According to a Reuters exclusive, the AI lab’s initial public offering hinges on an internal forecast that it will book between $190bn and $200bn of revenue in 2028, a previously unreported figure that bankers and investors are already leaning on to justify a price tag that could brush $2tn.
The leap involved is the sort that only makes sense inside a boom. Anthropic’s revenue run rate sat at roughly $47bn as of May 2026, up from about $9bn at the end of 2025, so the reported 2028 target implies revenue more than quadrupling again in a little over two years.
The near-term numbers are, admittedly, moving fast. Second-quarter revenue for 2026 is projected to reach at least $10.9bn, more than double the previous quarter, and the company is said to be eyeing its first quarterly operating profit, a forecast figure of around $559m.
For a lab that has spent most of its life burning cash, that would be a genuine inflection point, and it is the kind of trajectory that lets bankers argue the forecasts are not fantasy but extrapolation.
Rather than pricing Anthropic on what it earns today, bankers are reportedly applying enterprise-value-to-revenue multiples to those 2028 forecasts, a two-years-forward approach that is rare in public markets but was used before the flotations of Cerebras and SpaceX. It is a method built for companies growing too quickly to value on the present tense.
Palantir trades at roughly 53 times its expected 2026 revenue, while SpaceX and Cloudflare sit at around 41.6 times. Stretch multiples like those across Anthropic’s projected 2028 haul and the eye-watering headline figures begin to look, if not sober, then at least internally consistent.
The catch is that the whole calculation borrows heavily from the future, and it assumes the enthusiasm on display today survives long enough to meet the revenue it is pricing in.
Not everyone is convinced they would hold. “Could they get a $2 trillion valuation, yeah they could and I just wonder if it would stay there over time,” said David Merkel of Aleph Investments, neatly capturing the gap between what a hot market will pay on the day and what it will tolerate a year later.
The caveat sits in the cost base. Heavy spending on GPUs, compute and model training is pressing on current margins, and the bull case rests on the assumption that those costs fall as a share of revenue as the business scales.
If they do not, the two-years-forward logic starts to wobble, because the whole exercise trades present profitability for future size.
The valuation talk has been climbing for months. Anthropic has already attracted investor offers at an $800bn valuation and been reported to be eyeing something closer to $900bn in a mooted $50bn round, with the IPO chatter now nudging toward the $2tn mark that once sounded fanciful.
The wider risk is that the whole edifice depends on a forecast holding in a market that could turn. Cheaper rivals are already threatening the economics that underpin these valuations, and a two-years-forward multiple is only ever as good as the year it is pointed at.
For now, the story is the sort that raises a European eyebrow: dazzling growth, real questions about margins, and a valuation resting almost entirely on a number nobody can yet check.
Anthropic may well hit $200bn in revenue by 2028, and its recent run of results gives the ambition more credibility than it once had. But the market is being asked to price the company as though that arrival is already booked, and to do so two full years before the receipts come in.
Get the TNW newsletter
Get the most important tech news in your inbox each week.