Analyst(s): Nick Patience
Publication Date: November 20, 2026
Anthropic priced the largest IPO on record at $83 per share and began trading on Nasdaq under the ticker ANT in November 2026. A concurrent $26 billion placement makes seven of its chip, cloud, and software suppliers shareholders, tying the newly public company even more closely to the vendors it buys compute from.
What Is Covered in This Article:
- The Anthropic IPO priced at $83 per share, above the $70 to $78 range, and sold 980 million Class A shares for $81.3 billion, the largest IPO on record.
- AMD, Broadcom, Google, Micron, Microsoft, NVIDIA, and TSMC bought $26 billion of nonvoting Class N shares at the IPO price in a concurrent private placement.
- The shares closed their first day of trading at $97.17, up 17%, giving Anthropic a market value of roughly $1.8 trillion.
- Futurum data shows enterprise demand for Claude remains strong, although most seat buyers also pay usage-based charges that make revenue harder to forecast.
- Public investors gain quarterly disclosure but little influence, because a founder-controlled share class holds 50.1% of voting power.
The News: Anthropic priced its initial public offering of Class A common stock at $83 per share in November 2026, above the $70 to $78 range it set at the start of its roadshow, and increased the offering from 825 million to 980 million shares. Anthropic sold about 956 million of those shares and existing stockholders sold the remaining 24 million, raising $81.3 billion in total and passing the $75 billion SpaceX raised in June as the largest IPO on record. Alongside the offering, AMD, Broadcom, Google, Micron, Microsoft, NVIDIA, and TSMC agreed to buy $26 billion of nonvoting Class N shares at the IPO price. Anthropic expects combined net proceeds of about $104.8 billion before any exercise of the underwriters’ option to buy a further 147 million shares. The shares began trading on the Nasdaq Global Select Market under the ticker ANT, opened at $111.50, and closed their first day at $97.17.
Anthropic IPO Makes Its Biggest Suppliers Shareholders
Analyst Take: Most coverage of the Anthropic IPO has focused on its size, $81.3 billion from public investors and close to $105 billion in net proceeds once the placement is counted, although we think the more consequential detail is who bought the placement. Seven companies that sell Anthropic chips, memory, manufacturing capacity, networking silicon, or cloud capacity now own part of it – or even bigger parts than they did before – and much of the money raised will go back to that same group as Anthropic buys compute. The deal shows how closely the frontier AI business now depends on a small set of suppliers that also act as investors, and it makes a dependence already visible in the S-1 harder for public shareholders to ignore.
Anthropic IPO Pricing Shows Appetite for a Frontier Lab
The $83 price valued Anthropic at about $1.5 trillion on a fully diluted basis, against $965 billion after its $65 billion Series H round in May 2026, and the first-day close lifted that to roughly $1.8 trillion, making it the seventh-most valuable US company less than six years after its founding. Anthropic is the first US frontier AI model developer to list, and until now investors who wanted exposure to the model layer had to buy it bundled with an advertising business or a cloud provider. The S-1 gave them reasons to pay a premium: first-half 2026 revenue of $16.3 billion, compute and infrastructure costs down to 55% of revenue, and an adjusted operating margin of 10%.
Enterprise Demand Supports the Anthropic IPO Valuation, With One Caveat
Futurum’s October 2026 Technology Spending Intentions Survey gives Anthropic a Net Score of 78 among 757 IT decision makers, a measure that subtracts the share of customers planning to cut or replace spending from the share planning to adopt or increase it. Its Pervasion, a measure of how widely it is deployed, reached 68%, up 39 percentage points from a year earlier, and although the Net Score slipped 6 points from the previous survey it remains 12 points higher year over year. Among 312 organizations already buying Claude seats in Futurum’s November 2026 AI Product Series survey, 71% expect their license counts to rise over the next six months, and 54% say Claude’s value outweighs its cost, against 19% who say the opposite.
The caveat concerns how those customers pay. Of the same 312 seat buyers, 74% also pay usage-based charges, and 53% say consumption is a meaningful or primary share of their Claude spend. Usage-based revenue grows quickly when customers expand coding and agentic workloads, and it can shrink just as quickly when they route work to cheaper models, optimize prompts, or tighten budgets. We expect public investors to scrutinize consumption trends each quarter as closely as the compute-cost ratio.
Suppliers on the Shareholder Register Tighten Anthropic’s Dependence
The placement buyers span the hardware Anthropic relies on, with TSMC manufacturing the chips, NVIDIA and AMD designing GPUs, Broadcom supplying custom silicon and networking, Micron supplying high-bandwidth memory, and Google and Microsoft providing cloud capacity, Google also making the TPUs Anthropic trains and serves on. Google and NVIDIA each contributed roughly $10 billion, according to press reports. Amazon, an Anthropic investor since 2023 and the supplier of its Trainium capacity, is not among the buyers.
The arrangement suits both sides in the short term, since suppliers lock in a customer whose S-1 discloses large multiyear compute commitments and Anthropic gets $26 billion of patient capital with no votes attached. However, it extends a pattern already familiar from NVIDIA’s investments in OpenAI and in neocloud providers such as CoreWeave, in which vendors invest in customers that then spend the money on the vendors’ products. When part of a supplier’s revenue depends on its own investment, demand becomes harder to read for everyone, including that supplier’s shareholders. Two of the seven buyers also compete with Anthropic at the model layer, since Google builds Gemini and Microsoft builds its own models alongside its large OpenAI stake. Their new holdings give them a reason to want Anthropic to succeed, but not one that outweighs their interest in their own models, at least in the case of Google, which is far ahead of Microsoft in model innovation and delivery.
Public Shareholders Get Disclosure, Not Control
The governance terms limit how much the Anthropic IPO changes for public shareholders. A founder-controlled LLC holds a single Class F share carrying 50.1% of voting power after the offering, the Long-Term Benefit Trust keeps its role in electing directors, and the placement shares carry no votes. Class A holders gain quarterly reporting and a market price, which will sharpen disclosure, but they will have little say in how Anthropic balances its public benefit purpose against returns. The co-founders’ pledge to give away 80% of their personal equity and the 5% of shares placed in a public benefit fund reinforce that purpose, although neither changes who controls the company.
For enterprise buyers, the Anthropic IPO brings one clear benefit: audited quarterly data on a supplier that many now depend on for coding and agentic work. We recommend that procurement and risk teams build that disclosure into vendor reviews, track the compute-cost ratio and consumption trends as early indicators of pricing pressure, and keep a multimodel strategy in place, since pressure on margins is likely to reach customers through pricing and usage limits.
What to Watch:
- Anthropic’s first quarterly report as a public company will show whether compute and infrastructure costs keep falling as a share of revenue, the ratio behind the 10% adjusted operating margin in the S-1.
- The underwriters’ option on a further 147 million shares, and later the expiry of lock-up restrictions on existing holders, will add supply that tests how durable first-day demand proves to be.
- OpenAI now faces pressure to list, and a second frontier-lab IPO would give investors a direct comparison of revenue quality, compute costs, and governance.
- The seven placement buyers must balance their roles as suppliers, investors, and in two cases model competitors, and how they do so will affect capacity allocation, pricing, and Anthropic’s plans to diversify into its own data centers and neocloud providers.
- Usage-based spending, which is a meaningful or primary share of Claude spend for 53% of current seat buyers in Futurum’s November 2026 data, will show whether coding and agentic workloads keep expanding through 2027 budget cycles.
See the company’s complete pricing announcement for the offering on the Anthropic website.
Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.
Other Insights From Futurum:
Frenemies in the Cloud: Anthropic’s IPO and Its Hyperscaler Reliance
Anthropic Files For IPO, Looking to Beat OpenAI to the Punch
Will Anthropic’s Draft S-1 Ignite a New Phase in the AI Platform Race?
Author Information
Nick Patience is VP and Practice Lead for AI Platforms at The Futurum Group. Nick is a thought leader on AI development, deployment, and adoption - an area he has researched for 25 years. Before Futurum, Nick was a Managing Analyst with S&P Global Market Intelligence, responsible for 451 Research’s coverage of Data, AI, Analytics, Information Security, and Risk. Nick became part of S&P Global through its 2019 acquisition of 451 Research, a pioneering analyst firm that Nick co-founded in 1999. He is a sought-after speaker and advisor, known for his expertise in the drivers of AI adoption, industry use cases, and the infrastructure behind its development and deployment. Nick also spent three years as a product marketing lead at Recommind (now part of OpenText), a machine learning-driven eDiscovery software company. Nick is based in London.
