
Updated:
18 NOV, 2026
By Joanna Piwko from RankiaPro Europe

Anthropic’s IPO is nearing the finish line. The company behind Claude will set its final price on the evening of November 18 (New York time), with its Nasdaq debut scheduled for tomorrow, November 19, under the ticker ANT. The price range filed with the SEC is $70 to $78 per share, but strong demand could push the price above the upper end.
For the first time, a generative artificial intelligence lab funded so far by venture capital and major technology companies is also seeking capital directly from the market. The stakes therefore go beyond a single company, since investors’ response will determine how they view the entire sector’s ability to turn explosive growth into profits, in the first real public test of a technology that until now has been valued only in private funding rounds.
Of the 825 million shares on offer, 801.4 million are newly issued and approximately 23.6 million are being sold by existing shareholders. At the top of the price range, gross proceeds would reach $64.35 billion, plus a $26 billion private placement at the IPO price, subscribed to by Google, Nvidia, AMD, Broadcom, Micron, Microsoft, and TSMC. In total, up to $90.4 billion: a figure that would surpass SpaceX’s IPO record.
Demand, meanwhile, has exceeded expectations. According to Bloomberg, orders amount to nearly seven times the shares available, and the Wall Street Journal reports that Anthropic could price its shares above $80. According to the newspaper’s calculations, this would bring the valuation to $1.5 trillion–$1.6 trillion, compared with approximately $1.4 trillion on a fully diluted basis implied by the top of the price range.
All this is unfolding against a fragile backdrop for tech, following the early-November sell-off and Jensen Huang’s resignation as Nvidia CEO, announced on November 12. According to Gianmarco Roncarolo, Sales Director at GraniteShares, compared with the valuations of up to $2 trillion speculated about in recent months, Anthropic seems to have opted for an almost cautious approach, leaving room for an increase after the listing.
| Anthropic’s IPO at a glance | Figure |
|---|---|
| Price range | $70–$78 per share |
| Final price | Expected on the evening of November 18 (New York time) |
| Nasdaq debut | November 19, ticker ANT |
| Shares offered | 825 million (801.4 million new shares and 23.6 million shares for sale) |
| Maximum proceeds | $64.35 billion + $26 billion private placement |
| Valuation at the top of the range | About $1.4 trillion (on a fully diluted basis) |
| Demand | Orders for nearly 7 times the offering |
| Revenue in the first nine months of 2026 | $33.24 billion |
| Compute and infrastructure commitments | About $895 billion |
| Cash and short-term investments (September 30) | $131.9 billion |
The revenue trajectory supports the valuation. According to Jack Amy, Investment Analyst at Moneyfarm, revenue rose from $386 million to $4.6 billion in 2025, and the updated prospectus reports $33.24 billion in the first nine months of 2026. Revenue in the third quarter alone was $16.96 billion, up 1,127% year over year: annualized, that exceeds $67 billion.
Governance is less straightforward, since after the IPO the seven co-founders would retain 50.1% of the votes through a Class F share, and four of the seven board members would be appointed by the Long-Term Benefit Trust. As Amy (Moneyfarm) notes, this means shareholders "would have limited power to influence its strategic choices".
For Violeta Todorova, Senior Research Analyst at Leverage Shares, the choice for investors is therefore "unusually clear-cut": buying ANT means holding a stake in "one of the fastest-growing companies ever recorded", while accepting that they will have "virtually no influence over its management".
While growth is impressive, the financials raise more than a few questions. In the first nine months of 2026, the net loss was 104 billion, of which 94.3 billion was related to fair value changes in convertible bonds and SAFEs, so not all of it represented cash outflows. For Amy (Moneyfarm), however, frontier AI remains "economically more akin in structure to sectors such as energy and infrastructure than to traditional software".
Todorova (Leverage Shares), by contrast, focuses on efficiency: computing costs fell from $2.51 for every dollar of revenue in the first quarter of 2025 to $0.48 in the second quarter of 2026, the first quarter to close with an operating profit ($974 million, with an 8% margin). If confirmed, these figures would be, in the analyst’s view, "the most compelling evidence to date" that AI can be an economically sustainable business.
The third quarter, however, cooled the picture: based on the prospectus figures, the ratio of computing costs to revenue rose again to around $0.58, and the operating result returned to a loss of $567 million. This is the key issue for Todorova: determining "whether the 8% margin will hold up as the new data centers come online".
The 85% gross margin should also be viewed with caution. According to Alberto Tocchio, Head of Global Equity and Thematics at Kairos Partners Sgr, it does not include all research and training costs: "The question is how much is left after funding not only the current model, but also the next one." Todorova points out that the figure comes from a note to the financial statements and that analysts' estimates indicate a margin closer to 65%.
On the customer front, however, the picture has improved. Amy (Moneyfarm) noted that nearly a quarter of 2025 revenue came from just two customers, while 47% came from sales through partners such as Amazon, Google, and Microsoft. In the first nine months of 2026, according to the prospectus, no customer accounted for more than 10% of revenue, and the top five accounted for about a quarter.
Political and legal risks remain. In March, the Pentagon classified Anthropic as a “supply chain risk”, a designation that, Todorova (Leverage Shares) notes, threatens its ability to sell to government agencies. As for copyright, the company still has to pay $1.2 billion of the $1.5 billion stipulated in the settlement agreement with authors.
The contest is also geopolitical. Roncarolo (GraniteShares) points out that AI is also a military technology and that, if necessary, "Washington will be hard-pressed to deny its champions in the ongoing competition with China financial support, more or less directly," despite tensions with the administration over the ethical limits imposed by Anthropic.
On the commercial front, Tocchio (Kairos Partners) warns: "Open-source models, particularly Chinese ones, are improving rapidly, and for many tasks the most powerful and expensive model may not be necessary." Alexis Bienvenu, portfolio manager at La Financière de l'Echiquier, quantifies the valuation gap: Moonshot AI, which develops the Kimi model and is reportedly preparing to go public, is worth about $50 billion—thirty to forty times less than Anthropic.
The technological comparison, according to Bienvenu, is more nuanced. In programming, Chinese models are almost on par (96 points for DeepSeek V4.1 Flash in the LiveBench 2.0 test, compared with 97.7 for Claude Fable 5.1), while the gap remains wide in long-running agentic tasks: around 78 points versus 41.
China’s advantage is cost: around $0.18 per successful operation with DeepSeek, compared with several dollars for Anthropic’s most advanced models, following DeepSeek’s breakthrough in 2025.
The biggest item, however, is infrastructure. The prospectus updated on November 2 puts spending commitments for computing and infrastructure at around $895 billion, of which $69.8 billion is due over the next twelve months and around $213.8 billion is subject to termination rights or conditions. At the end of September, the company had $131.9 billion in cash and short-term investments.
For Roncarolo (GraniteShares), the planning of these investments is the biggest risk, for Anthropic as well as its competitors, given the uncertainty over how long it will take to monetize them: free cash flow may not turn positive for several years.
Across the entire ecosystem, Tocchio (Kairos Partners) counts $2.1 trillion in debt and capacity contracts, compared with the $1.3 trillion in U.S. subprime mortgages in 2007. He clarifies that this is "a comparison of scale, not of identical risks", but it still leaves one question open: "who will pay if demand falls short of expectations?".
Complicating the picture is the circular nature of the financing, with suppliers funding the labs from which they receive computing orders. For Tocchio (Kairos Partners), it is necessary to "distinguish independent end demand from capital that flows back to suppliers in the form of orders", while Amy (Moneyfarm) points out that Anthropic's model "creates a dependency on partners that can simultaneously act as shareholders, suppliers and, in some segments, competitors".
Index mechanisms will also matter in the first few weeks. The fast entry rule, in effect since May 1, allows a newly listed company that ranks among the top 40 companies in the Nasdaq-100 on its seventh trading day to enter the index after 15 trading sessions. Amy (Moneyfarm) explains that its weighting would be calculated by multiplying the value of the shares offered by three, as provided for under the new index rules for mega-IPOs.
At the top end of the range, $64.35 billion, the initial weighting would be around 0.6% of an index that Amy estimates is worth approximately $31 trillion, and a higher price would increase it. Initial estimates point to an inclusion date of December 21, coinciding with the annual rebalancing. This would trigger purchases by ETFs tracking the benchmark, providing support for the stock, as also noted by Roncarolo (GraniteShares).
However, Amy (Moneyfarm) cautions that "the technical effects of a large IPO on prices can be significant in the short term, but are difficult to predict and potentially temporary". Once the technical flows have run their course, the market will return to fundamentals, because, as Tocchio (Kairos Partners) puts it, "revenue growth alone is no longer enough", and according to Roncarolo (GraniteShares), after the IPO the stock "could experience significant volatility".
In any case, the stakes go beyond Anthropic. For Todorova (Leverage Shares), a positive debut would confirm the economic viability of frontier AI and pave the way for OpenAI and other companies, while "a disappointing debut would force a reassessment of the valuations of all private artificial intelligence companies on the market".