Exclusive Report - Anthropic Executives Control AI Lab Through "Founders' Limited Liability Company," Prioritizing Public Interest

Echo Wang/Ross Kerber/Jeffrey Dastin

Reuters, September 29 - Anthropic has long positioned itself as an “ethical” artificial intelligence company. A preliminary review of its IPO filing shows that the company is creating a new entity called “Founder LLC”, designed to serve the public interest while shielding the company’s leadership from market pressures.

According to a copy of the prospectus seen by Reuters, this new LLC will initially be comprised of Anthropic’s seven co-founders, including CEO Dario Amodei, who are described as “uniquely qualified to bear the responsibility of stewarding our mission.” Anthropic’s mission is to benefit humanity with responsible AI.

Earlier on Monday, Reuters exclusively reported details of the AI company’s IPO prospectus (link).

The documents state that the developer of Claude AI will continue to operate as a public benefit corporation (PBC) under Delaware law, which formally allows the company’s leadership to balance investor interests with the interests of all humanity.

Under this arrangement, a majority of the seven co-founders will decide how a single Class F share votes, holding 50.1% of total voting power on key matters, including electing certain directors and other items submitted to investors.

The remaining four classes of Anthropic shares serve different functions or restrictions, including minimum voting rights for strategic partners. While Class A common shares available to general investors carry one vote per share, Anthropic’s novel capital structure may effectively weaken ordinary investors’ influence.

An Anthropic representative declined to comment. Some details of this structure were previously reported by The Information in early September.

According to the documents, Anthropic’s corporate structure may result in decisions that “conflict with the short-, medium-, or long-term financial interests and business performance, which could potentially negatively impact the value of our Class A common stock.”

In theory, this approach should not be controversial. Anthropic states in the filing that one of its core strengths is “a culture advocating humility and the pursuit of truth, where cutting-edge AI capabilities and safety go hand-in-hand.”

Meanwhile, corporate history is not short of such cases: visionary CEOs have created enormous value, but also sometimes ignored some investors’ perspectives and ultimately paid the price.

Meta Platforms META.O in August agreed to pay as much as $18 billion to address concerns about children’s safety after resisting shareholder activists’ (link) attempts for years to raise these policy issues at annual meetings. Electric vehicle maker Tesla TSLA.O is also known for its highly volatile stock price (link), driven largely by CEO Elon Musk’s (link) social media posts.

A tightly-knit leadership team

Anthropic has previously conducted business with Meta and SpaceX SPCX.O, where Musk is also CEO. While SpaceX’s latest IPO further centralized power around Musk (link), Anthropic’s plan involves spreading authority from CEO Amodei to a tightly-knit executive team, core to the company’s operations.

This seven-person core team remains united five years after the company’s founding, with some members having known each other for over a decade. The team left rival OpenAI in 2020 and has often differed with peer companies to advance its own vision for AI safety.

Besides Amodei, Founder LLC members include his sister and company President Daniela Amodei; Chief Compute Officer Tom Brown; and Chris Olah—who recently met Pope Leo (link) and leads critical research initiatives.

Daniela Amodei also serves as chair of Anthropic’s board. According to the filed documents, after the IPO, she, her brother, and another yet-to-be-announced director will be chosen by both Class F and Class A shareholders.

The remaining four board members will be elected by another oversight body—the startup’s “Long-Term Benefit Trust.” Current trustees include former Federal Reserve chair Ben Bernanke and national security expert Richard Fontaine.

Anthropic’s strategy aims to leverage its management team’s collective expertise to develop AI responsibly. For safety reasons, the cutting-edge lab has already restricted or delayed launching certain new features, for example, setting up restricted access for Mythos Preview—an AI model particularly potent in cybersecurity.

“Similarly, to focus compute resources on research and safety priorities, we have chosen not to develop certain commercially attractive products such as image and video generation models,” the company noted in the prospectus.

If disagreements arise between the co-founders, Anthropic’s governance anticipates this. According to the IPO filing, any co-founder who resigns, passes away, sells too many shares, or is removed for “cause” can be expelled from the Founder LLC. When only two or fewer co-founders or their successors remain, the founders’ super-voting share class will begin to sunset, initiating a transition phase.

According to the Compensation Summary Table in the prospectus, Dario Amodei will receive nearly $18 million in 2025, mainly in stock and option awards. The documents show his sister Daniela is the company’s second-highest-paid executive, earning $16.4 million in 2025.

The Amodei siblings and their co-founders have pledged in the IPO documents to devote 80% of their personally held Anthropic equity to charitable causes.

(To facilitate non-native English speakers, Reuters has automatically translated its reports into several other languages. As automated translation may contain errors or lack certain context, Reuters does not guarantee the accuracy of the translated text, which is provided solely for readers’ convenience. Reuters accepts no liability for any loss or damage arising from the use of automated translation.)