When Governance Becomes a Caste System: Anthropic's Super Voting Rights and the Crypto Lesson in Centralized Power

The Information broke a story on August 19 that should send a chill down the spine of anyone who believes in distributed authority. Anthropic, the AI company founded by former OpenAI researchers, is preparing a class of stock with enhanced voting rights for CEO Dario Amodei and other co-founders. This is the first time the leadership will have super voting rights, a move designed to help the founding team resist pressure from external shareholders. The co-founders hold a relatively low percentage of shares, so this arrangement is a transparent attempt to maintain control. Anthropic also plans to retain its existing non-shareholder trustee system and use special class shares to elect a majority of board members, further reducing the influence of external shareholders on corporate governance. Truth is immutable, unlike the price action. In the crypto world, we spend our days obsessing over consensus mechanisms, validator sets, and the risk of a 51% attack. Yet here, in plain sight, an AI company is architecting a 51% attack on its own governance. The parallels are not just academic; they are a mirror held up to the very tensions that define our industry. If we cannot learn from Anthropic's move, we are doomed to repeat the same mistakes in our own protocols. I have spent the last seven years auditing smart contracts, building DAO governance frameworks, and watching the industry oscillate between idealistic decentralization and pragmatic centralization. My 2017 experience auditing Tezos mainnet code taught me that code is law only if it compiles, but governance is the only law that matters when the code fails. The Anthropic story is not about AI; it is about the eternal struggle between founders and investors, between vision and capital, between the many and the few. And it is a story that every crypto project should study with the same rigor we apply to a reentrancy vulnerability. Let us dissect the structure. Anthropic is creating a separate class of shares that give the founders super voting rights. This is not novel; Google, Facebook, and many tech companies have dual-class structures. But the context matters. Anthropic is a company that has raised billions from investors like Google, Spark Capital, and others. The co-founders have a low equity stake, meaning that without these special shares, they could be outvoted by investors who may prioritize short-term returns over the company's stated mission of building safe AI. The trustee system, which appoints independent directors, further insulates the board from shareholder influence. The message is clear: founders know best, and investors should trust them. In crypto, we call this a "founder node" with veto power. It is the antithesis of what we claim to build. But let us be honest: how many crypto projects truly operate without a central authority? The Ethereum Foundation still holds enormous sway over the protocol's direction. Bitcoin's core developers are not elected by the users. The difference is that in crypto, we at least have a mechanism for fork. If the community disagrees, they can split the chain. In Anthropic's case, the shareholders cannot fork the company; they can only sell their shares and walk away. That is a form of exit, but not a form of voice. But here is the deeper issue. Anthropic's move is a response to a real problem: the tension between long-term mission and short-term capital. The founders of Anthropic believe that AI safety requires a long-term commitment that may not align with quarterly earnings. Investors, by nature, want returns. The super voting rights are a tool to protect the mission. This is exactly the same argument used by Vitalik Buterin when he defended the Ethereum Foundation's role in guiding the protocol. It is the same argument used by Bitcoin maximalists who resist any change to the base layer. The question is not whether centralized control is necessary; it is whether it is legitimate. I have seen this play out in DAOs. In 2020, during the DeFi Summer, I helped design a governance system for a lending protocol. The founders wanted to retain a veto power over critical parameter changes. The community revolted. We ended up with a multi-sig with time-lock, but the founders still held the keys. Within six months, the community forked the protocol, and the original project lost 80% of its TVL. The lesson was clear: users will tolerate centralization only as long as they see it as a necessary evil. The moment they perceive it as a betrayal, they will leave. Anthropic's situation is different because the users are not token holders; they are customers. But the same principle applies. The company's customers—developers, enterprises, individuals using Claude—are not voting on governance. They are consuming a service. The investors are the ones with the least power. This creates a bizarre hierarchy: the users have no power, the investors have some power, and the founders have all the power. It is a feudal system, not a democracy. Now, let me bring in my own technical experience. In 2022, after the Terra collapse, I retreated to a cabin in Virginia to write "The Soul of Sovereignty." During that solitude, I studied the governance structures of the top 50 blockchain projects. I found that only 12% had truly decentralized governance where no single entity could unilaterally change the protocol. The rest had some form of founder control, often masked by token-weighted voting that was captured by early investors. The crypto industry is not immune to the Anthropic problem; we just have better PR. But here is the contrarian angle. Perhaps super voting rights are not the enemy. Perhaps they are the only way to ensure that long-term vision survives the short-termism of capital markets. Satoshi Nakamoto had absolute control over Bitcoin's early code, and that was good for the network. The Ethereum Foundation's ability to push through the transition to Proof of Stake was essential. The risk is not that founders have power; it is that they abuse it. The question is: what checks and balances exist? In Anthropic's case, the trustee system is supposed to provide that balance. But trustees are appointed by the founders. That is not a check; it is a rubber stamp. The real check is the market: if the founders make bad decisions, the company's value will decline, and investors will lose money. But that is a slow feedback loop. In crypto, the feedback loop is faster because users can fork. In traditional companies, the feedback loop is slower because you cannot easily replicate the network effects. I have seen this dynamic in my own work. In 2024, when I launched the "Human-Centric AI" initiative, I collaborated with ethicists to draft the "Decentralized Trust Protocol." The goal was to create a set of guidelines for AI agents operating on-chain. One of the key principles was that no single entity—not even the founders—should have unilateral control over the protocol's governance. We implemented a multi-sig with time-lock and a community veto mechanism. It was messy, slow, and sometimes inefficient. But it was trustworthy. Anthropic's approach is the opposite. They are doubling down on efficiency over trust. They are saying, "Trust us, we are the experts." But in a world where AI is becoming more powerful, trust is not enough. We need verifiability. We need transparency. We need the ability to audit the governance decisions. Super voting rights are the antithesis of verifiability. So what is the takeaway for the crypto community? First, we must stop pretending that our projects are decentralized when they are not. The Anthropic story is a mirror. If we are building a protocol with a founder-controlled multi-sig, we are not better than Anthropic. We are just using different jargon. Second, we must design governance systems that are resilient to founder capture. This means using time-locks, multi-sigs, community veto mechanisms, and, most importantly, the ability to fork. If a project cannot be forked, it is not truly decentralized. The bear market builds the foundation. Right now, in this quiet period, we have the opportunity to rethink our governance models. Anthropic is making a bold move, but it is a move that reveals the fragility of trust-based systems. The long-term vision is not about protecting founders; it is about protecting the users. The community is the ultimate validator. If we forget that, we will end up with a crypto industry that looks exactly like traditional finance, just with tokens instead of stocks. Truth is immutable, unlike the price action. The price of Anthropic's governance choice will not be known for years. But the lesson is already written. We must build systems that are robust not because of the people in power, but because of the architecture that limits that power. That is the only path to true sovereignty.