At a closed-door investor meeting on 14 August 2026, OpenAI CFO Sarah Friar said the company’s enterprise business now generates more revenue than its consumer ChatGPT business — the first time that has been true. The annualised revenue run rate reached $40 billion, roughly double the ~$20 billion at the start of the year. When OpenAI closed its $122 billion round in March, it told investors enterprise was on track to reach parity with consumer by the end of 2026. It got there about six months early.

What happened

The reported numbers from that meeting:

MetricValue
Annualised revenue run rate~$40bn (from ~$20bn in January 2026)
Revenue mix at start of 2026~60% consumer / 40% enterprise
Revenue mix in August 2026enterprise ahead of consumer
Run-rate growth, July+20% month over month
Business customer growth, July+32%

Two caveats belong on these figures. They were disclosed to investors rather than published in audited accounts, and a run rate is a monthly figure multiplied by twelve — it describes a moment, not a year. Neither point undermines the trend, but the trend is what the numbers support, not the precision.

The crossover is not only about enterprise growth. OpenAI began showing ads in ChatGPT’s free and Go tiers , which is what you do when consumer subscription revenue is closer to its ceiling than the enterprise line is to its own.

Why it matters for builders

The product you buy is now the company’s main business. For most of ChatGPT’s life, enterprise customers were paying for an add-on to a consumer product. That relationship has inverted. In practice this means enterprise concerns — SLAs, data residency, model version pinning, procurement terms, deprecation notice periods — now carry real commercial weight inside OpenAI. It is a reasonable moment to renegotiate or to ask for commitments you were previously told were unavailable.

It also raises the stakes on model deprecation. Consumer users tolerate a model changing under them. Enterprise workloads do not. Watch whether version pinning and deprecation policy improve now that the revenue mix rewards it — and in the meantime, do not build on unpinned aliases.

Competitive context matters more than the headline. Reporting through August also indicated OpenAI gaining ground on Anthropic among business users, where Anthropic had held an advantage in coding and agentic workloads. If you are choosing between them, the useful comparison remains task-level rather than vendor-level: see OpenAI vs Anthropic and AI vendor selection .

Read it alongside the compute commitments. $40 billion of run-rate revenue is the other side of the ledger from the $105 billion Nvidia backstop on the Ohio campus . Revenue growth of this shape is what those obligations are underwritten against, which is precisely why both numbers get scrutinised together ahead of an IPO .

Sources

  1. CNBC, “OpenAI CFO Friar tells investors that enterprise business now bigger than consumer” (14 August 2026): https://www.cnbc.com/2026/08/14/openai-cfo-friar-tells-investors-that-enterprise-bigger-than-consumer.html
  2. The Next Web, “OpenAI CFO tells shareholders enterprise revenue has overtaken ChatGPT”: https://thenextweb.com/news/openai-enterprise-revenue-overtakes-consumer-friar
  3. TheStreet, “OpenAI’s CFO notes shift in business revenue”: https://www.thestreet.com/investing/openai-enterprise-revenue-passes-consumer-friar-ipo
  4. Unite.ai, “OpenAI Tells Investors Enterprise Revenue Has Overtaken Its ChatGPT Consumer Business”: https://www.unite.ai/openai-tells-investors-enterprise-revenue-has-overtaken-its-chatgpt-consumer-business/
  5. TechCrunch, “OpenAI is gaining on Anthropic with business users, new data indicates” (20 August 2026): https://techcrunch.com/2026/08/20/openai-is-gaining-on-anthropic-with-business-users-new-data-indicates/

Further reading