The company building Claude just set up a power structure that would make Zuckerberg blush.
The Summary
- Anthropic is heading toward an IPO this year with a $65B revenue run rate and potential $2T valuation, but founders will retain 50% voting control through dual-class shares
- Amazon stands to gain more than Alphabet from its larger stake, reshaping AI sector power dynamics
- The voting structure mirrors Meta and Google's playbooks, but may deter investor confidence and pressure valuation despite explosive revenue growth
- This is the first major AI foundation model company to go public with explicit founder control mechanisms baked into the governance structure
The Signal
Anthropic's revenue run rate hit $65B, putting it in rarified air for a company that was barely a household name two years ago. That's not just good. That's OpenAI-tier growth without OpenAI's drama. The market demand for Claude and the enterprise AI tools built on top of it has turned Anthropic from "the safety-focused AI company" into a revenue machine that can credibly eye a $2T public market debut.
But here's the tension. The founders are locking in 50% voting control through a dual-class share structure. Dario and Daniela Amodei, the siblings who left OpenAI to build something more aligned with their safety principles, are now building a governance moat that makes those principles functionally untouchable by public shareholders. You can own a piece of Anthropic. You cannot steer it.
"Founder control mechanisms are the new normal in AI, but Anthropic's 50% threshold is the highest yet for a foundation model company going public."
This isn't novel in tech. Meta, Alphabet, and a parade of other platform companies have used dual-class structures to keep founders in control post-IPO. But it's new for AI at this scale, and it may chill institutional investor appetite. Pension funds and index trackers hate governance structures they can't influence. The risk: Anthropic prices its IPO, the pop is smaller than expected, and the narrative shifts from "AI goldmine" to "founders who don't trust the market."
Meanwhile, Amazon is positioned to extract more value than Alphabet from this IPO because it holds a larger stake. Amazon's cloud infrastructure deal with Anthropic wasn't just a partnership. It was a bet that Anthropic would become the AI layer AWS customers choose when they don't want to hand OpenAI more leverage. If Anthropic hits a $2T valuation, Amazon's stake could be worth more than the entire market cap of some Fortune 500 companies.
Key stakes for the AI economy:
- First test of whether public markets will tolerate founder control in foundation model companies
- Amazon vs. Alphabet proxy war for AI infrastructure dominance playing out in Anthropic's cap table
- If the IPO succeeds despite governance concerns, expect every AI startup to copy the structure
The Implication
Watch how institutional investors respond to the S-1 filing. If BlackRock and Vanguard push back on the voting structure, Anthropic will either have to adjust or accept a lower valuation multiple. That outcome sets the template for every other AI company watching from the wings. OpenAI, xAI, Cohere, they're all taking notes.
For builders in the agent economy, this matters because Anthropic's model availability and pricing are now tied to public market expectations. If the IPO underperforms, cost pressure flows downstream to developers betting on Claude as their inference backbone. If it crushes, Anthropic has the capital to undercut OpenAI on price and expand model capability faster than anyone expected.