On-chain governance was never a democracy problem. It was an accounting one.
That is the line the 2020s missed, and the one that, from 2036, looks obvious. The pitch for blockchain governance was political: replace the closed boardroom with the open ledger, give every token holder a vote, let code execute the collective will. What the pitch never reconciled was the unit of account. The vote was denominated in the asset — and any system that measures voice in money will, given enough liquidity, return the answer money already had.
It did. Quietly, transparently, and on schedule.
What everyone predicted
The conventional wisdom of the early 2020s held that transparency was the cure for capture. Corporate governance failed minority shareholders because it happened in private — proxy fights behind glass, fiduciary breaches you learned about too late. Put it all on a public ledger, the argument ran, and coordination among insiders becomes visible, therefore checkable, therefore deterred. Sunlight as disinfectant, ported to a chain.
The prediction was half right, which is the most dangerous kind. Sunlight arrived. Deterrence did not. By the time anyone had a decade of records to study, the data told one story across nearly every protocol that mattered. A 2022 Chainalysis analysis of ten major decentralized autonomous organizations found that, on average, fewer than 1% of holders controlled 90% of the voting power.1 Academic work landed in the same place from another angle: in a typical proposal, the top decile of voters commanded 76.2% of realized voting power.2 Participation, meanwhile, ran in the low single digits — frequently under 5% of eligible holders.3
These were not edge cases. They were the median.
Force one: the unbundling
The first force was structural, and Vitalik Buterin named it in 2021, before most of the capital had arrived. A governance token, he wrote, bundles two things that are trivially easy to pull apart: an economic stake in a protocol, and the right to steer it.4 Lending markets unbundle them for a fee. Borrow the token, vote with it, return it — and you have wielded governance power while holding none of the downside. Buterin's conclusion was blunt: "the rich tend to dominate the decision-making process," and small holders have "very little incentive to participate and a fairly high incentive to accept bribes to back bad decisions."4
This is the enhancement and the obsolescence at once. On-chain voting genuinely sped up and widened collective decision-making — a proposal could move from draft to executed code in days, visible to anyone. And it swept aside the apparatus that used to slow that down: the registrar, the proxy statement, the annual meeting, the human board. What it did not sweep aside was the oldest rule in the room. Wealth still set the agenda. The ledger just made it faster.
Force two: transparency as a coordination tool
The second force is the one the prediction got exactly backwards. Transparency did not deter coordination among large holders. It equipped it.
The cleanest illustration is real and predates 2036 by more than a decade. In July 2024, a small delegate group nicknamed the "Golden Boys," led by a whale known as Humpy, accumulated COMP voting power and pushed through a proposal to route roughly $24 million — about 5% of the Compound DAO treasury — into a yield product the group itself controlled.5 It took three tries. The winning vote landed over a weekend, when turnout was thin, after two earlier attempts failed.6 Nothing about it was hidden. Every delegation and ballot sat on a public chain for anyone to read. Visibility was not the obstacle to capture. It was the operating manual.
The system's defenders were not wrong about the mechanics — only about what to call it. Coordinated voting by professional capital, they argued, was efficient stewardship, not collusion. And on a public ledger, "collusion" is an awkward word for something everyone can watch in real time. That was the trap. The transparency built to make manipulation detectable made coordination frictionless, then made it respectable. The 2024 episode ended in an awkward truce; the next coordinated bloc, and the dozens after it, simply learned that the ledger was a feature, not a risk.
The reversal
Here is where the tool turns into its opposite.
On-chain governance retrieved an ideal its advocates thought they were resurrecting: shareholder democracy, one stake one vote, participation reduced to a click. But the retrieval kept going — past the modern corporation, back to the plutocratic assembly, where the weight of a voice was openly indexed to wealth. The Roman comitia centuriata at least said so out loud. So, it turned out, did the chain.
That is the reversal, and the whole story: a technology sold as the decentralization of power produced the most efficient concentration of it on record. Not the largest — empires have done worse. The most efficient. Corporate governance, for all its failures, ran on friction: disclosure rules, fiduciary duties, a regulator who could be embarrassed into acting. The chain removed the friction and kept the concentration — oligarchy with lower transaction costs, faster settlement, and a complete audit trail. Plutocracy as a service, optimized for uptime.
And the legal vacuum was load-bearing. Wyoming became the first U.S. state to recognize DAOs as a form of LLC in 2021, with Tennessee and Vermont following.7 But most large protocols never incorporated; they ran as unincorporated associations whose members, by some readings, carried unlimited personal liability — which is exactly why Compound's largest delegate, the venture fund a16z, abstained from the contested 2024 vote it could have swung.6 The actors with the most power had the strongest reason not to be seen using it. Europe's Markets in Crypto-Assets regime, fully in force from December 2024, largely declined to reach genuinely decentralized governance, deferring the question to a later report.8 The result: organizations steering tens of billions with less binding oversight than a chartered bank — by design, and on the record.
What the reformers ran into
The fixes were known early, and good. Quadratic voting weights a voter by the square root of the resources committed, blunting whale dominance.4 Reputation systems and proof-of-personhood schemes try to denominate voice in something other than money. Delegation lets small holders pool their weight. By the 2030s, several protocols had adopted versions of all three.
But every one shared a defect the math could not route around: adopting it required a governance vote held under the old rules — a vote the whales controlled. Asking a concentrated electorate to ratify its own dilution is not a technical problem. It is a turkey-voting-for-Thanksgiving problem, and turkeys are consistent. The reforms that shipped were the ones large holders could tolerate. The ones that would have actually moved power stayed in the research forums, elegant and unpassed.
The lesson
The durable principle is not that blockchain governance failed. It is that it succeeded — at exactly what its denominator rewarded. A system that prices voice in capital will, at scale, return capital's verdict, faster and more honestly than the system it replaced. That honesty is the only genuine improvement on offer, and it is a real one: the chain stops pretending wealth and voice are separable.
Which leaves the question the 2020s thought they had answered. Democratize the interface and you change nothing; the interface was never where power lived. The only reform that matters is the one that changes the unit of account — and the unit of account is the one thing the holders of the unit will never vote to change.
Measure voice in money, and you will get the government money would have bought anyway. The ledger just hands you the receipt.
A note on this piece: This is speculative journalism written from a vantage of 2036. The 2036 framing, the figure captions, and the forward projections are informed extrapolation, not reporting — the future described here has not happened. Every claim about the actual past and present, including the Chainalysis and academic concentration findings, Vitalik Buterin's 2021 analysis, the July 2024 Compound DAO governance episode, the Wyoming/Tennessee/Vermont DAO statutes, and the MiCA regime, is real and sourced below. The boundary between the two is meant to stay visible.