The $190 Billion Mirage: Why Anthropic's IPO Forecast Fails the Math Test
The number is too round to be real. $190 to $200 billion in revenue by 2028. That is the headline forecast for Anthropic's IPO valuation, a figure that would make it the fastest-growing company in human history. Let's be clear: this is not a projection. It is a mathematical hallucination dressed up as an investment thesis.
Anthropic, the AI safety company behind Claude, is reportedly targeting a valuation that requires a 200x revenue increase in four years. From roughly $10 billion in 2024 to $200 billion by 2028. That is a compound annual growth rate of 365%—every year, for four years. For context, the fastest-growing software company in history, AWS, took 12 years to reach $100 billion in revenue. Anthropic is supposed to do it in half the time from a fraction of the base.
I have spent years auditing smart contracts and DeFi protocols. I have seen the same pattern: a number that looks impressive but collapses under the weight of basic arithmetic. When I first saw the $190-200B figure, my instinct was to check the units. It is almost certainly a typo. The real number should be $19-20 billion. That would still imply a 110% CAGR, which is aggressive but plausible. The author of the original report—likely a journalist or analyst—either misread the decimal or was fed a number designed to inflate the narrative.
Let's run the math. If Anthropic does $10 billion in 2024, and we assume a 3x growth rate (optimistic but possible), 2025 would be $30B, 2026 $90B, 2027 $270B, and 2028 $810B. That is still less than half of $200B. To hit $200B, you need a 4x growth rate every year, which implies the entire global AI market would have to shift to Anthropic alone. That is not growth; that is monopoly.
Now compare to OpenAI. Industry estimates place OpenAI's 2028 revenue at around $100 billion—half of Anthropic's forecast. And OpenAI is the market leader with a first-mover advantage and a massive installed base. For Anthropic to be twice as large as OpenAI in four years, it would need to capture not just enterprise AI but also a significant chunk of cloud infrastructure. The global AI software market in 2028 is projected at $2-5 trillion. If Anthropic captures $200B, it would own 4-10% of the market. That is plausible. But the article claims $200B, which is 40-100% of the market. That is not plausible.
Debug the intent, not just the code. Why would a report publish such a forecast? One possibility: it is a deliberate attempt to inflate IPO expectations. In blockchain, we see this all the time—projects project multi-billion dollar revenues to justify token valuations. The same psychology applies here. If you can convince investors that a company will be worth a trillion dollars in 2028, you can justify a $50 billion valuation today. The forecast is not a prediction; it is a marketing tool.
Another possibility: the revenue definition is inflated. Does the $200B include resold compute from Amazon and Google? If so, that is not real revenue. It is pass-through, like a grocery store selling milk. The actual margin is thin. Trust the hash, not the hype. We need to know the revenue composition before we can even begin to evaluate the forecast.
Contrarian angle: what if the forecast is correct? What if AI adoption accelerates beyond all expectations? In that case, Anthropic would become the dominant AI infrastructure provider, effectively replacing AWS as the world's largest cloud company. But that would require a shift in global IT spending unprecedented in history. It would also require Anthropic to maintain a lead over OpenAI, Google, Meta, and emerging competitors. Possible? Marginally. But the probability is so low that it should not be the base case for any investment decision.
From my experience in DeFi, I have seen many protocols project 1,000% APY yields. They always collapse. The same applies here. Sustainable growth rates in enterprise software rarely exceed 50% CAGR for more than a few years. A 365% CAGR is a Ponzi-like assumption. It relies on new capital entering the market faster than the company can deliver value. That is not a business model; it is a bubble.
So what is the real takeaway? Treat the $190-200B figure as a red flag. Either it is a typo, or it is a deliberate overstatement. In either case, it should not be used as a basis for valuation. The honest forecast—$19-20B in 2028—is still ambitious but achievable. That would imply a 10-15x revenue multiple, giving a valuation of $200-300 billion. That is in line with current private market expectations. Anything above that is noise.
Volatility is the tax on uncertainty. The real uncertainty here is not about AI's potential—it is about the integrity of the numbers being presented to the public. The next time you see a forecast that rounds to such a neat, large number, ask yourself: who benefits from this narrative? And then check the math. Always check the math.