The data shows that Anthropic is planning a 2026 IPO at a $965 billion valuation. That number is a narrative, not a reflection of fundamentals. In crypto, I've seen this playbook before—projects that price themselves on future hype rather than current revenue. The difference is that Anthropic is a legitimate AI company, but the valuation math is still a house of cards.
Context: The AI Darling with a Cloud Leash Anthropic, the company behind the Claude model series, has positioned itself as the 'safe' AI alternative—Constitutional AI, heavy focus on alignment, enterprise-first. It has raised over $18 billion, with Amazon committing $8 billion and Google $2 billion+. The IPO is expected to be the largest tech IPO since Google itself. But look closer: the revenue story is still unproven. The company's 2025 annualized revenue is estimated at $5-15 billion, a wide range that suggests even insiders can't agree on the trajectory.
Core: The Valuation Gap I Trade On Let me apply the same forensic skepticism I use when analyzing a DeFi protocol's total value locked. A $965 billion valuation at a 20-30x price-to-sales multiple requires 2026 revenue of $32-48 billion. That means Anthropic must roughly triple its 2025 revenue in one year. Is that possible? In a bull market for AI, maybe. But here's the catch: the company is deeply dependent on Amazon Web Services for both training and inference. That's not a partnership—it's a single point of failure.
I've audited enough smart contracts to know that when a protocol's liquidity is locked in a single exchange, the risk of a rug pull is higher. Here, the 'liquidity' is AWS compute. If Amazon decides to renegotiate terms or if the relationship sours, Anthropic's cost structure becomes unpredictable. The company's S-1 will need to disclose the AWS contract details. If the pricing is favorable, fine. But if it's a standard enterprise agreement, the margins are thin.
The quantitative analysis is stark: at a 10x P/S (still high for a non-profitable company), revenue needs to be $96.5 billion. That's laughable. At 20x, it's $48 billion. Let's be generous and assume Anthropic hits $30 billion in 2026 revenue (a 3x from 2025). That gives a P/S of 32x—still above Snowflake's IPO P/S of 100x? No, Snowflake's was 100x but on a much smaller base. The point is that $965 billion is a narrative-driven number, not a data-driven one.
Contrarian: The Smart Money Is Already Hedging The mainstream narrative is that Anthropic is the 'safe' AI bet because of its security focus. But I see the opposite: the valuation is a trap for retail investors who don't understand the unit economics. Look at who is selling. Amazon is the largest shareholder. They have every incentive to see the IPO priced high to maximize their return. But ask yourself: why would Amazon, a company that literally runs the cloud infrastructure, let Anthropic go public at a valuation that implies a massive market share? Because they can diversify their risk. If Anthropic fails, Amazon still has AWS. If it succeeds, Amazon's $8 billion investment becomes $80 billion. The asymmetry favors Amazon, not the IPO buyer.
In crypto, we call this a 'smart money exit.' The insiders know the real value is lower. They'll sell into the retail demand. The contrarian trade is to short the IPO or wait for a post-IPO dip. The same pattern happened with Coinbase—opened at $381, dropped to $50 within a year. The fundamentals were there, but the hype was overpriced.
Another blind spot: the AI model race. Anthropic's Claude is good, but it's not dominant. OpenAI's GPT-5 and Google's Gemini are closing the gap. In this industry, technological leadership can vanish in six months. The market is pricing Anthropic as if it's a monopoly, but it's a competitive oligopoly. I've seen this in DeFi where a protocol with a temporary lead gets forked and surpassed. The same applies here.
Takeaway: The Gap Between Expectation and Execution I trade the gap between expectation and execution. Right now, the expectation is $965 billion. The execution is a company that hasn't proven it can scale revenue without burning cash. My advice: wait for the S-1 filing and scrutinize the AWS revenue recognition. If the unit economics are fuzzy, the trade is to short. If they are clear, the trade is to wait for a more reasonable entry point. Trust the math, verify the chain, ignore the hype. Uptime is a promise; downtime is the truth. And in this case, the promise is a valuation that will break under the first sign of decelerating growth.
