AI: Anthropic’s mega-IPO Math Banks on 2028. AI-RTZ #1181

The mega-AI IPO parade this AI Tech Wave is accelerating its pace. Elon’s SpaceX/xAI went first in June. Anthropic is next, still pointed at this fall. And OpenAI now likely follows as late as next year. Over the weekend, the Anthropic leg of the parade got its two most concrete data points yet: the revenue actually in hand, and the revenue the bankers are asking investors to price.

We served up our updated thoughts on Anthropic’s IPO prep last week in AI-RTZ #1176, when the company was rehearsing its IPO narrative with investors. This weekend moved it from narrative to numbers.

Start with the numbers in hand. Bloomberg, from documents shown to prospective investors, building on its Claude Code/Cowork momentum through several model increments:

“The Claude chatbot maker reported a preliminary revenue figure of more than $11.5 billion in its latest completed quarter, compared to $787 million in the corresponding period in 2025, and $4.73 billion in the first quarter of this year… The second quarter of 2026 saw Anthropic report positive adjusted operating income.”

That is 14-fold year over year, and well more than a double quarter over quarter. It also comes with the first whiff of operating profitability, which matters more than the growth number for what follows. The run rate had already crossed $47 billion in May, up from $9 billion at the end of 2025.

Then the number being pre-prepped with investors in terms of possible IPO valuations. Reuters lays it out in ”Anthropic IPO valuation hinges on $190-200 billion 2028 revenue forecast”. It reports the projection underneath the valuation talk:

“Anthropic is projecting 2028 revenue of roughly $190 billion to $200 billion, according to two people familiar with the company’s financials… Bankers and investors are using enterprise value-to-revenue multiples based on forecasts.”

Two things stand out in the Reuters reporting. The first is how far forward the Street is being asked to look: two years, which is unusual even for high-growth software. The second is the comp set being assembled for the analyst day: Palantir at 53 times this year’s expected revenue, and SpaceX and Cloudflare both at 41.6 times expected 2026 revenue.

The arithmetic explains the approach. On current-year numbers, any mega valuation looks over-ambitious. On the 2028 projection, a $2 trillion outcome is roughly ten times forward revenue, a multiple the Street can say out loud. The entire question becomes whether the $190 to $200 billion is real. As Reuters notes, the market has already accepted this forward-year framing this cycle: Cerebras backers cited 2028 revenue before its IPO, and SpaceX projections ran to 2029 before its record June debut.

The bear push-back writes itself, and Reuters includes it: could the valuation get to $2 trillion, “yeah they could and I just wonder if it would stay there over time,” as one investor put it. Underwriting a four-times-plus jump from the current run rate, while margins are still being built out of enormous compute spending, is a leap of AI faith however it is framed.

My take: this is the supply and demand series playing out in a prospectus. We spent four parts in recent days, arguing that the price of AI Compute falls in wide swings. All while the work done by AI expands beyond what fixed-quantity thinking allows.

Anthropic’s 2028 projection is management asking investors to underwrite exactly that expanding pie, and the margin story, training and inference getting cheaper as a share of revenue, is the same curve from the supply side.

The discipline in the pre-marketing is worth noting too. The company has kept the story it rehearsed last week intact: enterprise and AI Coding focus, frontier models, safety as both shield and brand. What changed this weekend is that the growth claims now come with a printed quarter and an operating profit attached. Aggressive, but aggressive with more backup.

The fall timing matters as much as the numbers. Bloomberg notes an IPO this fall would put Anthropic out not just before OpenAI but before DeepSeek in China. Which could file as soon as this year. In a listing year that has already raised over $256 billion, the most since 2021, going first among the mega-AI names is its own kind of pricing power. We covered that tight IPO clock in June, and it has only tightened since.

Tomorrow we look at the other side of the parade: OpenAI’s pre-marketing week, which came with its own investor meeting, its own numbers, and rather more churn.

An impact of the AI Tech Wave worth tracking. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here.)