$852 billion bought OpenAI the largest private fundraising round in history. $380 billion bought Anthropic the momentum.

One investor who has backed both companies told the Financial Times that justifying OpenAI’s latest round required assuming an IPO valuation of $1.2 trillion or more — a figure that makes Anthropic’s $380 billion look like the deal of the decade.

The math is uncomfortable. OpenAI generates roughly $24 billion in annualized revenue against its $852 billion valuation, a price-to-revenue multiple of about 35x. Anthropic, valued at $380 billion, reported $14 billion in run-rate revenue in its own Series G announcement, though TechCrunch and SaaStr put the figure at $30 billion. Using the more generous estimate, the multiple is roughly 12.7x, and OpenAI costs nearly three times what Anthropic does per dollar of revenue. By Anthropic’s own figure, the gap narrows significantly to about 27x versus 35x.

This isn’t supposed to happen to the market leader.

The Strategy That Unraveled

OpenAI’s $122 billion raise, closed in late March, was supposed to be a coronation. Amazon, Nvidia, and SoftBank committed $110 billion combined. OpenAI CFO Sarah Friar pointed to the round as evidence of continued investor confidence.

But the product portfolio tells a different story. OpenAI abruptly shut down Sora, its video generation platform, and ended a $1 billion partnership with Disney that was supposed to anchor its push into entertainment and social media. Instant Checkout, a shopping tool that let users buy from retailers like Walmart through ChatGPT, was quietly scrapped after a five-month trial. CEO Sam Altman declared a “code red” in December to refocus on ChatGPT following advancements in Google’s Gemini, according to The Guardian.

The pattern is clear: OpenAI spread itself across video, commerce, entertainment, and consumer apps, and watched most of those bets fail to stick.

Anthropic’s Sharper Edge

Anthropic went the other direction. No video platform. No hardware. No shopping tools. No entertainment partnerships. The company bet almost entirely on coding tools and enterprise AI — and the revenue figures suggest the bet is paying off.

Claude Code’s run-rate revenue has surged past $2.5 billion, more than doubling since the start of 2026 alone. A recent analysis estimated that 4% of all public GitHub commits worldwide were being authored by Claude Code — double the percentage from just one month prior. Overall annualized revenue exploded from $1 billion at the end of 2024 to $30 billion by the end of March 2026, according to TechCrunch and SaaStr. That is a 30x increase in fifteen months.

Brad Gerstner at Altimeter, an investor in both companies, noted that Anthropic added the equivalent of Databricks plus Palantir combined in revenue in a single month. He told a recent podcast he wouldn’t be shocked if Anthropic exits 2026 at $80 to $100 billion in revenue.

The headcount story deepens the contrast. Anthropic employs somewhere between 3,000 and 5,000 people, depending on the source. Google needed 32,000 to cross $30 billion in revenue. Salesforce needed 79,000. Anthropic’s revenue-per-employee ratio is 6 to 10 times greater than Google’s at the same scale, according to SaaStr. Inference costs have fallen 90% year over year, and gross margins are expanding as revenue scales against a relatively fixed compute cost base.

The Secondary Market Speaks

Private markets have already rendered their verdict. Ken Smythe, founder of secondary marketplace Next Round Capital, told Bloomberg that roughly a half-dozen institutional investors — hedge funds and venture firms holding large stakes — approached his firm in recent weeks looking to sell about $600 million in OpenAI shares. Demand has sunk.

Anthropic shares have become what TechCrunch described as “nearly insatiable” on the secondary market.

Iconiq Capital partner Roy Luo made his firm’s position unambiguous. Iconiq has invested over $1 billion in Anthropic while holding a smaller OpenAI stake. “There’s room for both, but there is fundamentally a number one and a number two dynamic, and the number one will win disproportionately,” he told the FT. “We picked.”

The IPO Test

OpenAI is widely expected to go public later this year in one of the most anticipated offerings in decades. A successful IPO at or above $852 billion requires public market investors to believe that OpenAI will not only sustain its current trajectory but outpace competitors already growing faster with fewer people.

OpenAI generates $2 billion a month in revenue but loses billions of dollars annually and, according to the Wall Street Journal, does not expect to turn a profit until 2030. A looming court battle with co-founder Elon Musk — who is suing the company over its shift to a for-profit model — adds legal uncertainty to the financial picture.

Altman has navigated this kind of pressure before. During his tenure leading Y Combinator, aggressive valuation inflation left some portfolio companies financially stranded while others proved worth every penny. The question now is which category OpenAI belongs to — and whether the $1.2 trillion IPO that investors quietly require is a reasonable expectation or a collective delusion.

Sources