OpenAI's Enterprise Business Overtakes Consumer as CFO Friar Courts Investors Ahead of IPO

For most of its history, OpenAI’s financial identity has been simple: a consumer app company with a rocket attached. ChatGPT’s hundreds of millions of subscribers and paying Plus users generated the lion’s share of revenue, while the enterprise business — API customers, ChatGPT Enterprise seats, and dedicated cloud deals — grew fast but from a smaller base. On Friday, August 14, 2026, that picture officially inverted. Speaking to investors, OpenAI CFO Sarah Friar said the company’s enterprise business is now bigger than its consumer business by revenue — a quiet but consequential milestone on the road to what is expected to be one of the largest public offerings in market history.

What Friar Told Investors

Friar’s message, delivered in an investor setting and reported by CNBC on August 14, was blunt: enterprise has overtaken consumer as OpenAI’s largest revenue source. That is a first for the company. It means the bulk of OpenAI’s income now flows from businesses — API consumption, ChatGPT Enterprise and Business seats, premium offerings like the recently launched $125-per-user-per-month ChatGPT Business Premium tier, and longer-term contracts with cloud and platform partners.

The timing is not accidental. OpenAI has been laying IPO groundwork for months: a $7 billion employee share buyback at an $852 billion valuation completed earlier in August, a wave of C-suite changes, and a deliberate effort to reframe the company’s revenue mix for public-market investors who prize predictability. A consumer app with churn and fickle usage patterns commands lower multiples than an enterprise software business with multi-year contracts and expanding seat counts. Friar, who joined OpenAI from Nextdoor and previously led finance at Square, has spent her tenure pushing exactly this narrative — and now has the number to back it.

Why the Crossover Happened Now

Three forces converged to drive the crossover.

First, consumer growth matured — and was deliberately reshaped. OpenAI recently made GPT-5.6 Luna the default model for Free and Go users with unlimited text chats, trading near-term consumer monetization for engagement scale. Free users don’t directly add revenue; they add distribution and habit. Meanwhile, the paid consumer tiers have grown more slowly relative to the base than they did during ChatGPT’s explosive first years.

Second, enterprise demand for agents went vertical. 2026 has been the year AI agents moved from demo to deployment, and agents are voracious API consumers. An employee using ChatGPT might send a few dozen messages a day; a single deployed agent fleet can generate millions of API calls daily. OpenAI disclosed that its business customer base has surpassed two million paying users, and agentic workloads — coding agents, customer-support automation, security operations — scale consumption in ways human chat never can.

Third, OpenAI rebuilt its enterprise go-to-market around security. The company replaced its first chief revenue officer, Denise Dresser, with Dali Rajic, the former Wiz president and COO who helped build the cloud-security firm Google bought for $32 billion. The signal was unmistakable: OpenAI intends to sell AI the way Wiz sold security — to risk-averse CIOs, through compliance and trust, at enterprise contract sizes. That approach is beginning to show up in the revenue line.

The IPO Context: A $2 Trillion Shadow Race

The enterprise milestone lands amid a blistering race to public markets. Anthropic, OpenAI’s chief rival, confidentially filed for an IPO in June at a $965 billion valuation, and investors now expect it to target $2 trillion or more when it prices — potentially in October. MarketWatch noted Anthropic’s run-rate revenue stood at roughly $49 billion as of May, with investors expecting it to double by year-end. If Anthropic lists anywhere near $2 trillion, it would eclipse SpaceX’s record $1.77 trillion listing and become the largest IPO ever.

OpenAI, whose $852 billion buyback valuation now looks conservative next to Anthropic’s rumored float, has every incentive to demonstrate that its revenue is not just large but structurally sound. “Enterprise is bigger than consumer” is precisely the sentence a CFO wants in investors’ heads before an S-1 becomes public. It answers the two questions that have haunted AI valuations: is the revenue durable, and does it compound without linear marketing spend?

What It Means for the Market

The crossover has implications beyond OpenAI’s own banker materials.

For the AI software landscape, it confirms that the center of gravity in generative AI revenue is shifting from prosumer subscriptions to B2B workloads. Anthropic has said for over a year that the majority of its revenue comes from business customers — Claude’s strength in coding and agentic tasks made it an early enterprise favorite. OpenAI matching and overtaking that mix means the two frontier leaders now compete head-to-head on the same revenue terrain: enterprise agents, seat expansion, and platform lock-in.

For unit economics, enterprise revenue carries higher gross-margin risk — API pricing has been in a race to the bottom, with Luna at roughly $0.10 per million input tokens and rivals like DeepSeek and Google’s Flash tier pushing prices lower still. Volume is offsetting price, but investors will probe how sustainable margins are when inference costs are falling industry-wide.

For the broader market, two IPOs of this scale — Anthropic potentially at $2 trillion, OpenAI presumably larger — would represent an unprecedented transfer of AI risk into public markets. Retail pension funds and index investors would hold direct exposure to the AI buildout’s economics, including its enormous compute obligations: OpenAI’s compute commitments run into the hundreds of billions of dollars across deals with Oracle, Microsoft, and the Stargate consortium.

The Road Ahead

Friar’s disclosure is a snapshot, not a victory lap. Consumer revenue is still growing in absolute terms, and OpenAI continues to invest heavily in consumer surfaces — including a rumored $300-plus smart speaker designed to act as a smartphone replacement. But the strategic weighting is now explicit: OpenAI sees its future the way Salesforce, Microsoft, and ServiceNow do — as an enterprise platform with a beloved consumer product on top, rather than the reverse.

For an AI industry that has spent three years justifying trillion-dollar private valuations on the strength of chatbot subscriptions, the enterprise crossover is the clearest evidence yet that the business model is maturing. Whether public markets agree — when Anthropic prices this fall and OpenAI eventually follows — will be one of the defining financial events of the decade.