OpenAI Annualized Revenue Nears $70 Billion

Estimated reading time: 5 minutes · Last updated: 29 September 2026

OpenAI's annualized revenue is on pace to reach nearly $70 billion, up roughly 70% from the beginning of the third quarter, driven by sharply stronger enterprise contracts and faster consumer uptake. Company sources say business-to-business sales have climbed more than 100% over the same interval, and that OpenAI added more consumer revenue in Q3 than it did over the whole of last year. At the same time, public filings and prospectuses from rivals show the sector is still carrying very large losses and planned infrastructure spending, and OpenAI has paused a major model launch for safety testing.

“For anything [regarding safety and alignment], there’s a tradeoff,” Jain said.

Saachi Jain, OpenAI head of safety

What the $70 billion run rate means and where it came from

The nearly $70 billion figure is a run-rate estimate based on recent earnings patterns and was reported to reporters by people familiar with the company’s business activity. The figure is described as an annualized revenue number, a projection that extrapolates current short-period receipts to a full year. That projection reflects two trends flagged by reporters: a strong jump in enterprise contracts and faster consumer monetization during the third quarter.

Sources told Reuters that OpenAI's B2B business is the main driver and has risen more than 100% over the stated period. Separately, the company recorded an unusually large increase in consumer revenue during the quarter — larger than the consumer revenue added across the whole of last year — which lifted the short-term run rate. Those items together explain the sharp jump in the annualized total.

Revenue growth versus costs and why profitability is still unclear

Coverage makes a distinction between headline run rates and underlying profitability. Reporters note that details about OpenAI’s expenses were not immediately clear, so a high run rate alone does not show whether the company is profitable on a GAAP basis or whether margins are widening. Bloomberg’s August reporting that OpenAI had been on track for more than $40 billion earlier this year underscores how rapidly the run rate has moved within months, but it does not disclose expense lines.

Axios reported that sources pointed to the continuing high cost of cloud compute and infrastructure: building and operating models remains expensive, and capital needs will matter for any public valuation. That cost backdrop is one reason some companies in the sector are disclosing large projected spending in IPO documents, even as revenue growth accelerates.

Anthropic and the IPO context: losses, spending plans and timing

OpenAI's revenue headline arrives as competitors move toward public markets. Reuters coverage of Anthropic’s IPO prospectus shows the rival recorded a $42 billion net loss last year and is anticipating $518 billion in spending on cloud, compute and infrastructure over coming years. Those figures underline why investors and underwriters will review both growth and capital intensity when pricing any offering.

Both OpenAI and Anthropic have filed confidential paperwork toward public listings, according to reporting. Media coverage expects Anthropic could list this fall, with some reports saying a listing might occur after the November 2026 midterm elections; OpenAI has told reporters it will not go public this year, citing the need to address safety concerns first. Those timing signals will shape the sector’s near-term public-market scrutiny and valuation benchmarks.

Model safety and product pauses that can slow monetization

Shortly before the revenue item surfaced, the Wall Street Journal reported OpenAI had halted the planned launch of its next model, called GPT-6.1 Astra, because the new build had regressed on alignment tests compared with its predecessor. Saachi Jain, identified in coverage as OpenAI’s head of safety, warned that safety and alignment work involves tradeoffs when tuning models to follow human intent.

A paused launch can delay revenue tied to a new model tier or enterprise features, and it can change the pace at which new product monetization appears in short-period earnings that feed an annualized projection. That tradeoff between faster release and more rigorous safety testing is now an explicit part of OpenAI’s near-term commercial calculus, and it is a factor investors and customers will watch alongside run rates and cost forecasts.

Case for and against continued tailwinds

The bottom line

The run-rate headline places OpenAI among the most rapidly monetizing AI platform providers, driven by both enterprise and consumer revenue gains. But the figure is a projection and reporters note that expense data were not released alongside it, so it does not on its own show profitability. At the same time, sector filings from rivals show very large losses and planned spending, and OpenAI’s pause on a major model release for safety testing highlights the tradeoffs between speed and alignment. Investors, customers and partners will need expense detail and product timelines to judge whether the run rate translates into durable earnings.

What to watch

• Watch for reports of Anthropic's IPO listing or pricing, which coverage says could occur after the November 2026 midterm elections.
• Watch for OpenAI's announcement on whether and when it will resume or schedule the GPT-6.1 Astra launch; no date has been set for that decision.

This article is information, not financial advice. Anyone acting on it should do their own checks.