- OpenAI is pivoting to enterprise and coding after internal all-hands where executives called Anthropic a ‘wake-up call.’
- Anthropic’s business adoption surged to 33% of US companies in March while OpenAI’s stayed flat at 35%.
- Both companies are preparing IPO filings for late 2026—and the first to list sets the valuation benchmark for every AI company that follows.
OpenAI held an all-hands meeting on March 16 that amounted to a corporate confession. Fidji Simo, the company’s chief of applications, told staff to stop chasing “side quests” and pivot aggressively toward coding and business users. The Wall Street Journal reviewed a transcript. CNBC reported that Simo framed the opportunity bluntly: “We have 900 million users. Our job is to turn them into high-compute users.”
Last year’s product blitz had OpenAI launching Sora, a web browser called Atlas, a hardware device, and e-commerce features for ChatGPT. CEO Sam Altman reportedly compared the approach to running a portfolio of startups inside one company. Current and former employees told the WSJ the sprawl made it nearly impossible to articulate a core strategy—which is awkward when you’re planning an IPO.
OpenAI’s Enterprise Pivot Comes After Anthropic Ate Its Lunch
The revenue math explains the urgency. OpenAI carries a $25 billion annualized run rate entering 2026, but only about $10 billion of that comes from enterprise clients. Anthropic has surged past $19 billion in annualized revenue, with roughly 80% from enterprise. Dario Amodei confirmed at a Morgan Stanley event that $6 billion in revenue was added in February alone—almost entirely from Claude Code. That is a revenue doubling in two months.
Enterprise adoption data from Ramp tells a starker story. PYMNTS reported that 33% of American businesses paid for Anthropic’s AI in March, up more than 6 percentage points from the prior month. OpenAI sat flat at 35%. ChatGPT downloads rose 5% in the month to March, while Claude downloads tripled to 21 million. In the US, ChatGPT’s weekly active users declined month-over-month for the first time in roughly two years, per Apptopia data cited by the Financial Times.
OpenAI disputed the figures. The company told the FT it didn’t recognize the data, noting that its coding agent Codex now has 3 million weekly users (up from 2 million) and that ChatGPT’s ads pilot hit a $100 million run rate in six weeks. “Our APIs now process more than 15 billion tokens per minute,” the company said. “ChatGPT has six times the monthly web visits and mobile sessions of the next largest AI app.”
Simo explicitly referenced Anthropic in the all-hands meeting, calling it a “wake-up call.” The pressure isn’t just commercial. The US Defense Secretary designated Anthropic a supply-chain risk after it declined to give the Pentagon unrestricted model access. PYMNTS reported that some businesses have grown hesitant to work with Anthropic—and OpenAI has moved to capture those vacated government contracts.
The losses are equally real. OpenAI recorded an $8 billion net loss in 2025. Internal projections reviewed by investors suggest cumulative losses through 2029 could reach $115 billion, with profitability not expected before the early 2030s. The company has revised its infrastructure spending target from the $1.4 trillion figure Altman once floated to $600 billion in total compute spend by 2030.
Amazon led a $110 billion funding round at a $730 billion pre-money valuation in February. One structural condition stands out: $35 billion of Amazon’s $50 billion commitment is contingent on OpenAI either achieving AGI or completing its IPO by year-end. That is not a funding clause—it is a deadline.
The Economist reported that if OpenAI, Anthropic, and xAI each float 15% of their shares, the combined sum would roughly equal the total raised across all American IPOs over the prior decade. Both companies are preparing S-1 filings for the second half of 2026. Anthropic retained Wilson Sonsini Goodrich and Rosati—the firm that managed Google’s 2004 IPO. OpenAI has held early discussions with Cooley.
The first to list will set the valuation benchmark for every AI company that follows. Public investors will decide whether 100-times-revenue multiples hold. Anthropic’s enterprise-heavy mix—precision work like agentic AI deployments and contract analysis, not consumer subscriptions—may prove more durable in that test.