On-Chain Governance Is Dying, and Its Practitioners Know It

The most recent Uniswap governance vote — to allocate $48 million of protocol treasury to a delegate-administered grant pool — closed with 7.1% voter participation. That is not a failure. That is a normal vote, in the range of the protocol's trailing twelve-month average. The fact that we now consider 7.1% unremarkable is the point.

What we used to say about this

In 2020 and 2021, on-chain governance was the load-bearing argument for the entire DeFi edifice. Governance tokens were justified by their governance utility. The token was a ticket to the room; the room was where the protocol's future was decided. A protocol without governance was a protocol vulnerable to capture, rent extraction, or slow institutional drift.

I believed much of this. Many people who now write about governance's difficulties also believed it. We were wrong about the empirics and probably wrong about the theory.

Where decisions actually get made now

If you want to understand what is about to happen to a major DeFi protocol, on-chain votes are not where you look. You look at:

• The core contributor Discord, where the substantive discussion happens
• The delegate working groups, where the proposals are pre-negotiated
• The foundation or Labs entity, whose operational decisions do not go to a vote

The on-chain vote is a ratification of decisions that were made elsewhere. This is not necessarily bad. It might be how grown-up institutions function. But it is not the thing we described when we told the SEC, the public, and ourselves what governance tokens were for.

"Voting is downstream of everything. If the voters cared, they would show up. They don't, because the decisions that reach them are the decisions that everyone who cares has already agreed on." — Arthur Hayes, in a recent blog post

The delegate consolidation problem

Delegation was supposed to make governance scalable by allowing engaged stakeholders to vote on behalf of less-engaged token holders. In practice, it has produced a small cohort of professional delegates — venture firms, dedicated delegation services, and a handful of individuals — whose combined voting power routinely exceeds 60% on major protocols.

This is not a corruption of governance. It is the logical equilibrium of governance at scale. But it undermines the claim that on-chain governance is meaningfully different from the proxy voting systems used in traditional public companies, which delegate voting power to BlackRock and Vanguard in approximately the same concentration.

The case for honesty

The honest position, in 2026, is that on-chain governance for mature protocols looks a lot like corporate governance for mid-cap public companies. There are professional delegates. There is shareholder apathy. There is substantive decision-making that happens in working groups rather than plenary votes. None of this is dishonorable. It is the institutional form that large economic coordination systems naturally take.

What is dishonorable is continuing to pretend otherwise. The governance token pitch of 2020 was a fiction that we collectively agreed to believe. The fiction has consequences — it makes the SEC's Howey analysis much more difficult to rebut, it continues to produce token structures optimized for a participatory mode that does not exist, and it makes the actual institutions doing the actual governing less accountable because they operate in the shadow of a theater.

The protocols should say what they do. The governance token, where it remains, should be priced for what it is: a revenue share, a fee discount, a claim on residual treasury. The participatory claim should be retired. It has not served the industry well, and at this point, it is not even serving as a convenient lie.