OpenAI Triples Computing Power and Revenue
OpenAI's CFO announces over $20 billion in annual revenue, pointing to increased computing power, which remains crucial.
(Image: Shutterstock/ioda)
OpenAI, the developer of ChatGPT, more than tripled its annual revenue last year, surpassing the $20 billion mark. In 2024, the company's revenue was still six billion US dollars. These figures were stated by Sarah Friar, OpenAI's Chief Financial Officer, on Sunday in a blog post.
The expansion of computing capacity, in particular, fueled revenue growth. OpenAI's available computing power also increased threefold year-on-year, or 9.5-fold from 2023 to 2025: from 0.2 gigawatts (GW) in 2023, to 0.6 GW in 2024, and to around 1.9 GW in the past year. Both weekly and daily active user numbers continued to reach record highs, according to Friar.
“Investments in computing power enable cutting-edge research and groundbreaking advances in model capacity. More powerful models enable better products and broader adoption of the OpenAI platform. Adoption drives revenue, and revenue funds the next wave of computing power and innovation. The cycle reinforces itself,” Friar outlined the success formula.
ChatGPT with Advertising
However, what sounds like an unmitigated success story reveals some pitfalls upon closer inspection. At the end of last week, OpenAI announced that it would begin testing advertising in its AI chatbot ChatGPT starting this week to generate more revenue from ChatGPT and thus cover the high costs of AI development. Initial experiments with this new financing model are to be limited to the United States for now.
Financing through advertising represents a significant change in direction for OpenAI. As recently as October 2024, OpenAI CEO Sam Altman had reiterated his personal aversion to advertising. But there had been indications of a rethink for some time. In early December 2025, Altman declared “Red Alert” at OpenAI due to growing competitive pressure. Side projects like working on advertising were to be postponed for now. But considering the company's billion-dollar losses, this renewed shift likely occurred.
Access to Computing Power
The next stage of development will go beyond subscriptions and advertising, Friar announced. With the integration of AI into scientific research, drug development, energy systems, and financial modeling, new economic models are emerging. “Licensing, IP-based agreements, and outcome-based pricing will share the value created,” said the OpenAI CFO.
Access to computing power is crucial for future development. “Computing power is the scarcest resource in AI,” Friar stated. “Access to computing power determines who can scale.” OpenAI manages this by keeping its balance sheet “lean,” entering into partnerships rather than acquisitions, and structuring contracts with flexibility across vendors and hardware types.
Videos by heise
Focus on Practical Implementation
OpenAI's product platform currently includes text, images, speech, code, and APIs. “The next phase is agents and workflow automation that run continuously, transfer context over longer periods, and take action across tools,” Friar ventured a look ahead. For individuals, this means AI that manages projects, coordinates plans, and executes tasks; for organizations, it will become an operating layer for knowledge work.
For the current year 2026, OpenAI will focus on “practical implementation,” Friar announced. Major and immediate opportunities lie particularly in the areas of health, science, and business. “Infrastructure expands what we can do. Innovation expands what [artificial] intelligence can do. Adoption expands the circle of those who can use it. Revenue funds the next leap. This is how [artificial] intelligence scales and becomes the foundation of the global economy.”
(akn)