Race to a Trillion
The twin titans of AI are bringing their diametrically opposed economics to the public markets.
OpenAI and Anthropic, the two companies that have dominated frontier AI development for the past three years, are both preparing to list on public markets before the end of 2026. Their combined private valuations exceed $1.2 trillion. A series of financial disclosures over the past two weeks has revealed that the companies, despite competing for many of the same enterprise customers, operate on fundamentally different economic foundations.
A quarter of the cost
The Wall Street Journal last week published confidential financial documents from both companies, shared with investors ahead of the funding rounds that each completed this spring. Anthropic’s annualized revenue run rate reached $30 billion in early April, up from $9 billion at the end of 2025, overtaking OpenAI’s $24 billion. Anthropic arrived at that figure in under three years from its first dollar of revenue. Salesforce took roughly two decades to reach the same mark.
The leaked documents revealed a sharp divergence beneath those topline figures. OpenAI projects $121 billion in compute spending by 2028, a year in which it expects to lose $85 billion even after continued revenue growth. The company does not forecast positive cash flow until after 2030. Anthropic’s training costs peak at roughly $30 billion in the same period, approximately a quarter of OpenAI’s outlay, and the company projects reaching profitability by 2028 or 2029.
The cost gap reflects different business architectures. OpenAI serves 900 million weekly ChatGPT users, most of whom pay nothing. Serving that free base generates inference costs without corresponding revenue. Anthropic draws roughly 80 percent of its revenue from enterprise customers, counts over 1,000 businesses spending more than $1 million annually, and lists eight of the Fortune 10 as clients. Claude Code, the company’s agentic coding tool, alone generates over $2.5 billion in annualized revenue, having doubled since January. Data from spend management platform Ramp showed that Anthropic’s share of enterprise AI spending rose from 24.4 percent to 30.6 percent in March, while OpenAI’s fell from 46 percent to 35 percent. Inference costs consume more than half of revenue at both companies, but Anthropic’s paying customer base offsets that burden more directly.
Under the microscope
OpenAI closed a $122 billion funding round in March at an $852 billion valuation, the largest private capital raise in Silicon Valley history. Amazon contributed $50 billion, with $35 billion of that sum conditional on OpenAI going public or reaching artificial general intelligence. The company’s CFO has confirmed that OpenAI is building toward a public listing and completed its conversion from a nonprofit to a public benefit corporation earlier this year, a structural prerequisite for an IPO that has faced legal challenges from state attorneys general. Anthropic has engaged Goldman Sachs and JPMorgan for a potential offering as early as October, targeting a raise exceeding $60 billion. If both listings proceed on schedule, they would rank among the largest IPOs on record.