Hook
A $190 billion revenue prediction for Anthropic by 2028. That’s not a typo. It’s a data anomaly. The chain of logic collapses under the weight of a single decimal point. Every analyst who reads this number as fact is walking into a trap. I’ve seen this pattern before—in 2017, when I audited ICO whitepapers promising billions in ecosystem value with zero deployed code. The numbers looked clean. The narratives were tight. But the on-chain evidence told a different story. Here, the evidence is even simpler: basic arithmetic. Let the data speak.
Context
Anthropic is the AI company behind Claude, a large language model positioned as the safer, more aligned alternative to OpenAI’s GPT. Founded by former OpenAI researchers, it has raised over $7 billion from investors including Google, Spark Capital, and Salesforce. Its current annualized revenue is estimated at around $10 billion as of 2024, based on media reports and industry cross-references. The IPO is widely anticipated, with valuations floating between $30 billion and $60 billion in private markets. Then came the prediction: $190-200 billion in revenue by 2028, quoted in a recent article as the basis for a trillion-dollar valuation. This is the number that breaks the model.
Core
I ran the numbers through four independent checks. Each one returned a red flag.
First, the growth rate test. If Anthropic does $10 billion in 2024, reaching $200 billion by 2028 requires a compound annual growth rate of roughly 110%. That’s aggressive but not impossible for a software company in a hyper-growth phase. But $190-200 billion? That’s a CAGR of 270-280%. To put that in perspective, AWS took 12 years to reach $100 billion from a $7 billion base. Anthropic is being asked to do it in four, from a $10 billion base, in a market that is already crowded with OpenAI, Google, Meta, and xAI. The probability is near zero. Even the most optimistic scenario—tripling revenue every year—only gets to $810 billion by 2028. Still less than half of the prediction.
Second, the market share test. The global AI software and services market is projected to be between $200 billion and $500 billion by 2028. If Anthropic captures $190-200 billion, it would control 40% to 100% of the entire market. No enterprise software company in history has achieved that. Microsoft’s peak share of the global software market never exceeded 20%. The prediction implies Anthropic becomes the dominant AI infrastructure provider, surpassing AWS, Azure, and Google Cloud combined in a single segment. The data doesn’t support that.
Third, the peer comparison test. OpenAI, the current leader, is expected to generate around $100 billion in revenue by 2028, according to Wall Street projections. That’s half of what Anthropic is supposedly going to do. For Anthropic to double OpenAI’s revenue, it would need to capture not just the enterprise AI market but also the consumer AI market, the API market, and the compute resale market simultaneously. The competitive landscape is too fragmented. Google, Meta, and Microsoft are all investing heavily in their own models. The idea that Anthropic alone can outpace the combined resources of the world’s largest tech companies is a hypothesis that requires far more evidence than a single headline number.
Fourth, the unit error test. When I reverse-engineer the prediction, everything clicks if the number is actually $19-20 billion—a 10x error. At $19 billion, the CAGR drops to 17%, which is reasonable for a premium enterprise software company. The valuation that follows—around $200-300 billion at a 10-15x sales multiple—aligns with the $183 billion valuation Anthropic reportedly received in its 2025 funding round. The unit error is the most parsimonious explanation. But the article didn’t present it as a data point; it presented it as a projection. That’s a red flag for any analyst who relies on the raw data.
Contrarian
Now, let me play the contrarian. What if the $190-200 billion number is not a mistake but a deliberate narrative? In the crypto world, we see this all the time: a project announces a $100 billion market cap TGE based on a tokenomics model that has no real demand. The number is designed to attract institutional investors who are looking for a story, not a spreadsheet. Anthropic’s IPO roadshow might be using the same playbook. The prediction sets a high anchor, so when the actual revenue comes in at $20 billion, it looks like a disappointment but still justifies a high valuation. The narrative is the product, not the revenue.
But I’m an empirical skeptic. I’ve seen this before in the 2017 ICO boom. Projects would publish whitepapers with projected user bases of 100 million in three years, based on nothing but a word document and a logo. The on-chain data would show zero transactions. The same pattern is emerging here: a high-revenue projection without any supporting evidence of how the revenue will be generated. The article didn’t break down the revenue sources—API calls, enterprise subscriptions, compute resale, or licensing. It just gave a number. That’s not analysis. That’s marketing.
Takeaway
The real signal is the gap between the prediction and the data. The next-week signal is the IPO filing. When Anthropic files its S-1, look for the actual revenue breakdown, the customer concentration, and the growth rate assumptions. If the filing shows a linear growth path from $10 billion to $20 billion, the prediction was noise. If it shows a parabolic curve, then the market is being set up for a correction. My advice: follow the chain of evidence, not the headline. The liquidity pool of investor capital is a mirror, not a reservoir. It reflects what the data shows, not what the narrative promises.
Tracing the ghost coins back to the genesis block. Every transaction leaves a scar on the ledger. Whales don’t swim in shallow water. The chain doesn’t lie—it just waits for someone to read it.
