The term sheet was already enormous. Then it got bigger. Anthropic’s latest fundraising round will raise more than the $86 billion SpaceX collected in its stock market debut earlier this year, according to people familiar with the matter, and it will give existing investors and employees a chance to sell shares in the deal.
The round’s size would rank among the largest private financings in technology history, larger than any single venture round ever completed. The secondary component matters as much as the primary: allowing existing holders to sell in the deal provides liquidity ahead of a public offering whose timing has not been set, and it lets early investors bank gains without waiting for an IPO that may still be years away.
Anthropic has been explicit about the scale it is aiming for. On August 25, the company told investors it is targeting a valuation of roughly $2 trillion for its eventual public listing, and it disclosed a potential market of more than $30 trillion, a figure that spans the software, agents and infrastructure the company believes its models can reach. The $30 trillion figure has drawn skepticism from analysts who note that it counts markets the company does not participate in yet, but it has also given investors a number large enough to justify the round’s size.
The same week brought a different kind of signal about the company’s priorities. Anthropic announced that starting September 14 it will cut the weekly usage cap for Claude Code, its coding agent product, by 17 percent, ending a temporary 50 percent boost the company had granted earlier in the year. The company framed the change as a return to normal service levels; the industry read it as something else: before an IPO, costs matter, and coding agents are among the most compute-hungry products in the company’s lineup.
The usage-cap decision is the tell in the whole transaction. Anthropic is raising more money than any startup has ever raised, at a valuation that puts it among the most valuable companies in the world, and it is simultaneously tightening the spigot on its most popular product rather than letting usage run ahead of revenue. The message to investors is that the company is done subsidizing growth and is now managing toward profitability, the standard posture of a company preparing for the public markets.
The financing arms race among AI labs has become a defining feature of the market. OpenAI has raised in the tens of billions across multiple rounds, xAI has closed enormous financings of its own, and each successive round sets a new record that the next one beats. Anthropic’s round extends the pattern: the sums have grown so large that they no longer finance product development so much as they finance the credibility of the companies themselves, signaling to customers and partners that the lab will be around to honor its contracts.
Critics call the dynamic circular. The same investors, and the same cloud providers, appear on both sides of the ledger: funds that own cloud shares also own lab stakes, and the labs pay those same clouds for computing. The cloud companies, in turn, invest in the labs whose workloads they host. Anthropic’s counterargument is the market disclosure: $30 trillion of addressable opportunity, agents that will eventually do a meaningful share of knowledge work, and a technology curve that has not flattened yet.
For employees, the round solves a practical problem. Anthropic’s staff hold options and restricted stock at valuations that have risen steadily, and the secondary component of the round gives them a rare chance to convert paper wealth into cash before an IPO. The company has said it will set aside a portion of the round for employee sales, a decision that also serves retention at a moment when rival labs are paying aggressively for talent.
The round also changes the competitive math of the AI industry. A lab with this much capital can fund its own infrastructure, its own models and its own go-to-market without depending on the generosity of cloud partners, and Anthropic has been explicit that it wants to control more of its own computing. The company has signed infrastructure agreements with several providers and has been building out its own capacity, and this round gives it the balance sheet to accelerate that effort.
The valuation question is where investors will focus. A $2 trillion target prices Anthropic as one of the most valuable companies in the world before it has completed a public listing, and the $30 trillion market figure is the justification. Skeptics argue that the figure counts potential rather than revenue, and that the AI market will be shared among many competitors rather than captured by one. The company’s answer is that the market is growing fast enough that even a modest share is worth the price.
There is also a regulatory dimension. A financing of this size, with a secondary component that lets insiders sell, will draw scrutiny from securities regulators and from lawmakers who have questioned whether AI labs’ valuations reflect fundamentals. The company has said it is structuring the round in compliance with applicable rules, and its lawyers will spend the coming weeks parsing the details with regulators’ preferences in mind.
The deal is expected to close in the coming weeks, according to people familiar with the matter, with final terms still being negotiated. When it closes, Anthropic will have done something no startup has done before: raised more money than one of the largest IPOs in history, and done it while telling its most dedicated users to use its product a little less. The two facts belong to the same strategy, and investors will be watching both.
