reads more like "please don't look at our books."
Enterprises still prefer closed models. Only 6.4% of AI-spending businesses on Ramp use open-source ones. But Anthropic, not OpenAI, leads business adoption, and OpenAI doesn't expect positive cash flow until 2030.
Open models are winning on volume. Closed models are winning on dollars and still losing money. How long can that last? The U.S. economy is betting on the answer.
