Skip to main content

OpenAI’s annualized revenue has surpassed $20 billion in 2025, more than tripling from $6 billion in 2024, as rapid user growth and a sharp expansion in computing capacity fuel the artificial intelligence company’s scale-up, Chief Financial Officer Sarah Friar said in a blog post on Sunday.

Friar said OpenAI’s computing capacity increased to 1.9 gigawatts in 2025 from 0.6 gigawatts a year earlier, closely tracking revenue growth as demand for AI models continues to surge. The Microsoft-backed company’s weekly and daily active user numbers are also hitting record highs, underscoring the growing adoption of ChatGPT and its developer tools.

The revenue milestone comes as OpenAI steps up monetization efforts to support the soaring costs of building and running large-scale AI systems. Last week, the company announced it would begin testing advertisements in ChatGPT for some U.S. users on its free tier and lower-priced Go plan, marking a shift beyond its traditional subscription-based model.

Friar said OpenAI’s platform now spans text, images, voice, code and application programming interfaces (APIs), with the next phase of development focused on AI “agents” and workflow automation. These systems are designed to run continuously, retain context over time and take actions across multiple tools, moving beyond simple question-and-answer interactions.

Looking ahead to 2026, OpenAI plans to prioritize what Friar described as “practical adoption,” particularly in sectors such as healthcare, scientific research and enterprise software, where AI tools can be embedded into everyday workflows.

She also emphasized that OpenAI is maintaining a “light” balance sheet by partnering rather than owning most infrastructure outright. The company is structuring contracts to remain flexible across different cloud providers and hardware types, a strategy aimed at managing risk while scaling rapidly.

Separately, Axios reported on Monday that OpenAI policy chief Chris Lehane said the company is on track to unveil its first hardware device in the second half of 2026, signaling a possible expansion beyond software and cloud-based services.