Situation report: Anthropic wants its seven co-founders to keep control of the company after it goes public. According to The Information, the Claude maker is looking at a Palantir-style voting structure that would give the founding group outsized say over company decisions. That could hold even after public shareholders own a big share of the equity.
The report comes as Anthropic moves toward an IPO. It points to the question every frontier AI lab has to answer eventually: once public money comes in, who decides what the company builds?
🎯 What’s on the table
The Information’s headline says it plainly: voting control for seven co-founders, modeled on Palantir. Here’s what that model looks like in practice.
- Palantir’s template. Palantir’s founders hold a special class of shares that lets them control close to 50% of the company’s voting power. That holds no matter how much of the company they actually own, so their vote stays steady even as their economic stake shrinks.
- The seven. Anthropic’s co-founders include CEO Dario Amodei, President Daniela Amodei, Jared Kaplan, Sam McCandlish, Tom Brown, Chris Olah and Jack Clark. Most of them left OpenAI together in 2021.
- The group, not one boss. This isn’t a single-founder setup like Meta’s, where Mark Zuckerberg holds the controlling vote himself. The plan reportedly gives control to the founding group as a whole.
The short report doesn’t lay out every detail, such as exact vote thresholds or sunset clauses. Those details will matter a lot once a filing becomes public.
🧭 Why it matters
What stands out here is how the plan fits with Anthropic’s existing governance. The company is a public benefit corporation, which means it has a legal duty to weigh its mission alongside profit. It also has a Long-Term Benefit Trust, an independent body that can elect some of its board members.
A founder voting block would add a third layer of protection. The logic is simple: if you believe you’re building one of the most consequential technologies around, you don’t want quarterly earnings pressure or an activist fund steering safety decisions.
The counterargument is just as simple. Dual-class shares tend to weaken accountability. Public investors would be putting billions into a company where they have little real power to change leadership or strategy.
📊 How we got here
In the AI race, governance has become a real business risk. OpenAI’s 2023 board crisis showed how fast a mission-driven structure can clash with commercial reality, and OpenAI has spent a lot of time since then reworking its own structure.
Anthropic has taken the opposite approach. Rather than loosening its safety-focused governance to attract capital, it looks set to lock that governance in before the public markets get a vote.
Founder control is also normal among big tech IPOs. Google, Meta, Snap and Palantir all went public with some version of it. Index providers and governance advocates have pushed back over the years, but investors keep buying the shares anyway.
⚠️ What to watch
- Investor reaction. Big institutional investors and proxy advisers often criticize dual-class structures. Watch whether they push for sunset provisions that end founder control after a set number of years.
- Index eligibility. Some index providers have limited or set conditions for companies with unequal voting rights. That affects how much passive money flows into the stock.
- Trust vs. founders. It’s unclear how founder voting power would interact with the Long-Term Benefit Trust. If the two ever disagree, the answer to who wins will tell you a lot.
- Rival moves. OpenAI and other labs heading toward public markets will face the same question. Anthropic’s structure could become the template.
🔍 Assessment
For practitioners and enterprise buyers, the practical takeaway is stability. If the plan goes through, the people who set Anthropic’s current direction on safety, model releases and responsible scaling will likely stay in charge for years. That makes the company’s roadmap easier to predict, for better or worse.
For investors, it’s a familiar trade: you get exposure to one of the fastest-growing AI companies, but you don’t get a real say in how it’s run.
More detail should come out as Anthropic gets closer to a filing. For now, The Information has the full report.