Ethereum's move to proof-of-stake changed who runs the network. Instead of miners with big electricity bills, validators hold stake as collateral. Liquid staking made that role available to anyone with a few dollars, and the total amount of ETH securing the network went up sharply. That's a genuine win.
The same tool also concentrated a lot of that stake into a small number of protocols. Whether that's a real problem, a manageable one, or something the community should shape more actively is a live debate. This is a look at both sides — no preaching, just the shape of the trade.
What liquid staking changed
Before liquid staking, becoming a validator meant locking 32 ETH, running a node 24/7, and knowing enough about servers to keep uptime high. Plenty of people didn't have 32 ETH or the technical patience, so they stayed on the sidelines. Liquid staking let anyone deposit any amount, get a token back, and quietly earn a share of validator rewards.
The result: total staked ETH has grown from roughly 10% of supply at launch toward 25-30% today (numbers move). That's more security for the network overall. It's also good for retail users who now have an on-chain alternative to leaving ETH on an exchange.
Where the concentration shows up
The awkward part is how that new stake got pooled. A large share flowed into a handful of protocols. Illustrative rough shares that have been reported at various points:
| Protocol / channel | Approximate share of staked ETH |
|---|---|
| Lido | ~25-30% |
| Coinbase (cbETH & institutional) | ~10-15% |
| Binance | ~4-6% |
| Rocket Pool | ~2-3% |
| Frax, Stader, others | Smaller shares |
| Solo & unpooled stakers | The remainder |
These numbers shift constantly. The point isn't the exact percentages — it's that a single protocol has at times sat close to the informal one-third threshold that gives it certain powers on the Ethereum consensus layer.
Why concentration matters
Ethereum's security assumes many independent validators. Concentration raises a few concerns:
- Censorship. A protocol that controls a large share of blocks could, in theory, choose which transactions to include.
- Governance capture. Decisions about validator selection, fee splits, and upgrades sit with the protocol's governance token holders.
- Correlated failure. A bug in one dominant liquid staking contract could affect a huge share of stake at once.
- Regulatory pressure. A single dominant issuer is easier for regulators to lean on than thousands of independent operators.
None of these are theoretical panics — they're the same worries that shaped Bitcoin's mining-pool debates for years.
The more nuanced picture
The concentration story is real, but it's not as neat as "one entity runs a third of the network":
- Lido isn't one operator. It delegates to dozens of independent node operators with different geographies and setups. A DAO chooses the set.
- Coinbase's staked ETH is split across many internal validators and is subject to legal jurisdictions that limit certain actions.
- Solo staking has grown too, aided by tools that lower the technical bar.
- Distributed validator technology (DVT) lets a single validator's duties be split across multiple machines and operators, which can reduce correlated failure risk.
You can still hold a strong view about whether the current mix is healthy — reasonable people land in different places. See staking vs liquid staking for how the trade-offs compare.
What users can do about decentralization
If network decentralization matters to you, a few practical choices exist. You can split stake across multiple liquid staking protocols instead of defaulting to the biggest. You can pick protocols with permissionless node operator sets (like Rocket Pool) instead of curated ones. You can run a solo validator or a shared one with DVT tools, keeping your stake out of any pool. Or you can accept the current mix and know what you're accepting. There's no single right answer. But the fact that individual users can shift these percentages, one small decision at a time, is one of the more interesting features of proof-of-stake — the outcome is not fixed. And keeping staking risks in mind while making those choices is worth it either way.