Super-voting shares are the crypto equivalent of a private key without a backup: absolute control, single point of failure. Anthropic, the AI startup behind Claude, plans to issue such shares to its CEO ahead of a potential 2025 IPO. The market cheers. I see a structural flaw hiding in plain sight.
Context: The Governance Hype Cycle Anthropic, valued at over $60 billion, is preparing for a public listing that could redefine AI investment. Super-voting shares—typically 10 votes per share vs. 1 for ordinary stock—allow founders to retain control despite diluted economic ownership. This is not new. Google, Facebook, and Snap did it. But in an era where “decentralized” is the crypto mantra, Anthropic’s move is a blunt reminder: trust is a variable you must solve, not a feature you can assume.
The crypto community often romanticizes DAOs as the future of governance. Yet here, a centralized AI company adopts the most centralized capital structure possible. The irony is lost on those who see IPO as validation. From my audits of DeFi protocols, I’ve learned that control concentration is the first variable to stress-test. Centralization hides in plain sight metadata.
Core: Systematic Teardown of Super-Voting Shares Let’s quantify the risk. Assume a standard dual-class structure: Class A (public) 1 vote, Class B (founder) 10 votes. If the CEO holds 10% of total equity but 100% of Class B shares, they control >50% of voting power with a minority economic stake. The mathematical inevitability: misaligned incentives. The CEO can prioritize long-term vision over quarterly profits—or personal projects over shareholder returns. The mechanism is identical to a rug pull: the controller can extract value without consent.
In crypto, we call this “admin key risk.” Every DeFi audit I’ve conducted flags single-signer ownership. When a protocol’s upgrade key is in one wallet, the protocol is a hot wallet waiting to be drained. Anthropic’s super-voting shares are that key—but for a $60 billion entity. The difference is regulatory approval. The SEC allows it. The market accepts it. Logic does not bleed; only code fails.
Consider the Terra collapse. Do Kwon’s control over UST’s mechanism was not absolute—but his influence was. The peg broke because one person’s confidence was the only collateral. Anthropic’s CEO, Dario Amodei, is not Do Kwon. But the structural risk is identical: a single point of failure in governance. Super-voting shares create a “too big to fail” narrative that masks fragility. Liquidity is a mirror reflecting greed. If the CEO makes a misstep, selling pressure on Class A shares will be immediate. The voting structure will not save the stock price.
From a quantitative perspective, I modeled the impact of a 20% CEO misstep (e.g., a failed product launch or regulatory fine). With super-voting control, the CEO cannot be removed by shareholders. The only recourse is a hostile takeover, which requires buying out the super-voting shares—a premium that may never materialize. The result: locked-in value destruction. Precision cuts through the noise of hype.
Contrarian: What the Bulls Got Right Super-voting shares do have a defensible logic. They insulate the CEO from short-termist activist investors. In AI, where R&D cycles are long and breakthroughs are unpredictable, long-term vision is critical. Anthropic’s focus on AI safety could benefit from governance that prioritizes ethics over quarterly earnings. The bull case: the CEO is a visionary, and the market trusts that vision. Historically, concentrated founder control has driven innovation at Alphabet, Meta, and Tesla.
Moreover, the IPO market rewards familiarity. Investors see dual-class structures as a signal of founder commitment. They are willing to accept lower governance rights in exchange for exposure to high-growth narratives. Anthropic’s IPO may be oversubscribed precisely because of this structure. Decentralization is a promise, not a feature. The market is voting for the promise.
But here is the blind spot: the same argument was used for FTX. Sam Bankman-Fried had super-voting control through a share class that gave him 100% of board seats. The result was a $8 billion black hole. The difference is regulatory oversight? FTX was regulated. The difference is transparency? FTX had audited statements. The difference is culture? Culture is a code that can be patched. Trust is a variable you must solve.
Takeaway: The Accountability Call Anthropic’s super-voting shares are a bet on the CEO’s character. But character is not code. Code is deterministic. Character is probabilistic. In a bear market—or a recession—the same governance structure that enables long-term bets can also enable long-term losses. The question every investor should ask: Is the CEO’s vision worth the single point of failure? Silence is the sound of exploited flaws.
I will not invest in Anthropic’s IPO. Not because I doubt the technology, but because I cannot audit the CEO’s intent. Volatility exposes the architecture of fear. Until the market demands decentralized governance for centralized entities, super-voting shares will remain a permissioned ledger with a single validator. And we all know how that story ends.