OpenAI brought in $10 billion in revenue in the third quarter, The Information reports. Its losses grew over the same period. Both facts are true at once: OpenAI is one of the fastest-growing companies in tech, and also one of the most expensive to run.
The headline from The Information doesn’t break out the size of the loss or what drove the revenue. The pattern is still clear, and it’s worth unpacking.
📌 The Quick Read
- Revenue: $10 billion in a single quarter, according to The Information.
- Losses: Growing, not shrinking, even as sales climb.
- What it means: Demand for OpenAI’s products is real and big. The cost of serving that demand is growing at least as fast.
- Who should care: Investors, enterprise buyers who rely on OpenAI’s APIs, and every rival lab running the same playbook.
📈 How Fast This Grew
For perspective, OpenAI’s full-year revenue for 2025 was projected at around $13 billion. Now it’s doing close to that in a single quarter. Few companies of any kind have scaled this fast.
That growth most likely comes from the usual sources:
- Consumer subscriptions to ChatGPT’s paid tiers
- Enterprise and team plans sold to businesses
- API usage from developers building OpenAI’s models into their own products
The Information’s headline doesn’t give the mix. What matters is that the top line has gotten too big for anyone to call AI demand hype.
🔥 Why Losses Keep Growing
This is the part that stands out. Normally, as a software company grows, its margins get better. Code costs almost nothing to copy, so each new customer is mostly profit.
Frontier AI doesn’t work that way, for three reasons:
- Compute isn’t cheap to copy. Every ChatGPT answer and every API call runs on expensive GPUs. More users means more chips, more power and more data center capacity.
- Training keeps getting pricier. Each new generation of models costs more to build than the last, and that spending happens long before it earns a dollar.
- Infrastructure commitments are huge. OpenAI has signed multi-year deals for data center capacity and chips worth hundreds of billions of dollars. Those bills land whether revenue keeps up or not.
Put simply, OpenAI is spending ahead of demand. It’s betting that today’s losses buy tomorrow’s dominance.
🧭 Why This Matters for the AI Industry
OpenAI is the reference point for the whole sector. When it posts numbers, everyone else gets measured against them.
For investors, the debate shifts. “Is there demand?” has been answered. The open question is whether AI can ever earn software-like margins, or whether it’ll look more like a capital-heavy utility business.
For competitors like Anthropic, Google and xAI, the message is that scale costs a fortune. Matching OpenAI on revenue probably means matching it on spending too.
For infrastructure players like Nvidia, CoreWeave and Oracle, OpenAI’s growing losses are, in a sense, their revenue. As long as OpenAI keeps spending, their order books stay full.
For businesses building on OpenAI, watch pricing. A company losing more money as it grows will eventually push hard on monetization, whether through price changes, new premium tiers, ads or tighter usage limits.
🔮 What to Watch Next
- Pricing moves. Expect more experiments with tiered plans and usage-based billing as OpenAI chases margin.
- Efficiency gains. Cheaper inference through custom chips, smaller models and better serving tech is the fastest way to close the gap.
- Fundraising and financing. Losses this size need constant capital. More funding rounds, debt deals or structured partnerships are likely.
- IPO chatter. Numbers like these will feed the ongoing talk about when OpenAI might go public, and how public markets would value a company growing this fast while losing this much.
A $10 billion quarter shows how much people want AI. The growing losses show what it costs to supply it. How long OpenAI can keep both lines going up together is the biggest open question in the industry. The full financial breakdown is in The Information’s original report.