Each holds roughly 2% of the shares, yet demands 50.1% of the voting rights: The IPO calculation of Anthropic's founders

On the cusp of its upcoming trillion-dollar-level debut IPO, Anthropic is brewing a highly controversial yet heavily protective governance structure adjustment: let the seven co-founders retain control of the company after going public.

Citing people familiar with the matter, The Information reported that Anthropic is seeking shareholder approval for a brand-new corporate structure that creates a special class of shares for CEO Dario Amodei and six other co-founders, granting them collective voting control over most corporate matters.

Under the proposed plan, as long as at least three of the seven founders continue to hold the stipulated number of shares, they will collectively own 50.1% of the voting rights.

This means that even if Anthropic continues to raise funds and issue new shares in the future, and the founding team's economic equity stake is further diluted, they will still be able to steer the company's major decision-making direction.

And this move by Anthropic draws reference from the "Founder Voting Trust" model of Palantir, the big data analytics giant.

The Balancing Act Between Extremely Low Equity Stake and Absolute Control

In fact, over the past decade or more, it has not been uncommon for founders of tech companies to lock in control through a "dual-class share structure" at the time of IPO. Typical representatives include Meta's Mark Zuckerberg and SpaceX's Elon Musk, who have both used super voting rights to allow founders to retain control of the company's strategic direction after listing.

However, the problems Anthropic faces are more unique.

Founded in 2021 by a group of former OpenAI employees, Anthropic has put AI safety and long-termism at the core since its inception, adopting the Public Benefit Corporation structure and setting up the Long-Term Benefit Trust, which consists of non-shareholder members including former Federal Reserve Chair Ben Bernanke and other figures, focusing mainly on AI safety and social impact issues. The purpose is to consider the potential long-term social impact of AI technology in addition to commercial interests.

But once it goes public, it will inevitably bring new pressure. Public market investors usually focus on revenue growth rate, commercialization efficiency, profit level, stock price performance and so on. To maintain its technological leadership, Anthropic needs to continuously invest in large-scale model training, computing power infrastructure, AI safety research... There is a natural "tension" between the two sides.

What is more disparate is the inversion of equity ratio and voting rights. Public information shows that Dario Amodei, the chief executive officer, only holds about 2% of Anthropic's shares, and other co-founders (including his sister, company president Daniela Amodei) have roughly similar shareholding ratios.

For a trillion-dollar AI giant about to go public, such a low shareholding ratio of founders is extremely rare. In addition, the founding team has previously publicly promised to donate 80% of their own wealth. With further equity dilution, external public market shareholders and Wall Street capital are very likely to exert huge profit pressure on the management after the listing.

And this will lead to a question: Can the founding team still decide the long-term path of the company?

This new special share structure cleverly separates "economic benefits" from "control": the newly created special class of shares does not come with additional economic rights such as dividend distributions, but provides the founding team with a solid moat, ensuring that the core team firmly controls the company's strategic direction while donating wealth and facing equity dilution.

Unique Governance Puzzle: The Deciding Vote of the Board of Directors, Trust and Employees

Of course, although Anthropic will grant the founding team 50.1% of the voting rights, this does not mean that they will have full control of the company.

According to The Information's report, on the most crucial board seat election, Anthropic still retains its unique supervision and balancing mechanism:

First, the Long-Term Benefit Trust (LTBT) maintains checks and balances: Anthropic's existing board of directors has a total of seven seats (one is currently vacant). According to the existing articles of association, the power to appoint most directors does not belong to the shareholders, but to the "Long-Term Benefit Trust".

Under the new structure:

The trust will retain the power to appoint the majority of directors;

The number of director nominees from the founders will be slightly expanded from the original 2 to 3.

It can be seen that Anthropic is trying to strike a balance between two forces: on one side is the founding team's long-term vision, and on the other side is the independent governance mechanism and supervision from external capital.

Second, beyond the founders, people familiar with the matter disclosed that Anthropic plans to grant employees a special class of shares, which will play a tie-breaking, decisive role to break deadlocks when there is a voting stalemate or tie on certain key corporate matters.

Why Palantir?

As mentioned earlier, this move by Anthropic is not original, but a reference to Palantir's founder control model.

When Palantir went public, it designed a special share structure: three founders Peter Thiel, Alex Karp and Stephen Cohen held special class shares through the Founder Voting Trust. This type of share allows founders to obtain nearly half of the voting rights.

According to Palantir's public documents, its Class F shares are designed to: when the minimum shareholding requirement is met, allow the founding team to own up to 49.999999% of the voting rights in total.

Anthropic's plan is similar to this:

Ordinary investors get economic benefits;

Founders get extra voting rights;

Control is separated from the economic equity stake.

The difference is that Palantir has a three-founder control structure, while Anthropic is trying to form a collective control mechanism for seven co-founders.

This design is relatively rare among large technology companies.

However, Anthropic's attempt actually reflects the new problems the AI industry is facing after entering the capital market: How should an AI company that needs more than ten years of investment and may affect the future society be governed?

For traditional technology companies, the logic is often: entrepreneurship — growth — IPO — shareholder return. But AI companies face a different cycle: foundational model R&D — computing power investment — safety verification — long-term commercialization.

This cycle may be far longer than the quarterly assessment of the capital market. Therefore, Anthropic is trying to establish a new listing model: Capital can enter the company, but cannot easily change the company's long-term direction...

The latest news shows that this high-profile IPO was originally expected by many investors to take place as early as this month, but now it is more likely to be postponed to the end of October or November.

What about you? What do you think of Anthropic's move? Feel free to leave a message and exchange ideas in the comment section!

Reference links:

https://www.theinformation.com/articles/anthropic-seeks-palantir-style-voting-control-seven-co-founders-ahead-ipo?rc=jn0pp4

https://www.investing.com/news/stock-market-news/anthropic-seeks-palantirstyle-voting-control-for-seven-cofounders-ahead-of-ipo-the-information-reports-4916355?utm_source=chatgpt.com

This article is from the WeChat Official Account "Synced" (ID: almosthuman2014), the author focuses on AI-related fields, and this content is published with authorization from 36Kr.