On August 19, The Information broke the news: Anthropic is preparing a class of stock with enhanced voting rights for CEO Dario Amodei and other co-founders. This is the first time the AI company’s leadership will hold super voting rights. The move is designed to insulate the founding team from external shareholder pressure. Given the co-founders’ low proportional equity, this structural tweak is a clear signal: control, not capital, is the final frontier.
I’ve seen this pattern before. In 2017, I audited the Parity Wallet multisig contracts. The ownership transfer logic had a single point of failure—a centralized admin key that could be exploited. The fix was a decentralized multi-signature scheme. Anthropic’s playbook flips that script: instead of distributing control, they are consolidating it. This is not a bug; it’s a feature designed for a world where the founder’s vision must survive the quarterly earnings hammer.
Context
Anthropic, founded by former OpenAI employees, has raised billions from investors like Google, Salesforce, and Spark Capital. The company’s valuation hovers around $18 billion. Yet, the co-founders hold a minority stake. To prevent a hostile takeover or dilution of their strategic direction, they are introducing a dual-class share structure. This is a common practice in tech—Google, Facebook, and Snap have done it. But for a company that positions itself as a safety-first AI builder, the optics are tricky.
The article also mentions that Anthropic will retain its existing non-shareholder trustee system and use special class shares to elect a majority of board members. This means external shareholders—even those with deep pockets—will have limited influence on governance. The trustees, appointed by the founders, will control the board.
Core: The Mechanics of Control
Let’s dissect the engineering. From a structural perspective, super voting rights are a form of asymmetrical control. The class of stock (let’s call it Class B) will have, say, 10 votes per share, while investor Class A shares have 1 vote. The founders will hold Class B. The trustee system adds another layer: even if investors accumulate Class A shares, they cannot elect a majority of the board because the trustees—appointed by the founders—will fill the seats.
This is a centralized governance model. It mimics the “admin key” problem in DeFi. In crypto, we call this a “multisig with a single signer.” It’s efficient for decision-making but catastrophic for decentralization. The difference is that Anthropic is not a protocol; it’s a corporation. The goal is to protect the mission (building safe AI) from short-term profit motives. But the mechanism is the same as a DAO with a single whale controlling the vote.
Based on my experience with DeFi leverage traps, I know that control concentration creates a moral hazard. In 2020, I deployed $150k into a compound strategy. The monitoring dashboard I built tracked liquidation thresholds. When the market spiked, I manually adjusted. The lesson: when a single entity holds the keys, they can override the system’s rules. Anthropic’s founders can now override shareholder votes. That’s a feature, but it’s also a risk.
Contrarian: Why This Is Good for Crypto (and Bad for AI)
Here’s the contrarian angle: Anthropic’s move validates the core thesis of decentralized governance in crypto. The AI industry is retreating to centralized control because it works. The founding team has a vision—safe AGI—that requires long-term bets. Shareholders, especially in a bear market, want short-term returns. The super voting rights prevent the company from being dismantled by activist investors.
But crypto’s promise is the opposite: trustless, transparent, and distributed governance. We don’t need super voting rights because we have code, multisigs, and on-chain voting. The irony is that while Anthropic builds centralized AI, the crypto community builds decentralized AI agents. The governance models are diverging.
I saw this during the Terra/UST collapse. The protocol’s algorithmic stablecoin had a centralized oracle feeding the peg. When the oracle failed, the system collapsed. Anthropic’s super voting rights are the same: a single point of failure. If the founders make a bad decision, there is no check from the board or shareholders. The trustee system is a rubber stamp.
Takeaway
Anthropic’s governance structure is a case study in centralized control. For crypto investors, it’s a reminder that “decentralization” is not a feature; it’s a foundation. The market doesn’t owe you an exit, only a price. If you want control, you need to hold the super voting shares or write the code.
Trust is a variable I solve for, never assume. Anthropic’s founders are solving for trust by removing the variables. That works until the founder is replaced or the vision changes. The crypto community should watch this closely—it’s the most honest example of “not your keys, not your coins” applied to corporate governance.
Security is not a feature; it is the foundation. Anthropic’s super voting rights are a security measure for the founders, not the users. The underlying technology—AI—is powerful, but the governance is fragile. In a bear market, survival matters more than gains. Anthropic is ensuring its survival, but at the cost of the very openness that makes crypto valuable.
I trade the structure, not the story. The story here is “founder control for safety.” The structure is a centralized oligarchy. I’ll bet on the crypto protocols that distribute power, not the AI companies that concentrate it.
Final Thought
Anthropic’s move is a governance playbook for centralized AI. For the crypto world, it’s a warning. If you want to build systems that last, don’t copy the corporate structure. Build the code that cannot be overridden. The market will reward the structures that survive the next crash.
Speculation is gambling with a spreadsheet. Anthropic’s investors are gambling that the founders will make the right calls. I’d rather audit the code.
- Emma Garcia