Anthropic Files IPO Terms: 825M Shares, $70–78 Range, $26B Backing

Anthropic filed amended terms for its initial public offering on November 10, 2026, covering 825 million Class A shares at an estimated $70 to $78 each, alongside a contingent $26 billion private placement from seven of the biggest names in chips and cloud. At the top end of the range, the base offering would raise $64.35 billion in gross proceeds. The filing turns one of AI's most closely watched private companies into a public-market test of whether investors will fund the compute buildout behind frontier models.

The company itself is selling 801,351,354 shares, and selling stockholders are offering another 23,648,646 — with no proceeds from those secondary sales going to Anthropic. Anthropic's primary portion would generate roughly $62.5 billion before underwriting discounts and expenses. Underwriters also hold an option to buy up to 123,750,000 additional shares directly from Anthropic at the IPO price less the underwriting discount.

The stock is headed for the Nasdaq Global Select Market under the symbol ANT. The standard caveat applies: the prospectus remains preliminary, its terms may change, and the securities cannot be sold until the registration statement becomes effective.

What happened before this filing?

The public terms follow a retracted draft. Reuters reported on November 10, 2026, that an earlier version sent to banks proposed about 866.7 million shares at $70 to $80. Anthropic then sent a follow-up saying the pricing details were inaccurate and that it would update the filing. The amended terms — slightly fewer shares, a tighter band — are the corrected version.

The retraction episode is a reminder that even the largest IPOs of this cycle are being assembled under intense scrutiny. Every number in the filing now carries weight, because one mispriced draft already made headlines.

Who is putting up the $26 billion?

The private placement is the structural heart of the deal. Advanced Micro Devices, Broadcom, Google, Micron Technology, Microsoft, NVIDIA and TSMC Partners agreed — severally, not jointly — to buy an aggregate $26 billion of Class N stock at the IPO price.

The placement depends on the IPO's completion and other conditions, and it is scheduled to close immediately afterward. There is no underwriting discount or commission on those shares.

The Class N shares are not registered in the IPO and are not covered by the underwriters' lock-up agreements. They are restricted securities, which means resale requires an effective registration statement or an available exemption.

Separately, certain directors and affiliated entities indicated interest in buying up to $303 million of IPO shares. Those indications are nonbinding — they permit more, fewer, or no purchases at all.

The buyer list reads like a map of Anthropic's supply chain. Chip designers, a chip manufacturer's partner, and three of the largest cloud and AI platform companies are all taking direct equity positions, tying their financial interests to Anthropic's compute-driven growth.

Why does the revenue growth matter?

Anthropic reported $17.0 billion of revenue for the three months ended September 30, 2026 — up 1,127% year over year. Over the same period, compute and infrastructure expenses grew 382%, and the company presented that disparity as evidence of compute efficiency. Revenue is outrunning the cost of the machines that power it, at least for now.

The customer base disclosed in the filing shows where that revenue comes from:

• More than 11,900 customers generating at least $100,000 in trailing-12-month revenue as of September 30, 2026
• More than 1,900 customers generating at least $1 million
• Approximately 250 customers generating at least $10 million

Anthropic uses the $100,000 tier as its measure of enterprise adoption, on the reasoning that spending at that level generally reflects organizational rather than individual usage. In plain terms: this is not a consumer app with a thin subscription layer — it is a business selling to businesses at scale.

How big are the infrastructure commitments?

Very big. Anthropic's purchase commitments and lease agreements total approximately $900.9 billion, substantially all related to compute and infrastructure. That figure covers commitments in place on September 30, 2026, others entered into afterward, and leases not yet commenced.

The timing breaks down like this: approximately $451.6 billion is expected to be incurred before December 31, 2030, substantially all of the remainder falls between 2031 and 2045, and approximately $70.7 billion comes due in the 12 months after September 30, 2026.

Those numbers explain why the IPO and the $26 billion placement exist. A company with nearly a trillion dollars in commitments needs capital channels beyond private rounds — and the balance sheet already reflects that scale. As of September 30, 2026, Anthropic held $95.6 billion in cash and cash equivalents plus $36.3 billion in short-term investments, and had raised $141.1 billion in aggregate capital by that date.

The commitments stretch out to 2045. Public investors are effectively being asked to underwrite a two-decade compute buildout.

Who actually controls the company?

Anthropic is a Delaware public benefit corporation, which requires its board to balance stockholders' financial interests, the interests of those materially affected by its conduct, and a charter purpose of maintaining and developing advanced AI responsibly for humanity's long-term benefit.

Control runs through two vehicles:

• The Long-Term Benefit Trust. Through Class T stock, the Trust generally can elect and remove the Class T directors, who will constitute a majority of the authorized board. Class T stock carries no economic rights and no general stockholder voting rights.
• The founder vehicle. Immediately before the offering completes, Anthropic will issue one Class F share to a limited liability company organized by its co-founders. That single share gives it 50.1% of aggregate voting power on general stockholder matters — including elections of directors other than the Class T directors — but no economic rights.

Anthropic qualifies as a controlled company under Nasdaq's governance rules because the Trust can generally elect a majority of the authorized board.

The Class N shares, including those in the $26 billion private placement, generally lack voting rights. They convert into Class A shares upon sale or transfer to an unaffiliated third party.

What does this mean for public investors?

The structure creates a clean division of risk and power. Public shareholders would bear the economic exposure to Anthropic's growth and its roughly $900.9 billion in infrastructure commitments, while board appointments and general voting control stay with the Long-Term Benefit Trust and the founder vehicle. Investors buying ANT shares get the upside and the obligations, not the governance.

That arrangement is not unique among tech listings, but its scale is. A $64.35 billion offering, a $26 billion strategic placement, and a trillion-dollar commitment stack make this one of the largest capital raises in AI's history — and a direct referendum on whether public markets will finance the next decade of compute.

The pricing terms remain preliminary, and the registration statement must become effective before any shares change hands. If the deal closes on these terms, the chip and cloud giants buying in at $26 billion will have aligned themselves with Anthropic's compute roadmap well into the 2040s.