OpenAI CFO says July annualised revenue exceeded all of Q2
OpenAI CFO Sarah Friar told staff that July’s annualised recurring revenue exceeded the whole of Q2, CNBC reports. The claim credits GPT-5.6, ChatGPT Work and Codex but lacks public verification.
OpenAI’s chief financial officer, Sarah Friar, told employees that the company’s annualised recurring revenue in July outpaced the entire second quarter, according to an internal meeting transcript reported by CNBC on July 29, 2026.
That declaration — delivered alongside remarks from board chair Bret Taylor — credits recent product pushes for the momentum, naming GPT-5.6, ChatGPT Work and Codex as drivers. The claim, if sustained, would mark a sharp acceleration in OpenAI’s commercial trajectory, but the company has not published matching public figures to verify the scale of the jump.
Friar credits GPT-5.6, ChatGPT Work and Codex for the lift
Friar and Taylor told staff the revenue acceleration followed releases and enterprise rollouts, with Friar saying, “And Q2 was no slouch,” in the partial transcript cited by CNBC. The company has repeatedly pointed to paid enterprise features and newer model versions as the primary levers for revenue growth; the meeting framed July as the month when those initiatives began to compound materially.
CNBC’s piece is based on what it describes as an internal all-hands and a partial transcript; OpenAI has not supplied a public breakdown of July ARR that would allow outside verification. That matters because independent benchmarks for the company’s quarterly or annualised revenue remain scarce outside of occasional reporting by trade outlets and analyst estimates.
Competition with Anthropic and the strategic timing
The staff meeting also touched on rivalry with Anthropic, a competitor that has been pitching its Claude models to enterprise customers, according to CNBC’s account. OpenAI’s emphasis on ChatGPT Work — its enterprise suite — and model upgrades helps explain the timing: product-led enterprise pushes often deliver sharp near-term revenue effects when billing and adoption inflect together.
Still, the competitive picture is mixed. Anthropic has publicly and privately sought to position its models as differentiated on safety and customer control, while rivals including Microsoft and Google continue to bundle AI capabilities into broader cloud and productivity offerings. CNBC’s account does not include a counter-statement from Anthropic.
Questions remain about scale, margins and public verification
Beyond the headline claim, key details are missing: CNBC’s report does not quote a specific dollar figure for Q2 revenue or for July ARR, and OpenAI has not filed public financial statements that would confirm the trajectory. DatacenterDynamics reported separately on OpenAI’s statements about compute scaling and revenue speed, noting the company’s comments about ending 2025 with higher compute usage, but that reporting likewise relies on company disclosure rather than audited filings.
Analysts and industry watchers caution against reading an internal remark as conclusive financial proof. “Internal ARR statements can reflect forward-looking bookings or annualised run-rates that differ materially from GAAP revenue,” said a former cloud CFO who asked not to be named, adding that public-market scrutiny typically demands audited quarters and clear definitions of ARR.
OpenAI has previously resisted traditional public disclosure; the firm remains private and has shifted governance and leadership roles in recent months, including the consolidation of certain responsibilities under its finance team, which Friar heads alongside other senior leaders.
The company’s internal optimism sits against uncertainty about profitability and long-term costs. Reports and secondary estimates circulated in trade outlets have offered widely varying ARR guesses for 2025–26, underscoring how opaque private financials remain outside company-controlled disclosures.
Looking ahead, investors and competitors will watch for any formal financial disclosure or a regulatory filing that pins down Q3 and year-to-date revenue trends. A public quarterly update, an S-1 filing, or audited results would provide the clearest test of whether July’s internal annualised run-rate represents a sustained step-change or a momentary spike tied to product timing.
