The Ethereum Foundation has never issued a press release about losing power. That’s the point.
In June 2024, a routine core developer call (ACDE #187) revealed something unusual: a proposed EIP to adjust the blob gas pricing mechanism was stalled not by Vitalik’s team, but by a coalition of L2 representatives and large staking pools who demanded a month-long simulation period before any vote. The Foundation’s own research team had already validated the change. It didn’t matter. The decision chain had shifted.
This is the quietest structural transition in crypto history—Ethereum migrating from a de facto foundation-led monarchy to a multi-node governance federation. No fork. No blog post. Just a slow, creeping redistribution of veto power.
Context: The Architecture We Were Sold
For years, Ethereum governance was simple: Vitalik and a handful of core researchers wrote EIPs, the Foundation shepherded them through client releases, and stakers validated the consensus. The Foundation controlled the purse strings—over $1.2 billion in ETH at peak—and the narrative. When EIP-1559 sparked controversy, it was the Foundation’s political muscle that pushed it through.
But that model assumed benevolent centralization. After the Merge, the power levers multiplied. Stakers now control block production. Client teams control execution-layer interpretation. Infrastructure providers control access. And protocols like Lido control the largest pool of stake. Each of these groups now possesses a structural veto: stakers can fork, clients can delay upgrades, infrastructure can censor transactions, and protocols can redirect liquidity.
The Foundation remains wealthy and loud, but it no longer dictates outcomes. It has become first among equals—a research grant-maker, not a commander.
Core: Mapping the Multi-Node Reality
Let’s take this apart systematically. The architecture of trust, engineered for failure, is now being retrofitted with distributed brakes.
1. Staker Power: Not Just Consensus, But Governance
Stakers don’t only validate blocks; they vote on chain upgrades via client adoption. During the Dencun upgrade, a significant minority of stakers ran older client versions despite coordinated release schedules. The delay wasn’t malicious—it reflected a lack of economic incentive to upgrade immediately. The Foundation cannot force stakers to update; it can only persuade. When persuasion fails, the upgrade either stalls or risks a split.
On-chain data from beaconcha.in shows that for EIP-4844 implementation, 7% of validators took over four weeks to adopt the required client updates. In a foundation-dominant model, that number would be near-zero within a week. The market’s reaction? Nothing. Because the market hasn’t noticed this governance latency yet.
2. Client Team Sovereignty
Geth now commands roughly 70% of execution-layer clients. Nethermind, Besu, and Erigon split the rest. A single critical bug in Geth could halt the chain—or, more subtly, give its maintainers enormous leverage over which EIPs get implemented. During the Shanghai upgrade, Geth team unilaterally delayed support for one EIP due to resource constraints. The Foundation didn’t override them; it adjusted the timeline.
Client teams are no longer contractors. They are independent nodes with their own roadmaps, funding sources (Ethereum Foundation grants, but also external sponsors), and internal politics. Their power comes from code: if they refuse to implement an EIP, that EIP is effectively dead.
3. Infrastructure as Gatekeeper
Infura and Alchemy together serve over 80% of Ethereum’s RPC traffic. They are not validators, but they control what most users see. If Infura decides to filter certain transactions or block access to a dApp for compliance reasons, the user experience is broken—regardless of what the Foundation or validators want.
In 2023, Infura briefly blocked IP addresses from sanctioned regions following OFAC guidance. The Foundation issued a statement supporting decentralization, but took no action against Infura. Why? Because the Foundation itself relies on Infura for its own infrastructure. This is not a conspiracy; it’s a structural dependency that turns infrastructure providers into de facto policy enforcers.
4. Liquid Staking Protocols: The New Block Producers
Lido now controls over 33% of all staked ETH. That’s a single protocol’s node operators deciding which blocks get proposed. Lido’s governance token (LDO) is controlled by a DAO with its own interests. If Lido’s DAO decides to support a contentious EIP, it can direct its 200,000+ validators to adopt that EIP’s client fork. The Foundation has no equivalent lever.
Data from Dune shows that Lido’s node operator set includes professional staking firms, large exchanges like Coinbase, and individual operators. Each has signing keys. But all follow Lido’s governance decisions. That’s a centralized coordination layer within a supposedly decentralized staking set.
The Result: A Fedora, Not a Monarchy
What we have is essentially a federation of five power blocs: Foundation (influence), stakers (consensus), client teams (implementation), infrastructure providers (access), and liquid staking protocols (block production). Each can block or delay decisions. None can unilaterally push an upgrade through.
This is healthier than a single foundation controlling everything—it reduces the risk of a corrupt or captured leadership. But it introduces a new class of risk: governance gridlock. The architecture of trust, engineered for failure, now has multiple failure points that can halt progress.
My forensic experience with the Celsius collapse taught me to look at where power actually resides, not where whitepapers claim it resides. On-chain trace analysis of Celsius showed a $2.1 billion shortfall hidden behind PR spin. Similarly, Ethereum’s current governance has a shortfall: the illusion of Foundation dominance masks a more complex reality with no clear resolution mechanism when blocs disagree.
Contrarian: What the Bulls Got Right
Let me be fair. The dispersion of power has real benefits. It makes Ethereum more resilient to regulatory attack. If the Foundation were ever designated a “controlling entity” by the SEC, the network would survive because the Foundation no longer controls the upgrade process. This was the primary argument against classifying ETH as a security—and it just got stronger.
Second, multi-node governance improves the quality of decisions. When L2s demand simulation periods, they are protecting users from gas fee volatility. When client teams push back on incomplete EIPs, they prevent buggy deployments. The EIP-4844 blob mechanism was delayed several times by validator and client feedback, and the final implementation was measurably more stable than the initial proposal.
Third, the market has priced this in implicitly. ETH’s premium over other L1s is partly due to its perceived institutional-grade governance. A federation is harder to subvert than a monarchy. Institutional investors like that.
But the blind spot is critical: a federation without a constitution is a recipe for paralysis. There is no formal mechanism to resolve disputes between blocs. The Foundation can convene meetings, but it cannot force compliance. If Geth team and Lido DAO disagree on the next hard fork, who breaks the tie? Currently, no one.
Takeaway: Watch the Signals, Not the Noise
Ethereum’s silent power transition is not a bug; it’s a feature of maturation. But maturation comes with new failure modes. The network survived the transition from PoW to PoS because stakers aligned incentives. The next test is whether these five power blocs can align incentives without a central arbiter.
I’ll be watching three signals. First, client diversity: if Geth’s share exceeds 80%, the infrastructure bloc becomes too dominant. Second, Lido governance turnout: if LDO holders start vetoing L2-friendly proposals, expect friction. Third, Foundation funding shifts: if the Foundation cuts core developer grants to fund external DAOs, it’s formally ceding power.
Until then, the architecture of trust, engineered for failure, remains fragile—not because the engineers are incompetent, but because they built a system so decentralized that no one can steer it. That’s both the triumph and the terror of Ethereum’s new governance.
