OpenAI posted a net loss of about $38.53 billion in 2025, according to audited financial documents reviewed by the Financial Times, a sharp jump from roughly $5.09 billion in 2024. The figures offer a rare look at just how expensive the company’s breakneck expansion has become as it prepares for a possible IPO, and they also show a business where spending is still outrunning revenue by a wide margin.
Revenue reached $13.07 billion in 2025, but total costs climbed to $34 billion, leaving little room for optimism if investors are expecting near-term profit. That mismatch is familiar in AI, where infrastructure, talent, and model training burn cash fast; the difference here is the scale. OpenAI is no longer a research lab with a breakout product. It is a capital-hungry platform trying to behave like a public company before it actually becomes one.
R&D spending dwarfed everything else
The largest cost item was research and development at $19.18 billion, followed by sales and marketing at $5.73 billion and administrative expenses at $1.57 billion. The company’s operating loss reached $20.92 billion, which is the kind of number that makes even aggressive growth stories look a little sweaty.
For comparison, OpenAI’s 2024 revenue was $3.7 billion against $12.48 billion in total expenses. The jump in losses over one year was nearly eightfold, and that is before the usual AI-industry excuses about compute costs and frontier-model arms races are even considered. Competitors such as Anthropic and Google DeepMind are spending heavily too, but OpenAI’s public-facing ambitions make its burn look especially exposed.
Microsoft remains the biggest financial counterpart
Microsoft sat at the center of the cost structure. OpenAI paid about $10.59 billion to the company for R&D services, while related obligations reached $17.2 billion. At the same time, Microsoft’s payments to OpenAI totaled around $303 million, underscoring how lopsided the relationship remains.
SoftBank also contributed about $867 million in investments. Add in the company’s structural shift toward a more commercial setup and the revaluation of convertible obligations, and the headline loss starts to look less like one clean number and more like a pile of accounting and expansion effects stacked on top of each other. That does not make the loss smaller; it just explains why the bottom line looks even uglier than the operating figures alone suggest.
OpenAI’s cash position leaves room for expansion
OpenAI ended the year with assets above $50 billion, with a significant portion in cash. That gives the company breathing room, and plenty of it, but public-market investors usually ask a less polite question than private backers do: how long can growth outrun monetization before the bill arrives?
The answer will probably depend on whether OpenAI can keep converting product demand into reliable revenue faster than its infrastructure, partnership, and talent costs keep rising. If it cannot, the IPO story shifts from ”how big can this get?” to ”how much cash does it burn on the way there?”