Anthropic Is Now Worth $800 Billion. Is That Insane or Just the New Normal?

The Claim

Anthropic’s $800 billion valuation is not the bubble people think it is. At its current revenue trajectory, the number is aggressive but defensible. The real valuation question in AI right now is not about Anthropic. It’s about OpenAI.

Why Most People Get This Wrong

When investors hear $800 billion for a private startup that didn’t exist five years ago, the instinct is to call it a bubble. That instinct is understandable but lazy. Valuation multiples only mean something when you hold the denominator constant. Anthropic’s revenue denominator is not standing still. It went from $9 billion at the end of 2025 to $30 billion by the end of March 2026. That is a 3.3x jump in roughly 90 days. No serious investor looks at that growth rate and sees a bubble. They see a land grab.

The Evidence

• Anthropic’s annual run-rate revenue hit $30 billion by end of March 2026, up from $9 billion at the end of 2025 and $19 billion just a few months before that.
• At $800 billion, the implied revenue multiple is roughly 27x. OpenAI’s $852 billion valuation against approximately $24 billion in annualized revenue implies a multiple closer to 35x.
• Demand for Anthropic shares on the secondary market is described as “nearly insatiable.” OpenAI shares are trading at a discount to their last round.
• Anthropic’s growth is driven by enterprise, not consumers. Coding tools and cybersecurity products are its core revenue. Enterprise contracts are stickier, higher margin, and more forecastable than consumer subscriptions.
• One investor told the Financial Times that justifying OpenAI’s latest round requires assuming an IPO at $1.2 trillion or more. That is the number that should make people nervous.

The Counterargument and Why It Does Not Hold

The bear case on Anthropic is straightforward: neither it nor OpenAI is profitable at scale, open-source models are improving fast, and regulatory risk around AI is real. All of that is true. But the counterargument misreads the market structure. Enterprise customers are not switching to open-source models for mission-critical tasks. The liability, reliability, and support requirements push large companies toward commercial providers. Anthropic’s Mythos model, which the company called too dangerous for public release because of its ability to find software vulnerabilities, is exactly the kind of capability that defense agencies and Fortune 500 security teams will pay significant premiums for. Revenue doesn’t lie at this scale.

What Follows From This

Watch Anthropic’s IPO timing. Current and former employees passed on selling shares at the February $350 billion valuation, holding out for a public listing reportedly targeted for as early as October 2026. If they list at $800 billion or above, it will be one of the largest tech IPOs in history and will force a serious repricing of the entire AI sector. The company to watch more carefully is OpenAI. Its valuation requires a $1.2 trillion IPO outcome to make early investors whole. If Anthropic goes public first and trades well, OpenAI’s clock starts ticking. If Anthropic trades poorly, every AI valuation in the market gets marked down overnight.