The meetings are starting to appear on calendars. Anthropic executives have begun sitting down with prospective investors, according to a person familiar with the plans, as the artificial intelligence company behind the Claude chatbot moves toward an initial public offering. The sessions are being organized by the banks leading the deal, and they amount to an early test of demand before any formal roadshow.
Goldman Sachs, Morgan Stanley and JPMorgan Chase, the three biggest Wall Street banks by revenue, are involved in the offering, according to people familiar with the matter. The lineup was formalized in recent weeks, and the banks have begun scheduling meetings between investors and Anthropic’s management, the person said. Anthropic filed its IPO prospectus confidentially with the Securities and Exchange Commission in June, and the company could list as soon as October, though the timing could slip.
An October debut would make Anthropic the first of the major AI labs to reach the public market, ahead of OpenAI, which is said to be eyeing 2027, and DeepSeek, which has also been preparing. The race matters because the company that lists first sets the valuation benchmark for the AI sector and gets first claim on institutional allocation budgets. It is a prize the banks are competing hard for.
The valuation talk has already escalated. Anthropic was valued at about $965 billion in a fundraising completed in May, a level above the roughly $852 billion OpenAI carried in its most recent private round, according to people familiar with the figures. The company has raised roughly $65 billion in private capital since its founding, and its backers include Amazon, Google and a roster of venture and sovereign funds.
The money goes out almost as fast as it comes in. Anthropic’s most distinctive cost is a computing contract with SpaceX, revealed in the rocket company’s own IPO filings, under which it buys about $1.25 billion of Nvidia chips per month, roughly 325,000 processors across the life of the arrangement. Analysts said the deal gives Anthropic the compute it needs to train frontier models but also locks it into a fixed cost base that will be a central question in every investor meeting.
The meetings will test how much of that story investors are willing to pay for. Anthropic has said its revenue is growing quickly, and its enterprise business, built on selling Claude to companies, has become its biggest revenue line. But the company is also spending heavily on compute, talent and data centers, and its path to sustained profitability depends on pricing power it has not yet demonstrated at public-company scale.
The cautionary example is fresh. SpaceX, which listed in June at a valuation around $1.8 trillion in one of the largest IPOs in history, has since traded below its offering price, a slump that bankers and analysts have attributed in part to the expiration of lock-up periods. The slide has put the durability of the IPO pipeline in question and raised the bar for the next big debut.
The banks are not treating Anthropic as a sure thing. People familiar with the deal said the underwriters are still deciding on the offering’s size, price range and structure, and that some details remain unsettled. JPMorgan’s role has been the subject of particular attention because the bank’s chief executive, Jamie Dimon, has publicly criticized parts of the AI industry, and because JPMorgan is also working with OpenAI on its own preparations.
The dual role of the banks illustrates how crowded the AI IPO lane has become. Morgan Stanley and Goldman Sachs have been battling for the top position on both the Anthropic and OpenAI offerings, a competition that has turned the two companies’ preparations into a closely watched race for fees, prestige and the right to call one of the largest technology IPOs in history.
For investors, the pitch is straightforward. Anthropic argues that it is a leader in frontier model development, that its safety-first positioning gives it an advantage in enterprise sales, and that the market for AI software is expanding faster than almost any software market before it. The company has also been widening its product line, adding tools for businesses to deploy its models in production, in an effort to show that its technology generates recurring revenue rather than one-time curiosity.
The questions will be harder. How much of revenue goes to compute? What happens to margins when competitors cut prices? Can Claude hold share against OpenAI’s ChatGPT and Google’s Gemini, both backed by companies with deeper pockets? And what does a public Anthropic do when its biggest customer relationships, including its compute deals, are also with its biggest investors?
The answers will start to emerge in the coming weeks as the meeting schedule fills up. If demand is strong, Anthropic could file a public prospectus in the fall and price the offering quickly. If investors balk, the timetable slips and the banks go back to work.
Either way, the process has already changed the AI industry’s conversation. A company valued near a trillion dollars, burning cash on compute and preparing to face quarterly scrutiny is now a concrete prospect, not a rumor. The roadshow, when it comes, will be the first public reading on whether investors believe the AI boom can pay for itself.
