OpenAI has reached a major shift in its business model, with enterprise customers now generating more revenue than its consumer-focused ChatGPT business, according to finance chief Sarah Friar.
During a meeting with investors on Friday, Friar said enterprise growth had accelerated faster than expected, allowing business revenue to overtake consumer revenue ahead of the company’s previous forecast.
Enterprise becomes OpenAI’s largest revenue source
OpenAI began the year with an estimated 60%-40 revenue split favoring consumer products, but enterprise adoption has grown rapidly throughout 2026.
“The majority of our revenue is now enterprise,” Friar said, according to a person who attended the meeting.
The development puts OpenAI ahead of its earlier expectations. The company had previously projected that enterprise and consumer revenue would reach parity by the end of 2026.
OpenAI’s annualized revenue run rate has now reached $40 billion, with the company’s revenue increasing by 20% month over month in July. Business customers grew even faster, increasing by 32% during the same period.
Leadership changes during a period of growth
The investor meeting was scheduled before a series of executive departures that created uncertainty around OpenAI’s leadership team.
Revenue chief Denise Dresser stepped down after eight months in the role. Dresser, who previously served as CEO of Slack and spent more than a decade at Salesforce, said she was leaving to pursue new opportunities.
Longtime OpenAI executive Brad Lightcap also announced his departure after eight years with the company, saying he planned to start something new.
OpenAI president and co-founder Greg Brockman attended the investor meeting and highlighted Dresser’s contributions to building the company’s enterprise foundation.
Dali Rajic, former chief operating officer at cybersecurity company Wiz, will replace Dresser as OpenAI’s new revenue leader.
OpenAI addresses competition and IPO questions
Executives also faced questions about competition from open-source AI models developed in China and the possibility of an initial public offering.
Brockman reportedly pushed back on concerns around open-source competition, arguing that the assumption that open-source models are always cheaper is misleading.
Executives declined to discuss IPO timing because the company has confidentially filed documents with the U.S. Securities and Exchange Commission.
Enterprise customers focus on AI efficiency
Friar said enterprise customers are changing how they use artificial intelligence, moving away from unrestricted AI spending toward measuring the cost and value of AI-generated work.
She described the shift as a move away from “tokenmaxxing,” where companies allowed employees to use large amounts of AI computing without closely tracking productivity gains.
“Enterprise customers have moved from tokenmaxxing to focusing on cost per unit of intelligence,” Friar said.
She highlighted improvements in OpenAI’s latest models, including a reported 54% efficiency improvement for agentic coding tasks, along with recent price reductions across the company’s model offerings.
Advertising becomes another growth opportunity
Friar also said OpenAI has made significant progress in advertising, with the business approaching a $1 billion annual revenue run rate.
The company began testing advertisements within ChatGPT earlier in 2026 as it explores additional monetization opportunities beyond subscriptions and enterprise services.
