If an institution holds millions of ETH and stakes most of it, does Ethereum become more centralized? The question sounds simple. The answer, as framed in the source article, is not.
Over the past year, two changes have unfolded at the same time in Ethereum staking. More institutional capital has entered the market, and EIP-7251 has raised the maximum effective balance for a single validator from 32 ETH to 2,048 ETH. Those shifts may look similar on the surface, but they describe different things: larger pools of capital are becoming more concentrated, while more ETH is also being consolidated into fewer validators.
Institutional staking is taking a larger share of the market
According to Lido’s report for the first half of 2026, Ethereum’s staking market grew from about 36.3 million ETH at the start of the year to 43.1 million ETH by the end of June, a gain of about 19% over six months. The figure includes ETH that had entered the queue but had not yet been activated, while excluding ETH in the exit queue.
The bigger shift was in where the money came from. Institutional staking rose from 25.9% of the market to 35.3% over the same period, an increase of nearly 10 percentage points. In other words, a growing share of new staked ETH is coming from corporate treasuries, exchanges, custodians, and other professional market participants.
Market-share data points in the same direction. As of the end of June 2026, Lido remained the largest staking protocol with about 21.2% of all staked ETH, but that was down from 23.9% at the start of the year. Bitmine, Coinbase, and Binance reached about 11.5%, 10.9%, and 7.9%, respectively. Bitmine stood out, adding roughly 10 percentage points in just half a year.
That marks a change in the way centralization is being discussed. For years, debate around Ethereum staking concentration focused heavily on liquid staking protocols such as Lido, with concerns that too many users were routing staking activity through the same platform. The article says the structure of new inflows is now shifting as institutional ETH holders move into native staking at scale, making liquid staking protocols less dominant as the main destination for fresh capital.
Owning staked ETH is not the same as operating validators
The article argues that one of the easiest mistakes in this debate is to treat ownership of staked ETH as identical to control over validator activity. It breaks the issue into three layers.
- Asset control: who owns the staked ETH and who can decide to exit, withdraw, or switch service providers.
- Validator operations: who runs the nodes, manages validator signing keys, and handles block proposals and attestations.
- Infrastructure and decision-making power: whether large numbers of validators rely on the same clients, cloud providers, or MEV relays, and how much a platform’s governance structure or management team can coordinate validator behavior.
Those three layers do not always line up. One institution may own 1 million ETH and stake all of it natively, while distributing that stake across more than a dozen independent node operators. In that case, ownership is concentrated, but validator operations may be spread across different service providers, regions, and client software stacks.
The reverse can also be true. Tens of thousands of users may each hold 32 ETH and still rely on the same node operator to perform validator duties. Ownership would look dispersed, while signing, operations, and infrastructure could remain highly concentrated underneath.
That is why the article says the more important question is who actually executes consensus and whether different participants share the same points of failure or the same decision dependencies.
Large institutional positions still carry risk. If one institution not only owns a large amount of ETH but also controls all related validator signing keys, its influence over network consensus cannot be ignored. Even when outside node operators are used, the asset owner may still shape validator behavior indirectly through service agreements, capital allocation, exits, and redelegation.
Ethereum consensus assigns weight according to effective staked balance. Once a participant coordinates or controls enough staking weight, it can affect finality. The article notes that if a coordinated group controls more than one-third of active staking weight, it can under certain conditions prevent the network from reaching finality.
For that reason, the centralization risk of large-scale staking cannot be judged only by how much ETH a single institution owns. Nor does the use of multiple service providers automatically remove the risk. Market share alone is too blunt a measure.
EIP-7251 changes how large staking positions are structured
The source article places special emphasis on the Pectra upgrade and EIP-7251. The proposal raised the maximum effective balance for a single validator from 32 ETH to 2,048 ETH, while keeping the minimum activation threshold at 32 ETH.
Under the old structure, someone holding 1,024 ETH and wanting to stake it natively would usually split that amount into 32 separate validators. Each validator had its own signing identity and had to participate in consensus independently, creating additional overhead in key management, message propagation, and state maintenance.
After EIP-7251, the same 1,024 ETH can, under protocol rules, be carried by a single 0x02 validator. That reduces management costs and lowers extra load at the consensus layer.
The article also highlights a more direct change: native compounding. Traditional 0x01 validators have an effective balance cap of 32 ETH. Rewards above that level do not increase staking weight, so anyone who wants to restake those rewards natively has to arrange that separately.
Large staking pools could work around this by aggregating rewards from many validators, collecting another 32 ETH, and launching a new validator. For long-term holders with only one or a few validators, that process was less practical because rewards accumulate gradually over time rather than arriving as a fresh 32 ETH all at once.
EIP-7251 allows consensus-layer rewards to remain in the validator’s own balance and, under protocol rules, increase effective balance in 1 ETH increments until the 2,048 ETH cap is reached. A process that previously required users to withdraw rewards, accumulate them, and stake again can now happen inside the protocol.
For large institutions, that means fewer validators and lower costs in key management and operations. For ordinary long-term holders, it removes the need to wait and repeatedly rebuild another 32 ETH position just to restake smaller rewards. The article describes this as Ethereum making staking more flexible at the protocol level for larger-scale, more institutional, and more professional participation.
Fewer validators do not automatically mean more centralization
The added convenience comes with new operational requirements. A 0x02 validator does not, by default, periodically sweep rewards above 32 ETH out of the validator balance the way the traditional setup did. If a user wants to use rewards for spending, trading, or other asset-management purposes, a partial withdrawal has to be initiated under protocol rules.
The article also notes that converting a traditional 0x01 validator into a 0x02 validator is irreversible. For large holders managing validators on their own, consolidation requires caution. An error in the target validator or authorization details could lead to serious losses.
Stepping back, the article sees two trends moving in parallel. One is at the market level: institutional capital continues to enter, and ETH staking is becoming a more important part of long-term asset management. The other is at the protocol level: upgrades such as EIP-7251 reduce the old mechanical splitting of validators and let large pools of capital participate in consensus more efficiently.
Still, higher efficiency does not by itself mean stronger decentralization. A smaller validator count does not prove the network is more centralized. A large validator count does not prove the opposite if signing power and operational control remain concentrated underneath. The deciding factor is still the real distribution of control and whether the network can avoid excessive dependence on a small number of operators, clients, and infrastructure providers.
For users, APR is no longer the only question
The article closes by saying users comparing ETH staking options should not focus only on APR. It cites imToken’s non-custodial ETH staking as an example, saying the product offers both compounding validators and auto-withdrawal validators. Users can choose based on whether they want to accumulate ETH over the long term or receive staking rewards that are periodically available for use.
In that setup, withdrawal control remains with the user, while professional service providers handle validator signing and operations. The article presents that as a way to keep control over staked assets without requiring users to run nodes themselves.
From that perspective, the next phase of Ethereum staking is about giving ETH holders of different sizes a clearer, more controllable, and more efficient way to participate. The longer-term shift worth watching, the article suggests, lies in that structural change.
