The signal is unmistakable. Polymarket, the crypto-native prediction market that has become an unlikely oracle for tech capital events, now prices Anthropic's October IPO probability at 70%. By year-end, it climbs to 83%.
But here is the number that should make every analyst pause: the target valuation for this debut is $2 trillion.
Let me calibrate the data before we go further. The original report from CryptoPotato cited a $9.65 billion post-money valuation from Anthropic's May funding round, claiming the $2 trillion target was "more than double" that figure. That math is broken. The correct post-money valuation from May is approximately $965 billion — a trillion-dollar level, not a billion-dollar one. This transcription error matters because the actual valuation leap from May to October is still over 100%, but the base is an order of magnitude higher. The narrative is aggressive enough without the data errors.
I watched fortunes bloom and wither in real-time during the 2021 NFT mania, and the patterns here are eerily familiar. The market is pricing an option on the future of AI, not a company with a proven profit model. Speed is survival, but empathy is the signal — and right now, the market is showing empathy only for the narrative.
The Context: What Is Anthropic Actually Bringing to Market?
Anthropic is the developer of Claude, a family of large language models that has carved out a distinct identity in the AI landscape. Its "safety-first" positioning — constitutional AI, rigorous alignment research, and a public commitment to responsible deployment — has become its primary competitive differentiator. This is not just a technical choice; it is a business strategy that resonates with enterprise clients and regulators alike.
The company's revenue trajectory is genuinely impressive. Reuters reports that Anthropic's annualized revenue is approximately $47 billion, placing it in the global top two alongside OpenAI. By 2028, the company projects revenue between $190 billion and $200 billion. This is a 4x growth from current levels, implying a compound annual growth rate of roughly 60%.
But here is the first crack in the narrative: the revenue composition is opaque. How much comes from API calls versus enterprise subscriptions? What is the gross margin? Most critically, what percentage of that $47 billion comes from strategic partners Amazon and Google — who are also investors? If the concentration exceeds 50%, the "independent commercial story" begins to fracture.
The Core: Deconstructing the $2 Trillion Valuation
Let me run the numbers the way I would audit a smart contract's reentrancy vulnerability — layer by layer, with no assumptions left unchecked.
At $2 trillion market cap and $200 billion in forward revenue (2028), the price-to-sales ratio is approximately 10x. This is within the range of high-growth SaaS companies during the bull years of 2018-2021. Salesforce traded at 8-12x forward P/S during that period. So the multiple itself is not the problem.
The problem is the current P/S. At $2 trillion market cap and $47 billion annualized revenue, the current P/S is approximately 42.6x. This is far beyond any reasonable SaaS valuation and sits firmly in the territory of pre-revenue unicorns during the 2021 SPAC frenzy.
Now consider the earnings dimension. If Anthropic achieves a 25% net profit margin by 2028 — an extremely optimistic assumption given the crushing cost of AI inference — net income would be $50 billion, yielding a P/E of 40x. That is acceptable for a high-growth tech company. But if the margin is only 10-15%, the P/E balloons to 67-100x. That is bubble territory, and we all know how those stories end.
Code was the law, and I was its restless guardian. The code here is the revenue model, and it has not been audited by the public markets yet.
The Valuation Leap: A Five-Month Doubling
The May funding round valued Anthropic at approximately $965 billion (calibrated). The October IPO target is $2 trillion. That is a 107% increase in five months. This is not organic value creation; this is narrative acceleration.
I have seen this pattern before. In 2021, NFT projects would raise at a $10 million valuation in January and claim a $100 million valuation by June based on nothing but Discord hype and Twitter engagement. The market was pricing the narrative, not the fundamentals. The difference here is that Anthropic is a real company with real revenue, real technology, and real customers. But the velocity of the valuation increase is a warning signal that cannot be ignored.
The Competitive Blind Spot
The $2 trillion valuation implicitly assumes that Anthropic will maintain parity with OpenAI in the model race and dominate the enterprise market. This is far from guaranteed.
OpenAI's annualized revenue is significantly higher than Anthropic's, and its ecosystem is broader — ChatGPT's consumer reach, the developer platform, Sora, and the deep integration with Microsoft's Azure and Copilot suite. Anthropic's partnership with Amazon and Google is a "multi-bet" structure that provides stability but also limits the deep ecosystem synergy that OpenAI enjoys with Microsoft.
Stability isn't a feature; it's a protocol. Anthropic's stability as a business depends on its ability to maintain its safety differentiation as a competitive moat. But OpenAI is rapidly closing the gap in enterprise security and compliance features. The window of differentiation is narrowing.
The Infrastructure Black Hole
Here is the most underreported dimension of the entire IPO narrative: Anthropic's capital expenditure requirements for AI infrastructure.
To go from $47 billion to $200 billion in revenue, Anthropic will need to scale its inference capacity by 5-10x. Each frontier model training run costs hundreds of millions to billions of dollars. Between now and 2028, there will be at least two to three major model iterations. Where does this capital come from?
The IPO proceeds will almost certainly be earmarked for building proprietary compute clusters, reducing dependence on Amazon and Google. But the cost of this buildout is not disclosed in any of the IPO coverage. If capital expenditure consumes 30-40% of revenue, the free cash flow story collapses, and the valuation multiple becomes indefensible.
The Contrarian Angle: The Fragility of the Narrative
Let me offer a counter-intuitive perspective that the market is ignoring.
The $2 trillion valuation is not just aggressive; it is structurally fragile in ways that are not being discussed.
First, the Polymarket data. A prediction market with $50 million in total volume does not represent the wisdom of institutional crowds. It represents the sentiment of crypto-native traders who are structurally long on risk. Using this as a primary evidence source for IPO probability is a methodological error.
Second, the "IPO pricing discount" problem. Anthropic's private market valuation is already ~$1 trillion. To attract public market investors, the IPO price typically needs to be below the latest private valuation to offer a "first-day pop." If the IPO prices at $1.5 trillion instead of $2 trillion, the narrative shifts from "record-breaking debut" to "disappointing discount." The market's reaction to a $1.5 trillion pricing would be a critical stress test of the entire AI valuation thesis.
Third, the revenue composition opacity. If Amazon and Google account for more than 40% of Anthropic's revenue, the company is effectively a strategic supplier to two hyperscalers, not an independent AI platform. The $2 trillion valuation assumes independence, but the business model may tell a different story.
Fourth, the AI safety paradox. Anthropic's entire brand equity is built on being the "safe" AI company. But if a major AI safety incident occurs — a model leak, an agent misuse event, or a regulatory crackdown — Anthropic will face the highest scrutiny because of its positioning. The safety moat could become a liability if the industry's trust is shaken.
The Takeaway: What Comes Next
The market is not pricing Anthropic's current business. It is pricing a call option on the entire AI industry's trajectory from 2028 to 2035. The $2 trillion valuation is a bet that AI will transform enterprise software as profoundly as the internet transformed media, and that Anthropic will capture a disproportionate share of that value.
This bet may be correct. But the asymmetry of risk is uncomfortable. If the narrative holds, early investors will see returns that rival the early days of Amazon or Google. If the narrative breaks — due to a competitive miss, a revenue slowdown, or a macroeconomic shock — the downside is catastrophic.
I watched fortunes bloom and wither in real-time during the DeFi Summer of 2020 and the NFT mania of 2021. The protocols that survived were the ones with real users, real revenue, and real transparency. The ones that collapsed were the ones that relied on narrative momentum without fundamental backing.
Anthropic is not a rug pull. It is a genuinely impressive company with world-class technology and a clear mission. But the $2 trillion valuation is a narrative construct, not a fundamental one. The first quarterly earnings report after the IPO will be the moment of truth. If the revenue growth rate is below 50% annualized, or if the gross margin disappoints, the narrative will crack.
Until then, the market is buying a story. The question is whether the story can survive contact with reality.