Anthropic's $190B Revenue Forecast: The IPO Math That Defies Gravity

Anthropic is heading to market with a valuation anchored to a 2028 revenue forecast of $190 billion to $200 billion. That number does not exist yet. Neither does the company's ability to achieve it, but that particular detail appears to have survived only one read-through of the pitch book before someone spilled coffee on it.

Let us establish what we are actually discussing here. Anthropic publicized a $47 billion annual revenue run rate as recently as May. The forecast under discussion represents a fourfold increase in just over two years. The company is pushing toward a near-$2 trillion valuation on the back of it. For context, that would make Anthropic worth roughly 50 times its current run rate revenue. Palantir, a company that has actually grown and turned a profit, trades at 53 times this year's expected revenue. SpaceX and Cloudflare both sit at 41.6 times expected 2026 revenue. The bankers have done their spreadsheet work.

This is where the fiction begins in earnest. Valuation methodologies are only as useful as the assumptions bolted underneath them, and the assumptions here require something close to religious faith. The premise: Anthropic will grow revenue by 300 percent in 26 months while simultaneously expanding margins as it scales. That margin expansion is not currently happening. The company is being pressed by the same constraint that presses every large AI operation—the cost of compute, model training, and hiring talent at the top of the market does not decline as you grow. It rises. Ask Anthropic's bankers about this and they will hand you a spreadsheet showing the point at which scale economies kick in. Ask them when, and they will gesture at 2028 and hope you nod along.

The precedent for this kind of long-range forecasting in tech IPOs has not aged well. Cerebras Systems cited 2028 revenue expectations in the runup to its own IPO. SpaceX projections extended as far as 2029 before the company went public at a record valuation in June. Neither company faced the kind of immediate skepticism you might expect from seasoned capital markets participants. Wall Street prefers forecasts to fact-checking. Forecasts are negotiable; facts are not.

But facts, as it turns out, do eventually arrive. GEMI shares are down 90 percent from IPO following multiple Wall Street price target cuts. The stock is now trading at record lows after analyst repricing. This was a company that, much like Anthropic, benefited from generous initial valuations and forward-looking revenue projections. The difference between GEMI and Anthropic is not that one was speculative and one was not. The difference is that GEMI's forecasts met the market first.

Anthropric has one genuine edge that GEMI did not: it has actually grown revenue at a demonstrable pace and currently sits ahead of chief rival OpenAI in both annual revenue and valuation metrics. That is not nothing. But it is also not $190 billion in 26 months. That is a spreadsheet. And Wall Street has learned nothing from the last three years of watching spreadsheets collide with reality.

The question is not whether Anthropic's management believes these numbers. It is not even whether the bankers believe them. The question is what happens on the day the 2026 results arrive and they are something less than 300 percent of 2024's run rate. At that point, the $2 trillion valuation will need to find another foundation, and foundations built on forward projections tend to vanish when the forward arrives.

For now, Anthropic will price its IPO on the strength of those 2028 forecasts. The market will pay. Investors will buy. And someone, somewhere, will spill coffee on the page where the margin assumptions are spelled out. That is not a prediction. That is historical pattern recognition.