aiexpert
Home / News / Brief
Market · Aug 14, 2026, 07:33 PM · 4 sources

OpenAI Enterprise Revenue Now Exceeds Consumer, $40B Run Rate

OpenAI CFO Sarah Friar told investors Friday that the company's enterprise business has crossed a critical milestone: it now generates more revenue than consumer ChatGPT. The company entered 2026 at a 60-40 enterprise-to-consumer split, but enterprise growth accelerated dramatically, and the lines crossed before end of year. Annualized revenue reached $40 billion with enterprise customers growing 32% month-over-month in July, compared to 20% overall run-rate growth.

The crossover marks a fundamental shift in OpenAI's business mix. Two million enterprise customers doubled over the past year, driven by agentic AI workflows and a move away from what OpenAI calls 'tokenmaxxing' — companies are now focused on cost per unit of intelligence instead of racking up unlimited token spend. The newest models are 54% more efficient on agentic coding tasks, and OpenAI has cut prices across its model suite.

Friar also disclosed advertising is approaching a $1 billion run rate after testing ads in ChatGPT starting in February 2026. This diversification beyond subscriptions comes amid a turbulent week: revenue chief Denise Dresser departed after eight months, replaced by Dali Rajic from cybersecurity firm Wiz (acquired by Google). Longtime operating chief Brad Lightcap also left.

The talent churn and IPO uncertainty (Friar said she could not discuss timing due to confidential SEC filing) contrast sharply with the company's accelerating enterprise traction, which now operates more like infrastructure for enterprises than consumer software.

Sources

Everything this brief rests on
  1. 01 Primary source cnbc.com
  2. 02 cnbc.com cnbc.com “Enterprise has accelerated much faster than expected and those lines have now crossed”
  3. 03 cnbc.com cnbc.com “Denise Dresser stepped down just eight months into the job; Dresser took on added responsibilities when Brad Lightcap left”
  4. 04 cnbc.com cnbc.com “Enterprise customers have moved from tokenmaxxing to focusing on cost per unit of intelligence”