Anthropic Sets IPO at $83 as AI’s Computing Economics Go Public

News Summary
Anthropic priced its initial public offering at $83 per share on Wednesday, November 18, 2026, Eastern Time, above its previously stated range of $70 to $78. The maker of Claude expanded the offering to about 980 million shares, and the IPO valued the company at roughly $1.5 trillion on a fully diluted basis, according to reporting published that evening. Its Nasdaq debut under the ticker ANT was scheduled for Thursday, November 19, Eastern Time. The pricing is a market milestone, but for technology learners it also puts a practical question in focus: can demand for advanced AI software grow fast enough to support the computing infrastructure required to build and run it?
Pricing and scale
The $83 price is about 6.4% above the top of Anthropic’s initial $70–$78 range. The company increased the number of shares offered by roughly 19% to about 980 million, expanding the public offering as investor demand developed. At the reported price, the public share sale represents about $81.3 billion. A concurrent private placement of $26 billion at the same per-share price brought the combined transaction value to approximately $107.3 billion.
The valuation depends on how shares are counted. On a fully diluted basis, which includes potential shares from instruments such as stock options and restricted stock, Anthropic was valued at about $1.5 trillion. A calculation excluding those potential shares put the figure near $1.1 trillion. These are different measures of company value, not different prices paid for the listed Class A shares.
The AI business behind the listing
Anthropic’s growth has been linked to demand for Claude models and developer tools such as Claude Code. Its filing reported an annualized revenue run rate of $81 billion as of October 31, 2026, and 11,900 customers with at least $100,000 in trailing 12-month revenue as of September 30. A run rate extrapolates recent revenue over a year; it is a snapshot-based measure rather than a guarantee of future sales.
The company’s reported figures also show why AI economics are closely tied to infrastructure. Anthropic spent nearly $9 billion on computing capacity and infrastructure in the first half of 2026, almost four times the comparable prior-year amount, and incurred a further $5 billion in computing costs in the third quarter. Revenue rose 47% quarter over quarter to $16.96 billion in the third quarter, while operating results shifted from income in the previous quarter to a $567 million loss. Those figures illustrate how rapid sales growth can coexist with substantial and changing costs.
What computing costs mean
Training and serving large AI models require specialized chips, data centers, electricity, networking, and engineering. Training uses large computing clusters to adjust a model’s parameters; serving, also called inference, uses computing resources each time people or organizations make requests. As usage expands, a company must keep investing in capacity while improving how efficiently each unit of computing produces useful output.
Anthropic disclosed nearly $900 billion in long-term computing commitments with infrastructure partners, including Amazon and Google. Such commitments can secure access to capacity needed for development and customer services, while also creating long planning horizons and reliance on suppliers’ ability to finance and build facilities. The company’s public filings explain that infrastructure availability and cost can affect delivery of contracted capacity.
Why the IPO matters to technology education
A public listing makes more regular financial disclosures available to a wider audience. Students, developers, and researchers can use future reports to study how model adoption, subscription and enterprise revenue, computing expenses, and product efficiency move together. Comparing revenue growth with infrastructure spending can help distinguish software demand from the cost of delivering AI at scale.
The $83 IPO price records what investors agreed to pay at the offering; it does not establish what the shares will be worth later or resolve questions about long-term profitability. The company’s market debut therefore offers a live case study in the relationship between technical progress and business fundamentals: better models may attract more users, while each additional user can bring new computing costs. Future disclosures will show whether efficiency improvements and revenue growth can keep pace with that demand.
This article was compiled by the AIBARS editorial team with AI assistance. AI can make mistakes, so please check the original source for anything important. Spotted an error? Email [email protected].