Anthropic’s revenue passed OpenAI’s for the first time last quarter

OpenAI grew more slowly than Anthropic last quarter, and for the first time its rival’s sales came in higher. OpenAI told investors its revenue rose 18 percent from the first quarter to the second, to $6.7bn, The Wall Street Journal reported. Its losses deepened over the same stretch. Berber Jin and Corrie Driebusch wrote that the numbers disappointed some shareholders.

The gap with Anthropic is the story. Anthropic more than doubled its revenue to $11.6bn in the same period, the Journal reported. It was the first time its sales have passed OpenAI’s. Anthropic also swung to a small operating profit. OpenAI moved the other way.

OpenAI’s revenue rose to $6.7bn from $5.7bn in the first quarter, the Journal said. Its operating loss, which includes stock-based compensation, widened from $9.3bn to $12.3bn. That means the loss grew by $3bn while revenue added only $1bn, Jin noted in a post on X.

The margin sank further into the red just as the company prepares to go public.

Why a $7bn quarter still disappoints

For most startups, nearly $7bn in quarterly revenue would be remarkable. The bar for OpenAI is different. The company has sold investors on a stratospheric pace of growth, the Journal reported. It has signed large computing deals premised on generating hundreds of billions of dollars a year. The performance of Nvidia, Oracle and other tech giants hinges on OpenAI meeting those commitments.

By that yardstick, the quarter fell short. OpenAI’s sequential growth trailed Palantir over the same period, the Journal said. It also lagged other AI highfliers such as CoreWeave and Micron. Anthropic, by contrast, told investors it had made progress in how efficiently it uses computing resources.

Jin framed the exclusive bluntly. “OpenAI has been tossing out a lot of vague ARR numbers, so we decided to take a deeper look,” he wrote on X. The company, he added, “grew revenue by just 18 percent… while its losses sank further into the red.”

OpenAI says the growth is turning

OpenAI pushed a more upbeat account to investors. Its growth rate has picked up since it launched a set of new models in July, the people told the Journal. Revenue from business customers grew 32 percent in July from the month before, the New York Times reported. That outpaced its overall run rate.

Chief financial officer Sarah Friar told investors most of the company’s revenue now comes from business customers.

The company is also reshaping its product. It recently released a “super app” that folds its Codex coding tool, ChatGPT and a web browser together. The company says the product is drawing users fast. Co-founder and president Greg Brockman has taken a more active role over the product and business teams, the Journal reported, in a push to reaccelerate growth.

The push follows a bruising stretch for its leadership. OpenAI last week replaced its chief revenue officer, Denise Dresser, after less than a year, the Journal said. Her exit followed that of former chief operating officer Brad Lightcap. It also followed the departure of Fidji Simo, once seen as a possible successor to chief executive Sam Altman.

Two IPOs, two very different pitches

The results land as both labs prepare to list. Anthropic could go public as soon as this autumn, according to the Financial Times, cited by the Times. Investors are targeting a $2tn valuation. OpenAI is likely to follow next year. Both need to show strong growth to investors as their costs are set to climb for years.

Anthropic is already lining up the machinery. It is setting up a multi-class share structure that hands its founders outsized control, according to The Information, and expanding its credit line, according to Bloomberg. Its reported revenue carries a caveat, though.

The Journal said it is unclear how Anthropic calculated its adjusted profit, and noted the company has previously excluded stock-based compensation from that figure. Its preliminary revenue has also been reported at $11.5bn by Bloomberg, a shade below the Journal’s $11.6bn.

A pause, and a skeptical crowd

The financials arrived days after OpenAI said it was slowing down. Altman wrote on X that the company had “paused some frontier RL training to ensure that we can meet the appropriate alignment, security and monitoring standards for the new level of capabilities in front of us.” He said model progress was “extremely rapid.”

The desk has covered that pause, which followed a test in which OpenAI’s agents hacked other companies.

Not everyone took the stated reason at face value. “There’s no way this is the real reason right,” one user replied. Another, Ross Hendricks, offered his own reading: “we need to immediately stop torching cash to provide some semblance of a sustainable business model so we can rush this IPO out the door.”

OpenAI has given a safety rationale for the pause. The timing, next to the widening losses, is what the skeptics seized on. The desk takes no view on which reading is right.

The pressure underneath

Part of the squeeze is competitive. OpenAI subsidises hundreds of millions of users who do not pay for ChatGPT, the Journal reported, and it cut prices on two of its latest models after corporate customers grew cautious and shifted tasks to cheaper Chinese models. Anthropic has had to calm the same investor nerves about those models.

For now, the two companies are telling opposite stories into the same IPO window. Anthropic points to a doubling of revenue and a first profit. OpenAI points to July’s reacceleration and a growing base of business customers. The quarter itself, though, put Anthropic’s sales ahead of OpenAI’s for the first time, and left OpenAI explaining a loss that grew faster than its revenue.