OpenAI Hits $20 Bn ARR Mark as Compute Capacity Triples: CFO Sarah Friar

OpenAI’s annualised revenue has surged past $20 billion in 2025, up from $2 billion in 2023, as the company rapidly expands its compute capacity, according to a new statement by Sarah Friar, chief financial officer of OpenAI.

In a company blog post, Friar said OpenAI has structured its business model so that revenue growth increases in step with the practical value its AI systems generate, tying financial performance directly to the amount of real-world work carried out using its technology.

Compute capacity has grown roughly threefold year over year, reaching about 1.9 gigawatts in 2025, compared with 0.2 GW in 2023, while revenue expanded at a similar pace to exceed $20 billion ARR.

“Our ability to serve customers—as measured by revenue—directly tracks available compute,” Friar wrote, adding that greater access to compute in earlier years would likely have driven even faster adoption and monetisation.

OpenAI said both daily and weekly active users are at all-time highs, driven by ChatGPT’s transition from a consumer curiosity to what Friar described as “infrastructure that helps people create more, decide faster, and operate at a higher level.”

Initially launched as a research preview, ChatGPT is now embedded in everyday personal and professional workflows, from education and writing to software development, marketing, and finance. That usage shift shaped OpenAI’s commercial strategy, starting with consumer subscriptions, expanding to team and enterprise plans, and adding usage-based pricing for developers through its API platform.

“As AI moved into teams and workflows, we created workplace subscriptions and added usage-based pricing so costs scale with real work getting done,” Friar said.

More recently, OpenAI has extended its model to advertising and commerce, positioning ChatGPT as a decision-making platform where users move from exploration to action. Friar stressed that ads and commercial options are only introduced when they are “clearly labelled and genuinely useful,” arguing that monetisation must feel native to the product experience.

At the core of OpenAI’s financial strategy is compute management. Friar called compute “the scarcest resource in AI,” noting that OpenAI has moved from reliance on a single provider to a diversified ecosystem of partners. 

In January, OpenAI signed a $10-billion deal with chipmaker Cerebras Systems, turning its focus to inference infrastructure.

Looking ahead to 2026, Friar said OpenAI’s financial focus will be on practical adoption, particularly in health, science, and enterprise use cases where improved intelligence can directly translate into measurable outcomes. She also signalled future revenue models beyond subscriptions and APIs, including licensing, IP-based agreements, and outcome-based pricing, as AI expands into areas such as drug discovery, energy systems, and financial modelling.