- Anthropic plans to sell 825 million Class A shares at an indicative price of $70–$78, implying a gross offering size of $57.75–$64.35 billion.
- The offering would test public investor appetite for a stand-alone frontier-AI developer with massive infrastructure commitments and no clear path to profitability.
- Governance structure concentrates voting power among founders and a long-term benefit trust, limiting outside shareholder influence.
A Landmark Offering for an AI Unicorn
Anthropic is moving toward an initial public offering that would rank among the largest ever for a technology company, according to a draft prospectus reported by [Reuters (NASDAQ (NASDAQ:NDAQ):TRI)](https://www.roic.ai/quote/NASDAQ:TRI). The San Francisco-based developer of the Claude family of AI models plans to sell 825 million Class A shares at an indicative price range of $70 to $78 per share, implying a gross offering size of $57.75 billion to $64.35 billion. The filing also indicates that existing investors will sell shares alongside the company, meaning the total proceeds will not accrue entirely to Anthropic.
The share count and price range are proposed terms, not final pricing. Deal details were expected to be announced shortly, and Anthropic had not immediately responded to a request for comment. An earlier circulated document, which indicated roughly 866.7 million shares at $70–$80 and potential proceeds of $69.3 billion, was retracted after Anthropic told bankers those figures were inaccurate.
Anthropic formally announced its public S-1 filing on October 26 and intends to list on the Nasdaq Global Select Market under the ticker “ANT.” That announcement did not establish an offering price or final share count.
Rapid Growth, Mounting Losses
The offering comes against a backdrop of extraordinary revenue growth and equally extraordinary losses. Anthropic’s third-quarter 2026 revenue reached approximately $16.96 billion, up from $11.56 billion in the second quarter and about $1.38 billion a year earlier—a roughly twelvefold increase year over year and a 47% sequential gain. For the first nine months of the year, revenue totaled $33.24 billion, up approximately 1,200% from the prior-year period.
Yet the company’s net loss for the third quarter was $23.3 billion, and its first-nine-month net loss reached approximately $104 billion, including a $94 billion accounting adjustment related to financing agreements. Purchase and lease commitments stand at $895.4 billion, substantially related to computing and infrastructure, with about $70 billion due within 12 months. Those commitments are not all immediately payable debt, but they represent a substantial financing and execution burden.
Governance and Control
Public investors will have limited influence over Anthropic’s direction. Under the proposed structure, the seven co-founders’ “Founder LLC” will hold 50.1% of voting power on general shareholder matters. Separately, the Long-Term Benefit Trust generally has authority to elect and remove a majority of directors. Public Class A shares carry one vote each, but public investors will not collectively control ordinary shareholder votes.
Dario Amodei remains CEO, and Daniela Amodei is president and board chair. The company operates as a public benefit corporation with a stated mission of developing advanced AI for humanity’s long-term benefit—a structure that requires balancing shareholder financial interests with its stated purpose.
Market Reaction and Implications
The IPO would test whether public investors will fund a stand-alone frontier-AI developer at an enormous scale, rather than obtain AI exposure through diversified technology companies. It comes during a less favorable market backdrop: recent weakness in the Nasdaq, particularly in chipmakers and specialized AI-cloud companies, could weigh on demand for the offering.
The economic tension is straightforward: Anthropic’s customer revenue is growing rapidly, but supplying that demand requires extraordinary computing expenditure. Its $895.4 billion of commitments connect its prospects closely to data-center construction, semiconductor capacity, electricity availability, and financing conditions. A successful offering could support that investment chain; disappointing demand or tighter financing could expose its vulnerabilities.
Reports have also described a separate $26 billion placement involving Google (NASDAQ:GOOGL), Nvidia (NASDAQ:NVDA), AMD (NASDAQ:AMD), Broadcom (NASDAQ:AVGO), Micron (NASDAQ:MU), Microsoft (NASDAQ:MSFT), and TSMC (NYSE:TSM). However, that placement appeared in the earlier retracted material and was not confirmed in the latest Reuters report. It should therefore be treated as provisional.
Political and Regulatory Hurdles
Anthropic’s public-benefit status creates a governance issue for investors: the company must balance shareholder financial interests with its stated purpose and the interests of people affected by its conduct. Founder and trust control are designed to preserve that mission but materially limit outside shareholders’ influence.
Data centers have become a political constraint. Recent reporting described substantial opposition to AI infrastructure following the U.S. midterms, including voter rejection of data-center-related ballot measures in 40 of 41 cases. Concerns include utility bills, environmental effects, and potential employment disruption. Anthropic’s prospectus itself acknowledges increasingly organized opposition to data-center development. Raising capital does not by itself secure planning permissions, power connections, or public acceptance.
What’s Next
In the short term, the key milestones are publication of definitive offering terms, investor demand during marketing, final pricing, and confirmation of the primary-versus-secondary share split. The latest report does not establish a completed transaction or trading date.
Longer term, the central question is whether durable customer spending and improving economics can support the infrastructure commitments. The final prospectus and pricing announcement will be decisive.
Correction: An earlier version of this article misstated the implied gross offering size based on the retracted share count. The correct range, based on 825 million shares at $70–$78, is $57.75–$64.35 billion.