TechCrunch reported last week on an interesting development: some institutional investors who have stakes in both OpenAI and Anthropic are rethinking the valuation logic for the two companies.
This doesn't mean they are selling OpenAI. But when comparing the two, the numbers are becoming increasingly difficult to justify.
The Numbers
Anthropic's annualized revenue stood at $9 billion at the end of 2025. By the end of March 2026, just three months later, it had reached $30 billion.
That is more than a threefold increase. The primary driver is its coding tool — Claude Code's penetration in the enterprise sector has exceeded most expectations.
OpenAI's current valuation is $852 billion. One investor holding shares in both companies revealed that their internal estimate is: to make OpenAI's valuation make sense, one would need to assume its market cap at IPO would exceed $1.2 trillion.
That figure has no historical precedent.
What the Secondary Market Says
More telling than what investors say is what they do.
In the private secondary market, demand for Anthropic stock is strong, while OpenAI shares are trading at a discount. Neither company's stock is freely tradable, but price movements in the private secondary market are real market signals.
When one side commands a premium and the other trades at a discount, the market is speaking clearly.
The "Netscape" Comparison
Jai Das, President of Sapphire Ventures, made a remark that made OpenAI executives uncomfortable: he compared OpenAI to "the Netscape of the AI era."
What was Netscape? The absolute dominant player of the early web era, later crushed by Microsoft's Internet Explorer, and eventually disappeared.
This analogy may not be entirely accurate, but it reveals a concern beginning to circulate among investors: Is OpenAI the first mover, but will the profits ultimately be taken by later entrants?
OpenAI faces not just Anthropic, but also Google, Meta, and China's large model army. The pattern of first-mover profits being eroded by later competitors is not without precedent in the tech industry.
OpenAI's Response
OpenAI's CFO Sarah Friar responded simply: the company just completed a $122 billion funding round, which itself is a reflection of market confidence.
This logic has some merit — if no one believed in the company, that much money wouldn't have come in. But a large funding round and a secondary market discount are not contradictory. The primary market reflects negotiated outcomes, while the secondary market provides real-time pricing. The signals from both markets can coexist.
OpenAI is also accelerating its enterprise customer transformation, directly challenging Anthropic's strength. But Anthropic already has a lead in enterprise coding scenarios, and whether OpenAI can catch up remains uncertain.
Why Anthropic Is Growing So Fast
The core driver behind the $30 billion annualized revenue growth rate is one thing: enterprises are willing to pay real money for AI coding tools.
A key decision Anthropic made in 2025 was not to compete with OpenAI for the consumer market, but to bet on enterprise coding and knowledge work scenarios. After the launch of Claude Code, that bet has largely been validated. Surveys show 46% of developers prefer Claude Code, and Claude Code's code accounts for about 4% of commits in public GitHub repositories, with expectations of reaching 20% by year-end.
The returns from this focus are now reflected in the revenue data.
The Underlying Logic of This Competition
Anthropic vs. OpenAI is no longer just about "which company has better technology," but "which company's business model is more sustainable."
OpenAI's problem is not a lack of revenue, but that its expenses are expanding in parallel — computing costs, R&D, headcount, and the ongoing commercialization transformation all require continuous cash burn. The $852 billion valuation rests on the assumption that "future growth can support it." Once growth begins to slow, pressure will quickly emerge.
Anthropic, of course, has its own issues. The $30 billion annualized revenue is real, but it is built on a single scenario — coding tools — making it highly concentrated. If OpenAI can truly catch up in this scenario, or if a new player emerges, this foundation could be shaken.
Jai Das's "Netscape" comparison may be overly pessimistic. But the risk it describes is real.
Sources: CocoLoop, Anthropic's rise is giving some OpenAI investors second thoughts (TechCrunch)