Ethereum Validator Check: At 97 Percent, Every Reward Above 32 ETH Is Swept

If you stake Ether, a prefix two characters long decides whether your rewards keep working or leave your validator balance at regular intervals. It is called the withdrawal credential, and on every validator it sits in one of three states: 0x00, 0x01 or 0x02. Only with type 0x02 do rewards stay at stake and earn on top. With 0x01, everything above 32 ETH is deducted automatically. And with 0x00 you cannot reach your balance at all.

On September 9, 2026, cryptoticker.io counted the distribution across the validator set itself. The result is unambiguous: 97.0 percent of active validators still sit on the old type 0x01. The share held by the compounding type 0x02 stands at 2.1 percent. This analysis was carried out by cryptoticker.io on September 9, 2026.

Lido brings compounding mode to testnet: the trigger on September 8

What prompted this count was a report from the previous day. Crypto Briefing reported on September 8, 2026 that the staking provider Lido has launched a new version of its Community Staking Module built around credential type 0x02. The version runs on a testnet for now; a mainnet launch is scheduled for October 2026 according to the report and remains subject to approval by the Lido DAO. Nobody has guaranteed that this date will hold, and this article does not claim otherwise.

What makes it interesting is less the individual provider than the pattern behind it. The large operators are currently moving their validators to the compounding type because it lets them pool capital. Individuals who stake on their own often fall behind without noticing. That gap can be measured, which is exactly why it is set out here.

0x00, 0x01 and 0x02: what the three validator credential types mean

The withdrawal credential is the entry stored in the protocol that says where a validator's balance may flow. It sits within the validator record on the Beacon Chain and appears neither in an account statement nor in an app setting. Three prefixes are currently in circulation, and the names for them stand unchanged in the protocol parameters that this newsroom queried directly from a public beacon node.

Why the prefix, and not the wallet app, is what counts

0x00 (BLS credential): the original state from the period before the Shapella upgrade. No withdrawal target on the execution layer is stored. As long as that remains the case, the validator can neither pay out rewards nor release the capital that was deposited. The balance is locked.

0x01 (execution address): the state that has been standard since Shapella. An Ethereum address is stored and withdrawals work. The catch lies in the upper limit: the balance that counts for the reward calculation is capped at 32 ETH. The documentation on ethereum.org states that any balance above 32 ETH is swept away automatically.

0x02 (compounding credential): the type introduced with the Pectra upgrade. Here the upper limit sits at 2,048 ETH instead of 32 ETH. Rewards stay inside the validator and earn along in the next reward round. The parameter for this is called MAX_EFFECTIVE_BALANCE_ELECTRA in the protocol and stands at 2,048,000,000,000 Gwei, which is exactly 2,048 ETH.

One term that comes up several times deserves a clear definition here: the effective balance is the share of the balance, rounded to whole ETH, that the protocol uses to calculate rewards and penalties. Anything above it earns nothing extra under 0x01.

Counted in house: 97 percent of active validators still sit on 0x01

For this analysis the newsroom drew two mutually independent random samples from the entire validator registry of the Beacon Chain on September 9, 2026 between 15:56 and 16:00 UTC, 6,000 indices each, so 12,000 in total. The data was retrieved through the public beacon interface of a node operator, with every request returning HTTP status 200. At the time of measurement the registry held 2,362,534 validator indices ever issued; of those, 38.6 percent were active in the sample, while the rest had already exited and been paid out.

Among the 4,633 active validators in the sample, the credential types were distributed as follows:

• 0x01: 4,495 validators, 97.02 percent (95 percent interval: 96.53 to 97.51 percent)
• 0x02: 95 validators, 2.05 percent (1.64 to 2.46 percent)
• 0x00: 43 validators, 0.93 percent (0.65 to 1.20 percent)

Both draws arrived independently at the same figure of 97.0 percent for 0x01. Extrapolating the share of active validators to the registry gives roughly 912,000 active validators, with a statistical range of about 892,000 to 933,000.

A second figure from the same measurement explains why the headcount alone is misleading. Measured by capital, the compounding type is long past being a fringe phenomenon: the 95 validators of type 0x02 that were found held an average of around 820 ETH, while the 4,495 validators of type 0x01 averaged 32.008 ETH. In the sample, roughly a third of the staked Ether therefore sat with 0x02. This capital share is considerably less certain than the headcount, because it rests on only 95 observations with a very wide spread; a single large validator shifts it noticeably. As an order of magnitude it still holds: a few large operators compound, and the broad mass of individual validators does not.

The figure of 32.008 ETH is no coincidence, it is the proof of the mechanism. Ether traded at $2,489.61 at the time of measurement (Kraken, September 9, 2026, 16:00 UTC).

The sweep: why your balance keeps falling back to 32 ETH

The protocol calls the automatic pass that pushes surplus balance out of validators to the stored withdrawal address a sweep. It runs without your involvement, without a fee and without any confirmation from you. On a validator with a 0x01 credential, everything above 32 ETH is cleared out in the process.

How quickly this happens can be derived from two protocol values and the count carried out here. Per block the network processes at most 16 withdrawals (MAX_WITHDRAWALS_PER_PAYLOAD), and a block appears every 12 seconds (SECONDS_PER_SLOT). At around 912,000 active validators, a full pass through the registry therefore takes about 7.9 days. Roughly every eight days your validator comes up and is reset to 32 ETH.

The measured average of 32.008 ETH among the 0x01 validators is precisely what this produces: the small remainder that has built up since the last pass. Practically every one of these validators sat just above the mark, and practically none appreciably above it. The capital works, the rewards from it do not work along with it.

How to check which credential type your validator carries in two minutes

The check needs neither keys nor a login. All you need is the index or the public key of your validator, the identifier you also see in your monitoring.

Call up the validator record at a public beacon explorer or directly at a beacon interface and look at the withdrawal_credentials field. What matters are the first four characters of the value. If it reads 0x02, your rewards compound. If it reads 0x01, they are swept out at regular intervals. If it reads 0x00, no withdrawal target is stored at all, and then action is urgent.

Anyone running the validator on their own node finds the same value in the output of the beacon interface under the path for validator states. Anyone staking through a provider usually cannot see it there and has to ask; more on that further down.

A second look at the balance is worthwhile. If it sits permanently at just above 32 ETH even though the validator has been running for months, what you are seeing there is the sweep mechanism at work rather than a display error.

0x00: when no withdrawal target is stored, nobody can reach the balance

The most serious finding of the analysis concerns the smallest group. 43 of the 4,633 active validators in the sample still carried a credential of type 0x00, so 0.93 percent. These validators keep earning rewards, but there is no target that payouts could go to. Neither the rewards nor the 32 ETH originally deposited can be retrieved for as long as that state persists.

Extrapolated to the full set, the share corresponds to around 8,500 validators, with the statistical interval running from about 5,900 to 10,900. Their average balance in the sample was around 37 ETH, appreciably above the 32 ETH of the 0x01 group, precisely because nothing is swept from them and the rewards have been accumulating for years. Multiplying the two gives a locked amount in the order of 220,000 to 407,000 ETH. At the price on September 9, 2026 that corresponds roughly to $550 million to a good $1 billion.

This range is deliberately kept wide and rests on 43 observations, and the text does not smooth it into a tidy single figure. What can be said: this is a relevant amount of capital that would have been retrievable since the Shapella upgrade and is not, because nobody has carried out the switch.

Those affected are typically validators from the early days of staking, whose keys were generated in 2020 or 2021. If you started back then and have changed nothing since, this point belongs at the very top of your list. Switching from 0x00 to a withdrawal target requires the original withdrawal key, which is derived from your seed phrase. Without that key there is no way back.

Switching to 0x02 is possible, costs a fee and cannot be undone

Moving from 0x01 to 0x02 runs through a request that you trigger on the execution layer, for instance through the validator actions of the official launchpad interface or through the tools of your node setup. Technically it is a consolidation request pointed at your own validator, introduced with the EIP-7251 standard.

Three points are worth knowing beforehand. First, a network fee applies, and its size depends on network load. Second, the switch is final: the documentation on ethereum.org states explicitly that the step from 0x01 to 0x02 cannot be reversed. Third, the switch changes the payout behaviour fundamentally, because after it nothing flows out automatically any more. Anyone who relies on those regular inflows as ongoing income loses exactly that.

What comes first when switching out of a 0x00 state

For validators on 0x00 the order is a different one: there, storing a withdrawal target comes first, which is the step to 0x01. Only after that does 0x02 come into question at all.

Whether the switch pays off for you depends on things this text does not know: your holding period, your fee situation and your tax position. It therefore makes no recommendation. It says only that you should know the state before you decide.

Staking through an exchange or pool: the one question to put to your provider

Most investors in Germany stake through an exchange, a pool or a staking product rather than through a validator of their own. For them the credential type is not directly visible, yet it works in exactly the same way.

The question you can put to your provider is simply this: are my rewards left at stake and compounded, or are they paid out and then sit idle? Providers that have moved to 0x02 can answer it. Where the answer dodges, you have learned something too.

A related point concerns the payout side. How long you wait for your capital in the event of an exit depends on the network's exit queue and therefore on something other than the credential type. How that queue works and how much it fluctuates is set out in our analysis of the Ethereum staking queue. Which supervisory requirements have applied to staking offerings in Germany since MiCA is something we wrote up in a separate piece on staking under MiCA. Anyone wanting to see different providers and their terms side by side will find them in our overview of staking platforms.

Open in tax terms: why the moment of receipt under 0x02 is an unsettled question

Switching the credential type changes something that counts for tax: the moment at which rewards arrive at your address. Under 0x01 they flow in regularly, roughly on an eight-day rhythm. Under 0x02 they stay in the validator balance and merely increase a number there.

Whether that shifts the moment of receipt for tax purposes has not been settled conclusively. The German Federal Ministry of Finance circular of March 6, 2025 on individual questions of the income tax treatment of crypto assets is the relevant administrative guidance, but it does not address compounding validators explicitly. This piece therefore asserts no legal position where none is established. Clarify the point with tax advice before you switch, and document the moment of the switch properly.

Independently of that, the documentation duty remains. Tools that record this automatically are listed in the overview of crypto tax tools and portfolio trackers.

Limits of this analysis: what the sample does not show

The survey rests on 12,000 randomly drawn validator indices from a registry of 2,362,534 entries, measured at a single point in time on September 9, 2026. The result is therefore a snapshot and not a full census.

Three things this analysis expressly could not deliver. The data set assigns no operator to any credential, because the Beacon Chain does not contain that information; which shares fall to large providers and which to individuals therefore remains open. The measurement says nothing about the operators' place of residence, and so nothing about how many German stakers are affected. And the capital share of type 0x02 rests on 95 observations with a very wide spread, which is why it stands here only as an order of magnitude and not as a robust ratio.

All of it can be recalculated. The protocol parameters come from the configuration endpoint of a public beacon node, the validator data from the state endpoint of the same node, and the definitions of the three credential types are set out in the documentation on withdrawal credentials at ethereum.org. The trigger, the testnet version of the Lido module, can be read in the Crypto Briefing report of September 8, 2026.

Checking your Ethereum validator credential: what to take away

1. Look at the prefix, today. Call up your validator record and read the first four characters of withdrawal_credentials. With 0x00 you store a withdrawal target first, otherwise your capital stays locked. Anyone who does not stake yet and is weighing an entry should first get an overview of the available staking platforms and their terms.
2. Decide consciously about the switch to 0x02. It lets rewards compound, it is final, and it ends the regular inflows. Check beforehand whether you need those inflows, and record the moment of the switch with a tax and portfolio tool.
3. Ask your provider if you do not validate yourself. A single question is enough: are my rewards compounded or paid out? How the waiting times on exit are developing can be followed in our analysis of the staking queue.

(As of September 9, 2026. This article is not investment advice. Prices and fee structures change; check the terms with the provider before you buy.)