The news hit the terminal at 09:14 Dublin time. Anthropic, the AI safety poster child, plans to issue super-voting shares to its CEO and co-founders ahead of IPO. The market yawned. But I froze.

Because I've seen this before.

Not in AI. In DeFi. In the summer of 2020, when a certain protocol's multi-sig holders voted themselves veto power over liquidity pools. The code didn't bleed — the liquidity stayed cold. But the governance rot was real.

Let me be clear: this isn't a governance innovation. It's a trap. And if you're looking at Anthropic's IPO with dollar signs in your eyes, you need to understand the mechanics of the trap before you step into it.

Context: The House of Cards Built on Hope

Anthropic is the darling of the AI safety movement. Founded by former OpenAI employees, it has raised billions from strategic investors like Amazon and Google. Its Claude models are competitive. Its narrative is pristine: build AI that is safe, aligned, and accountable.

Now it wants to go public. And it wants to ensure that the founders — Dario Amodei, Daniela Amodei, and a few others — retain control even after selling shares to the public. The mechanism: super-voting shares, likely with 10x or 20x voting power per share.

This is not new. Google, Facebook, Snap all did it. But for an AI company that preaches transparency and accountability, it's a glaring contradiction. The code bleeds, but the liquidity stays cold.

Core: The Order Flow Analysis of Control

Let's break down the mechanics. In a typical dual-class structure, Class A shares trade publicly with one vote each. Class B shares, held by founders, carry multiple votes — often 10 or 20. The exact ratio for Anthropic hasn't been disclosed, but based on the pattern, I'd bet on 10:1.

Why does this matter? Because it means that even if the founders sell 90% of their economic stake, they can still control 100% of the voting power. They can appoint the board. They can approve mergers. They can override shareholder votes on executive compensation. They can, in theory, ignore the market's demands for short-term profits.

That sounds like a feature, not a bug, for an AI safety company. But here's the problem: power without accountability is a bug. And in the world of DeFi, I've seen how that bug exploits.

In 2022, I watched a DAO with a three-person multi-sig drain its treasury to fund a fork that the community didn't want. The multi-sig holders argued they were protecting the 'vision'. The token went to zero. The vision didn't survive the liquidity crisis.

Anthropic's super-voting structure is a slower, more elegant version of that same bug. The founders have absolute control over the AI safety narrative. They decide what is safe. They decide when to launch. They decide when to shut down. No external audit, no independent board, no shareholder vote can override them.

This is not alignment. This is centralization.

Contrarian: The Retail vs. Smart Money Blind Spot

Here's the contrarian angle: the market is pricing this as a positive. 'Founder-led companies perform better,' the narrative goes. 'Google and Facebook thrived with dual-class structures.' That's true — for the first decade. But look at the long tail. Facebook's governance battles with the FTC. Google's antitrust lawsuits. Snap's declining stock price despite founder control.

In each case, the super-voting structure protected the founders from being ousted, but it also protected them from being corrected. When the market wanted a pivot, the founders could ignore it. When the board wanted accountability, the founders could veto it.

In AI, the stakes are higher. If Anthropic's founders decide to rush a model to market despite safety concerns, no one can stop them. If they decide to slow down despite competitive pressure, no one can accelerate them. The risk is not governance — it's monopoly on judgment.

And here's the blind spot: institutional investors like BlackRock and Vanguard, who are already under pressure from ESG mandates, may discount Anthropic's IPO. Governance risk is real. Studies show dual-class companies trade at a 5-10% discount. For a company already valued at hundreds of billions, that's billions in lost capital.

But the smart money is already in. Amazon and Google are strategic investors. They don't care about voting rights — they care about access to AI talent and models. The retail investor, however, will be the exit liquidity. They'll buy the IPO at a premium, only to find that their shares have no voice.

Incentives align only when the risk is priced in. Here, the risk is not priced in. It's hidden behind a 'safety' narrative.

Takeaway: The Liquidity Stays Cold

So what's the play? If you're a long-term investor, skip the IPO. Wait for the first governance crisis — a founder launches a controversial model, the stock drops, and the discount becomes real. Then buy.

If you're a trader, exploit the volatility. The IPO pop will be driven by retail FOMO. Short the pops. Buy the dips. The code will bleed, but the liquidity will stay cold.

And if you're building in AI-crypto convergence, take note. This governance model is the antithesis of what DeFi stands for. Decentralization is not just about code — it's about power. Anthropic's super-voting shares are a reminder that the most dangerous centralization is the one that hides behind a mission.

Volatility is the only constant truth. But governance is the structure that determines whether volatility is a feature or a bug.

Audit trails don't lie. But the people who control the audit trails can.

Postscript: The Experience That Shapes This View

I've been on both sides of this table. In 2017, I reverse-engineered a Solidity contract with a reentrancy bug. The code was elegant. The exploit was inevitable. The governance was the last line of defense — and it failed because the multi-sig holders didn't want to upgrade.

In 2020, I pulled my liquidity from Uniswap V2 pools minutes before a flash loan attack. I didn't wait for a vote. I acted. That's the difference between a founder with absolute power and a protocol with distributed governance. One can save you. The other can trap you.

In 2022, I shorted Terra because I saw the governance failure before the code failure. The founders had super-voting power over the protocol. They used it to print LUNA. The rest is history.

Anthropic is not Terra. But the governance pattern is the same. Power without accountability. Vision without check. Safety without transparency.

I'll watch the IPO. I'll trade the volatility. But I won't trust the narrative. Because the code may bleed, but the liquidity stays cold.

This article is based on public information and my own trading experience. It is not financial advice. The market is a mirror, not a floor.