Todd Underwood, a former Anthropic engineering leader, spent the months before its stock market debut fielding calls from investment managers. They wanted an insider’s view of the business strategy behind Claude, the company’s artificial intelligence assistant. Underwood thought an asset he valued was getting little attention.

Few asked about the governance.

Anthropic debuts on the Nasdaq on Thursday under the symbol ANT with a corporate structure designed to let it put AI safety ahead of shareholder returns. The seven co-founders hold a combined 50.1% of the voting control through a share class created this fall.

The Breakdown

AI-generated summary, reviewed by an editor. More on our AI guidelines.

ANT on the Nasdaq, live

$83.00+$0.00 (+0.0%) vs. $83 IPO price
Last trade: Nov 19, 2026 10:08 AM ETSession: Open
Thu 7:14 AM PT$83.00

Checked every 30 minutes until Sat., Nov. 21, 6 a.m. PT; the box changes only when the stock trades. Prices come from Nasdaq's quote feed and can run up to 15 minutes behind. Not investment advice.

The price

The offering priced Wednesday at $83 a share, above the expected range of $70 to $78. Anthropic and selling shareholders expanded the sale to 980 million shares from a planned 825 million, raising $81.3 billion. The price puts the company’s valuation above the $1.4 trillion Anthropic had been targeting.

A separate $26 billion private placement of shares brings in companies including Google, Microsoft, Nvidia and AMD. Anthropic buys costly systems from Nvidia, Google, Amazon and AMD.

The prospectus lists revenue of $33.24 billion for the nine months ended Sept. 30, up roughly 1,200% from a year earlier, and a $104 billion net loss over the same period, most of it from the change in fair value of convertible notes and derivative liabilities.

The offering includes shares sold by the founders, and Anthropic receives none of the proceeds from those sales. Most of the shares being offered are newly issued by the company.

The prospectus gives investors a direct warning: “Our product and investment decisions may not prioritize short-term financial results.” Those choices “could adversely affect the market price” of its stock.

Underwood expects that warning to matter. The company “will choose to do things in favor of safety rather than in favor of making money,” he said. He agrees with that priority and thinks it could discourage some investors.

Paul Abrahimzadeh, a senior partner at the technology investment firm 1789 Capital, sees the elaborate governing arrangements as something investors may struggle to understand. He still expects demand. “You have to own it. The question is at what price?”

Where it came from

Anthropic began in 2021 as a public benefit corporation, founded by seven people who left OpenAI. Siblings Dario and Daniela Amodei led the group. Dario became chief executive; the other founders are Jared Kaplan, Sam McCandlish, Tom Brown, Jack Clark and Christopher Olah.

The public benefit designation puts a social mission into the company’s legal duties. Directors must balance shareholders’ financial interests with the company’s stated public purpose. Anthropic’s purpose is the responsible development of advanced AI for humanity’s long-term benefit. It pays corporate taxes like an ordinary company.

In 2022, Amodei described a business that would follow the leading AI developers closely enough to study their technology without pushing much further ahead. Anthropic would do safety research while others set the pace.

That summer, the company held back Claude because of internal safety concerns. OpenAI released ChatGPT first and gained a lead with consumers. Inside Anthropic, the decision became an example of accepting a commercial sacrifice for safety. Amodei later called it a “one-off.”

Since the release of Claude Code, Anthropic’s quarterly revenue has surpassed OpenAI’s, reaching $17 billion in the third quarter of 2026. Claude’s coding tools helped drive the growth, allowing programmers to direct AI agents in ordinary language rather than write every line themselves.

Amodei now describes the strategy as a “race to the top.” Leading the market would give Anthropic enough influence to persuade competitors to adopt safer practices. The company has embedded safety auditors with engineering teams and backed California legislation requiring greater transparency from leading AI developers.

Jasmine Sun, a writer who has interviewed dozens of current and former employees over the past two years, writes in her essay that the strategy assumes powerful AI will be built regardless of whether Anthropic participates. She also identifies its assumption that Anthropic is the most responsible developer among the companies that could lead.

How Anthropic got here

Inside the company

Some of the current and former Anthropic employees Sun interviewed described the IPO with a shrug. They treated it as another funding round. Conversations about newly liquid shares often concerned where charitable donation accounts should direct the money. “The money is secondary,” Sun wrote.

That picture comes from Sun’s interviews, rather than a survey of the workforce. Few employees she interviewed planned to sell immediately, and some expected their shares to become much more valuable.

Employees call themselves Ants. In the company’s bare, high-security San Francisco offices, books on display include Richard Rhodes’s “The Making of the Atomic Bomb” and printouts of writings by Amodei and the in-house philosopher Amanda Askell. Some employees share modest apartments and spend much of their time at work.

A hiring question asks candidates how they would feel if safety slowdowns drove the value of their stock to zero. Research teams can reject technically gifted applicants over cultural fit. Early employees question whether colleagues’ decisions satisfy the mission, and some worry that newer sales and marketing hires care less about superintelligence.

Some staff have asked whether Anthropic should stop training more capable models even without a matching commitment from competitors. Amodei rejected that proposal, saying the company could control its pace without surrendering its position.

Most former employees Sun interviewed considered Amodei deeply principled and willing to make sacrifices for safety. They also acknowledged that the pressure created by rapid growth could exceed a conscientious founder’s ability to control it.

Sun’s criticism of the strategy concerns how much the public is being asked to trust those judgments. Safety skeptics also dispute whether current evaluations can establish that a model is safe enough to release. Tests in which Claude only pretended to follow its instructions are part of that disagreement.

Who holds the votes

Anthropic created its Long-Term Benefit Trust in 2023 to give independent people a role in enforcing the mission. The trust now has five members, including former Federal Reserve chair Ben Bernanke. Its trustees do not own Anthropic stock.

The governing arrangements give the trust a special share with voting rights and no financial entitlement. The trustees cannot sell it or profit from a rising share price. Their authority works through appointing and removing directors.

At the IPO, the trust directly elects three board members. Other shareholders elect another three. A seventh seat is subject to the consent of a nominating committee expected to be controlled by the founders, limiting the trust’s ability to command a board majority independently.

The seven founders received a new share class this fall that gives them a combined 50.1% of voting control. Public investors buying into the offering therefore have limited power over who runs the business.

Jesse M. Fried, a Harvard law professor who studied the corporate structure, described the trust’s authority as “watered down” by the arrangements around the additional seat.

Stavros Gadinis, a law professor at the University of California, Berkeley, explained the practical limit: trustees who disagree with management must try to remove directors, and may be unable to remove a majority. The structure has not been tested under the pressure of public trading.

Glenn Davis, executive director of the Council of Institutional Investors, represents an organization whose members include public and union pension funds. It has long opposed arrangements that give founders voting power beyond their financial stakes. On Anthropic, the council will “wait and see,” he said.

The founders’ control means that “ultimately the founders are accountable to themselves,” Davis said.

Know someone who'd find this useful? ✉️ Email it to a friend in one click, or they can subscribe free here.

Douglas Chia, a senior fellow at Rutgers Law School who advises companies on governance, has struggled to find a precedent. “I don’t know if I can really compare it to anything else,” he said. “It’s going to be an interesting experiment.”

Who runs Anthropic

The fights

Anthropic’s refusal to accept unrestricted military use of Claude brought a direct commercial penalty this year. It sought safeguards against mass domestic surveillance and fully autonomous weapons. The Pentagon demanded access for any lawful use and disputed the company’s right to constrain military decisions.

In March, the Defense Department designated Anthropic a supply-chain risk, blocking military use and contractors’ use of Claude in work for the department. Anthropic sued in separate federal courts because the government had relied on distinct legal designations.

A San Francisco federal judge ruled one designation illegal in August. On Sept. 25, a divided federal appeals panel in Washington upheld the other. The appellate court delayed the decision taking effect to allow Anthropic to seek further review. An Anthropic spokesperson said the company was considering all options, including further review.

Thriller novelist Andrea Bartz brought a copyright suit with other authors in 2024 over books used to train Claude. A federal judge approved a $1.5 billion settlement in July 2026, providing about $3,000 for each covered book. The ruling covered more than 482,000 books.

The court had distinguished between training a model on copyrighted works and obtaining those works through piracy. Training was held to be fair use in that case; Anthropic’s acquisition of millions of books from pirate sites was wrongful. The settlement addressed the pirated copies.

A July 2025 internal memo exposed another difficult choice. Amodei said Anthropic would pursue investment from the United Arab Emirates and Qatar after earlier resistance to some regional funding. He acknowledged that accepting the money could benefit authoritarian rulers.

“This is a real downside and I'm not thrilled about it,” he wrote.

Amodei proposed keeping the investments narrowly financial, without moving training infrastructure into those countries. He also acknowledged that the prospect of future funding could give investors influence. Competitors’ access to Gulf capital had put Anthropic at a disadvantage, he wrote.

In 2024, Anthropic decided not to accept money from Saudi Arabia, citing national security concerns. The company has not said whether that stance has changed.

The controversies

The money and the mission

A draft federal AI bill circulated by Anthropic’s lobbyists in October angered some employees. It would preempt certain state safety laws and restrict the circumstances in which the government could block an advanced model’s release. Amodei defended the proposal publicly, and the policy team promised revisions.

Independent evaluators registered with the Commerce Department would periodically evaluate a lab and its frontier models, and the lab could pay them if payment was not contingent on the results. Evaluators finding an “imminent enumerated risk” would have to report it to Commerce within 72 hours of completing their assessment. Only a federal judge in the District of Columbia could order a model suspended, and the draft bars any other U.S. “department, agency, or officer” from suspending its development, deployment or internal use.

Anthropic has separately committed shares to a Public Benefit Fund, intended to support charitable donations or direct distributions to selected populations. The announced starting stake was 5% of company shares, subject to shareholder approval. The founders pledged at the company’s formation to donate 80% of their equity holdings.

The fund is separate from the trust that appoints directors. It can distribute economic benefits; it does not decide when Claude is safe to release.

Anthropic’s latest prospectus lists more than $900 billion in outstanding spending commitments. Much of the IPO funding will support the computing capacity needed for further AI development. Yang Wang, a financial analyst at Counterpoint Research, said the business must show it “can keep pace” as obligations come due.

The offering is expected to close on Nov. 20, subject to customary conditions. Underwriters have a 30-day option to buy up to 147 million additional shares.

Frequently Asked Questions

What did Anthropic's IPO raise, and at what price?

Anthropic priced its offering at $83 a share on Wednesday, above the expected range of $70 to $78. It expanded the sale to 980 million shares from a planned 825 million and raised $81.3 billion. A separate $26 billion private placement brings in companies including Google, Microsoft, Nvidia and AMD.

Who controls Anthropic after the IPO?

The seven co-founders, including Dario and Daniela Amodei, hold a combined 50.1% of the voting control through a share class created this fall. Public investors buying shares in the offering have limited power over who runs the company.

What is the Long-Term Benefit Trust?

A five-member body created in 2023 whose trustees, including former Federal Reserve chair Ben Bernanke, own no Anthropic stock. It directly elects three board members. A seventh seat needs consent from a nominating committee expected to be controlled by the founders, which Harvard law professor Jesse M. Fried said 'watered down' the trust's authority.

Is Anthropic profitable?

No. The prospectus shows revenue of $33.24 billion for the nine months ended Sept. 30, up roughly 1,200% from a year earlier, and a $104 billion net loss over the same period, most of it from the change in fair value of convertible notes and derivative liabilities.

Why did the Pentagon blacklist Anthropic?

Anthropic sought safeguards against mass domestic surveillance and fully autonomous weapons, while the Pentagon demanded access for any lawful use. In March, the Defense Department designated the company a supply-chain risk. A federal judge ruled one designation illegal in August, and an appeals panel upheld the other on Sept. 25.

AI-generated summary, reviewed by an editor. More on our AI guidelines.

Anthropic Hauls $81.3 Billion Out of a Jittery Market
Anthropic priced its initial public offering at $83 a share on Wednesday, above its expected range. The sale raised $81.3 billion. The stock is expected to begin trading Thursday on the Nasdaq under t
Can Anthropic's Revenue Outrun a $900 Billion Bill?
Anthropic’s amended filing on Tuesday, Nov. 17, puts Amazon and Google’s combined ownership after its IPO and private placement at 28.4%, as the company guides investors toward a price at the top of i
Anthropic aims for $1.4 trillion value; SoftBank pays for buying too early
IMPLICATOR .ai Morning Briefing · From San Francisco   Wednesday, November 11, 2026 9 stops From San Francisco 1 The Editorial   Good morning. AI's ambitions carry
AI News

San Francisco

Editor-in-Chief and founder of Implicator.ai. Former ARD correspondent and senior broadcast journalist with 10+ years covering tech. Writes daily briefings on policy and market developments. Based in San Francisco. E-mail: editor@implicator.ai