Sources: Sam Altman told staff that OpenAI is generating $1.3B in revenue per year, implying the company is now generating $100M+/month, up 30% in three months
ChatGPT maker OpenAI is generating revenue at a pace of $1.3 billion a year, CEO Sam Altman told staff this week, according to several people with knowledge of the matter.
Context & Ripple Effects
This report puts a concrete commercial marker on OpenAI's early ChatGPT-era business: annualized revenue had risen about 30% in three months, making monetization—not just product attention—a central measure of its progress.
The figure became a near-term baseline for subsequent coverage of OpenAI reaching a $1.6B annualized revenue pace, while much later reporting tied growth to API demand. The arc matters because revenue scale and the cost of serving AI workloads must advance together.
First-order effects
- OpenAI gains evidence of a sizable, fast-growing revenue base, strengthening its ability to treat paid AI products as a core business rather than an experimental add-on.
- Staff and prospective customers get a clearer signal that ChatGPT-related offerings are converting demand into recurring spend at meaningful scale.
Second-order effects
- Other AI developers selling model access or assistants face greater pressure to show paid adoption and durable revenue, rather than relying on user interest alone.
- The reported run rate raises the importance of AI unit economics: growth in paid usage is valuable only to the extent that serving those workloads can be supported efficiently.
Third-order effects
- If comparable growth persists, frontier AI competition will increasingly be organized around the ability to convert inference capacity into recurring revenue, not solely around model launches.
- The durable uncertainty is whether revenue growth can keep pace with serving costs; that balance will help determine which AI business models can sustain scale.
The trend: This is an early data point in the commercialization of AI compute, where recurring revenue must increasingly justify the cost of delivering model output.