Dr Kamran Jalali

Anthropic’s $119.5B IPO Meets $900B Compute Commitments

Anthropic IPO: the $119.5B deal reveals a surprising split between investors and control.

Introduction

Anthropic’s IPO and its private placement produced a combined gross deal worth $119.5 billion. Yet the company’s own prospectus describes roughly $900.9 billion in contractual obligations, most of them tied to computing power and infrastructure.

That is the part of the listing worth reading twice. The headline is a record-breaking stock market debut. The deeper story is how an AI company raises capital, locks in the machines it needs, and decides who gets a say when those two things collide.

The Anthropic IPO began trading on Nasdaq on November 19 under the symbol ANT. On November 20, Anthropic confirmed that the underwriters had exercised their full option and that its separate private placement had closed. The combined gross proceeds were about $119.5 billion before fees and other expenses, according to the company.

But this was not $119.5 billion paid entirely into Anthropic’s bank account, and the $900.9 billion figure is not due tomorrow. The details reveal how public investors, strategic suppliers and company control fit together.

Main Section: What the Anthropic IPO Actually Raised

Supporting Topic: Three numbers, three different meanings

Anthropic sold 1.127 billion Class A shares at $83 each after the underwriters exercised their option to buy another 147 million shares. The offering generated about $93.5 billion gross. Anthropic received roughly $91.6 billion, while selling stockholders received the rest.

That difference matters: proceeds from the 23.65 million shares sold by existing stockholders went to those sellers, not Anthropic.

Then there was a second transaction. Seven companies agreed to buy 313.25 million shares of Anthropic’s Class N stock in a private placement for about $26 billion. The buyers were AMD, Broadcom, Google, Micron, Microsoft, Nvidia and TSMC Partners.

Add the public offering and private placement, and the combined gross deal reaches about $119.5 billion. That is the figure Anthropic announced. It is a useful headline number, but it combines two different types of financing and includes money paid to selling shareholders.

This distinction shows who supplied the capital and what influence they received.

Supporting Topic: Why seven technology companies bought Class N

Class N shares have economic rights similar to Class A shares, but they do not carry ordinary voting rights, except where the law requires otherwise. The SEC filing says Class N holders are primarily strategic business partners. The shares generally convert into Class A shares if transferred to an unaffiliated third party.

The buyers sit across the AI supply chain, from chip design and memory to semiconductor manufacturing and cloud infrastructure.

The prospectus does not say that every buyer has the same commercial relationship with Anthropic, and it would be wrong to treat them as one coordinated group. But the structure is still worth watching. A company that needs access to chips, data centers and cloud services has brought several major technology suppliers into its shareholder base.

Advanced AI models need enormous computing capacity, and securing it years ahead can matter as much as developing the model. Strategic investors may bring capital and supply access, but the company can become more dependent on firms with their own AI products.

Main Section: The $900 Billion Number Needs Context

Supporting Topic: A long-term commitment, not a bill due today

The most striking figure in Anthropic’s final prospectus is its estimate of $900.9 billion in contractual obligations, short-term and later commitments, and leases not yet started. The figure combines the company's September 30, 2026 reporting with subsequent commitments.

Anthropic says substantially all of that amount relates to compute and infrastructure. The figure includes roughly $451.2 billion in estimated future lease obligations for leases signed but not yet started, and about $105.7 billion in commitments entered into after September 30. The company also listed around $17.4 billion in minimum payments under certain short-term contracts.

Do not read this as a $900.9 billion loan or an immediate cash payment. These are commitments spread across future years. Some are conditional, some depend on delivery, and the company said about $186.2 billion of the aggregate amount relates to arrangements with termination rights or conditions that must be met before they begin. It also excluded other arrangements without fixed dollar commitments, which it estimated could involve another $150 billion to $175 billion in payments.

The point is not that Anthropic is about to run out of money. Frontier AI requires infrastructure spending on a scale that makes even a record IPO one part of a much larger financing plan.

Supporting Topic: Why compute changes the business model

Compute is the processing power needed to train and run AI models. It depends on chips, servers, electricity, data centers and networks. Customer growth means little if the company cannot afford enough hardware to serve those customers.

Anthropic’s filing shows how quickly those costs can grow. For the nine months ended September 30, 2026, it reported $33.24 billion in revenue, compared with $2.55 billion in the same period a year earlier. Over that same nine-month period, compute and infrastructure expenses reached $18.81 billion, up from $4.38 billion.

The business is growing fast, but its cost base is expanding too. Compute and infrastructure represented 57% of revenue in those nine months, according to the prospectus. The company also reported that $6.5 billion of compute during that period came from short-term commitments that cost significantly more than comparable capacity secured through multi-year agreements.

Long-term contracts can secure supply and pricing, but they create obligations before demand is certain. Anthropic says many compute commitments must generally be met regardless of revenue or utilization.

That is the central business risk. If demand keeps rising, locked-in capacity can support growth. If demand slows, expensive capacity can become a burden. The IPO does not remove that risk. It gives Anthropic a much larger pool of capital with which to manage it.

Main Section: The Investors Do Not Get the Final Say

Supporting Topic: Public shares are not the same as public control

The capital structure is the other part of the prospectus that deserves attention. A buyer of Anthropic Class A shares receives one vote per share. The seven strategic investors buying Class N shares receive no ordinary voting rights.

Even Class A shareholders do not control the company in the usual way. Anthropic’s prospectus says a single Class F share, held by a company organized by the seven co-founders, carries 50.1% of the total voting power on matters put to shareholders generally, including the election of directors other than the special Class T directors.

There is another layer. Anthropic’s Long-Term Benefit Trust, through its Class T share, generally has the power to elect and remove the Class T directors, who together make up a majority of the authorized board. The trust’s share has no ordinary economic rights.

Public investors can own an economic stake without having the final say. The founders hold a majority of general shareholder voting power, while the trust selects most board members.

This is not a secret hidden in fine print. It is set out in the SEC prospectus. But it changes what the word “public” means in this case. Anthropic is publicly traded. That does not mean it has a conventional shareholder-control model.

Supporting Topic: The public benefit promise has a real trade-off

Anthropic is incorporated as a public benefit corporation, or PBC. That is a for-profit company whose directors must balance financial interests with the interests of people affected by its actions and a stated public benefit. Anthropic’s stated purpose is the responsible development and maintenance of advanced AI for humanity’s long-term benefit.

The argument for this design is that safety decisions may require delaying a launch, refusing a customer or spending more on testing. A trust with board appointment powers may protect those choices from short-term market pressure.

There is also a serious counterpoint. Public shareholders may disagree with decisions that reduce near-term returns, while the founders and trust have strong control over the company. The structure asks investors to trust the mission and the people charged with protecting it. It does not give them the same influence they might expect at a more conventional listed company.

A strong structure can protect a mission but make outside accountability harder. The test will be how the board handles conflicts between safety, competition, infrastructure costs and investor expectations.

Main Section: The Conflict Built Into the AI Supply Chain

Anthropic’s prospectus makes another unusually direct disclosure: some cloud service providers it relies on for computing are also strategic investors, customers and competitors offering their own AI models and services. The company warns that these relationships could create conflicts or competitive pressures that affect compute access, cost or workload priority.

That is the mechanism readers should watch after the IPO. The same companies can help finance Anthropic, sell it infrastructure, distribute its products and compete for the same customers. Those roles can support growth, but they do not always point in the same direction.

This is not automatically improper. A cloud provider can earn hosting revenue and invest in a customer at the same time. The risk is dependence, not proof of wrongdoing.

Anthropic says it is shifting toward a hybrid model that includes leasing chips directly and placing them in dedicated facilities, rather than relying only on fixed-term cloud agreements. That approach may give it more flexibility over time, but it also requires careful planning and more capital.

The IPO matters because Anthropic is financing the physical foundation of its business. Its success depends on Claude's appeal and on securing enough compute at a cost its revenue can support.

FAQ

How can I buy Anthropic shares?

Anthropic shares trade on the Nasdaq Global Select Market under the ticker ANT. Investors need a brokerage that supports US-listed shares, and purchases after listing take place at the prevailing market price rather than the original IPO price.

How much money did Anthropic raise in its IPO?

The public offering generated about $93.5 billion in gross proceeds, of which Anthropic said it received about $91.6 billion. A separate $26 billion private placement brought the combined gross deal to about $119.5 billion before fees and expenses.

Who invested $26 billion in Anthropic?

AMD, Broadcom, Google, Micron, Microsoft, Nvidia and TSMC Partners agreed to buy Class N shares. These shares do not carry ordinary voting rights, except where required by law.

What is the difference between Anthropic Class A and Class N shares?

Class A shares carry one vote per share. Class N shares have similar economic rights but no ordinary voting rights, and generally convert to Class A shares if transferred to an unaffiliated third party.

Who controls Anthropic after the IPO?

The founders’ Class F share carries 50.1% of the general shareholder voting power. Anthropic’s Long-Term Benefit Trust also generally elects and removes the Class T directors, who make up a majority of the authorized board.

Does Anthropic really have $900 billion in debt?

No. The roughly $900.9 billion figure refers to a broad set of contractual obligations and infrastructure commitments across future years, not a loan or an amount due immediately. Some commitments are conditional or subject to termination rights.

Conclusion

Anthropic’s IPO is more than a record fundraising event. It is a case study in the new economics of frontier AI: huge upfront infrastructure commitments, strategic investors who can also be suppliers and competitors, and a governance system that separates public ownership from much of the company’s control.

The $119.5 billion headline is impressive. The $900.9 billion commitment figure explains why the company needs so much capital, while the voting structure explains why buying its shares does not mean buying the final say.

My view is that investors should judge Anthropic on three things over time: whether Claude keeps attracting paying customers, whether compute costs become more efficient, and whether the governance structure can protect its stated mission without weakening accountability.

If you were considering ANT shares, would the founders’ and trust’s control make you more confident in Anthropic’s long-term mission, or less willing to invest as a public shareholder?

Disclaimer: This article is for information and education only. It is not financial advice or a recommendation to buy or sell ANT shares. IPO prices and market values can change quickly. Always review the official prospectus and consider your own risk tolerance before investing.

Did you enjoy this article?


Crypto Strategist
Crypto Strategist

I am Dr. Kamran Jalali, Crypto researcher & educator. Deep analysis on crypto trends, AI tokens, RWA, and smart money, in plain language. No hype. Just honest research to help you make smarter decisions.


Dr Kamran Jalali
Dr Kamran Jalali

Most people lose money in crypto not because the market is against them — but because nobody ever taught them the rules of the game. I am Dr. Kamran Jalali. I write about crypto in plain, simple language that anyone can understand — no confusing jargon, no hype, no false promises. Here you will find honest breakdowns of how crypto really works, why traders fail, how to protect your money, and how to make smarter decisions in the digital asset world. Whether you are completely new to crypto or have been in

Publish0x Publish0x

Reward the author with $0.01 in crypto, and earn yourself as you read!

20% to author / 80% to me.
Rewards are FREE. Publish0x pays them, not you.

Page not displaying correctly?