We didn’t just hunt alpha; we rewired the game. But when an AI lab built on the promise of safe, decentralized intelligence pivots to a governance model that concentrates power in the hands of a few founders, it’s time to ask: Are we rewiring the game, or just building a new cage? This week, the crypto-native press lit up with a report from Crypto Briefing: Anthropic, the vaunted AI safety company behind Claude, is planning to issue super-voting shares to its CEO and co-founders ahead of its IPO. On the surface, it’s a run-of-the-mill Silicon Valley power move—Google did it, Meta did it, Snap did it. But in the context of a mission-driven AI lab that positions itself as the antidote to OpenAI’s profit-over-principles drift, this move is a philosophical earthquake. And because I’ve been in the trenches of both blockchain governance and AI ethics, I want to tear open the implications from a crypto-native perspective—because this isn’t just about Anthropic. It’s about the soul of decentralized technology itself.
Let me set the stage. I’ve spent the last seven years building crypto education platforms, auditing smart contracts, and watching the line between code and trust blur. In 2017, I was auditing early Solidity contracts for a project called EtherHouse, catching re-entrancy bugs that would have cost investors $200,000. That experience taught me that “code is law” is a beautiful lie—unless you have a governance structure that can enforce it. Fast forward to 2020, I was forking Uniswap to build a localized AMM in Jakarta, only to realize that the real bottleneck wasn’t code but community trust. And in 2022, after the Terra/Luna collapse, I wrote a 50-page dissection of trustless systems that relied on infinite growth, realizing that economic confidence is not the same as cryptographic trust. So when I see a company like Anthropic—which has built its entire narrative around “responsible AI” and “long-term safety”—reaching for the same centralization tools that have historically corrupted everything from governments to protocols, I can’t stay silent.
Context: The Golden Goose Meets the Iron Fist
Anthropic is not just any AI company. It’s the one that explicitly spun off from OpenAI because of concerns about profit-driven safety shortcuts. Its founders—Dario Amodei, Daniela Amodei, and others—have positioned the company as the “safe” alternative, attracting a $7.5 billion valuation in 2023 and a $4 billion investment from Amazon. They’ve raised more than $10 billion to date, and the IPO is the natural next step. But here’s the kicker: according to the report, they plan to issue super-voting shares that give the CEO and co-founders control over strategic decisions, even after going public. The exact voting ratio is unknown, but precedent suggests 10x or 20x voting power per share.
This is a classic “founder protection” mechanism, but in the AI world, it’s a double-edged sword. On one hand, it protects the company from short-term activist investors who might push for faster monetization of AI models, potentially cutting corners on safety. On the other hand, it concentrates power in a small group of people who are not subject to external checks—exactly the kind of centralization that blockchain was designed to fight. And Anthropic’s entire PR machine is built on the idea of “alignment” and “distributed safety.” The irony is thick enough to mine.
Core: A Crypto-Native Deconstruction of Super-Voting in AI
Let me break this down through the lens of the decentralized governance principles I’ve been teaching for years. In blockchain, we have a continuum of governance models: from pure on-chain DAOs with one-token-one-vote, to Ethereum’s off-chain rough consensus, to the centralized control of a foundation. The sweet spot is somewhere between anarchy and dictatorship. Anthropic’s super-voting shares are a hard pivot toward dictatorship—call it “benevolent dictatorship” if you want, but it’s still a dictatorship. And the crypto world has a long history of benevolent dictatorships that turned sour. Remember the DAO hack? The Ethereum Foundation’s decision to fork was a centralized act that saved the ecosystem but broke the social contract. Now multiply that by a trillion-dollar AI industry.
From my work auditing smart contracts, I’ve learned that the most dangerous exploits are not technical but governance-based. The re-entrancy bug I caught in 2017 was a technical flaw, but it was fixable. The real poison is when a protocol has a hidden backdoor—like a multi-sig that can override the community. Anthropic’s super-voting shares are a permanent backdoor. The founders can, in theory, overrule any board decision, any shareholder vote, or any safety committee. And while they might be virtuous today, the history of power is that it corrupts, and absolute power corrupts absolutely. The crypto community has seen this with projects like BitConnect, where the founder’s unchecked power led to a $2.4 billion fraud. Even well-intentioned founders like Vitalik Buterin have voluntarily reduced their influence to avoid centralization. Why? Because they understand that the long-term health of a decentralized system requires guardrails against power concentration.
Now, let’s talk numbers. The analysis report I’m working from indicates that the market already prices a governance discount for dual-class shares—typically 5% to 10%, sometimes higher. But for AI companies, where the regulatory landscape is shifting and public trust is fragile, that discount could be larger. I’ve seen this play out in the crypto world: when a project like Solana had a centralized validator set, the market punished it with lower valuations during the 2022 crash. The same will happen to Anthropic if investors realize that the founders’ control is a liability, not a feature. The IPO pricing will be a battle between the “mission protection” narrative and the “power risk” reality.
But here’s the deeper layer: the super-voting structure is not just about governance; it’s about the nature of AI safety. Anthropic claims to be building AI that is “safe, helpful, and honest.” But if the definition of “safe” is controlled by a few individuals, then the safety is as robust as the founders’ judgment. I’ve seen this fallacy in crypto security audits: the assumption that the auditor is always right. In reality, auditors are fallible, and the best security comes from multiple independent checks. Anthropic’s super-voting shares effectively eliminate the checks. The company can ignore dissenting voices from institutional investors, from the public, or even from its own employees. This is a recipe for what I call “algorithmic autocracy”—a system where the rules are set by a small group and enforced by code, but the code is not open to democratic revision.
From my experience in the Jakarta Web3 hub, I’ve seen how local communities react to centralized control. When we launched a DAO for a reforestation project, the first thing we did was design a multi-sig with stakeholders from different backgrounds. The result was slower decision-making, but higher trust. Anthropic is going the opposite direction—faster decisions, but lower trust. In a world where AI safety is a global concern, trust is the most valuable currency.
Contrarian: The Case for Benevolent Dictatorship (and Why It Fails)
Now, let me play devil’s advocate—because I’ve been in the trenches long enough to know that simplicity is a trap. There is a legitimate argument for super-voting shares in the context of AI safety. The reasoning goes like this: AI is a transformative technology that could lead to existential risks. Short-term profit motives could push companies to deploy unsafe AI. By concentrating power in the hands of founders who are committed to safety, you insulate the company from market pressure. In other words, the super-voting shares are a “mission lock.” This is similar to how some crypto projects have a “foundation” with veto power over protocol upgrades—like the Ethereum Foundation’s role in the early years. And it worked, for a while.
But here’s the problem: the crypto ecosystem learned that even benevolent foundations can become bottlenecks. The governance of Ethereum is now a messy, but ultimately more resilient, process of community consensus. The Bitcoin governance model is even more decentralized—no one individual can change the protocol. The result is that both networks have survived multiple crises, including the 2022 crash. Anthropic’s model, by contrast, puts all the eggs in the founders’ basket. What happens if the founders disagree with each other? The report doesn’t specify how the super-voting shares are distributed among the co-founders. If they are held jointly, a single disagreement could paralyze the company. If one founder holds a majority, that person becomes a de facto AI emperor.
I’ve seen this governance failure in crypto firsthand. In 2021, I was part of an NFT project that gave the founder absolute control over the treasury. The founder was a visionary, but eventually, he made a bad call—investing in a rug-pull project—and the community lost everything. The lesson is that no one is immune to bad judgment. The blockchain industry has moved toward decentralized governance not because it’s efficient, but because it’s antifragile. Anthropic’s move is a step backward, toward centralization that is efficient only until it breaks.
Moreover, the assumption that super-voting shares protect against short-termism is flawed. In practice, founders can still be influenced by the need to deliver quarterly results to the market. The super-voting shares just make it harder for outsiders to fire them. But the founders themselves may still be driven by the same profit motives—especially if they hold significant equity. The safety narrative could become a shield for poor decisions. I’ve seen this in crypto when projects claim to be “decentralized” but the founders control the keys. The market eventually sees through the hypocrisy.
Takeaway: A Fork in the Road for AI Governance
So where does this leave us? Anthropic’s super-voting shares are a test case for the entire AI industry. If the market accepts them, we will see other AI labs—OpenAI, xAI, Mistral—follow suit. The result will be an industry where a handful of individuals control the direction of the most powerful technology in human history. If the market rejects them, we might see a new governance model emerge—one that borrows from blockchain’s decentralized governance, with independent safety boards, community voting, and transparent decision-making.
As someone who has spent years teaching that “education is the new mining rig for the mind,” I believe the crypto community has a moral obligation to speak out against this centralization. We have the tools to design better governance—multi-sig, DAOs, quadratic voting. Anthropic could use these tools to create a truly decentralized AI safety governance model. Instead, they are choosing the old Silicon Valley playbook. It’s a missed opportunity.
From core dev trenches to community heartbeat, I’ve seen both the promise and the peril of concentrated power. When the market sleeps, the architects wake up. And right now, the architects of AI are making a choice that will define the next decade. Let’s hope they don’t build a cage that locks us all in.
We didn’t just hunt alpha; we rewired the game. Education is the new mining rig for the mind. When the market sleeps, the architects wake up. Art is the interface; blockchain is the canvas. These are the signatures I live by, and they remind me that the most important battles are not about technology, but about trust. Anthropic’s super-voting shares are a test of that trust. Let’s see if the market passes.