EIP-8062 0x01 validator sweep withdrawal fee (Add sweep withdrawal fee for 0x01 validators)

No call decision · Consensus & staking · EIP status Draft · created 2025-10-28

Imposes a 0.05% fee on the withdrawn amount for 'sweep' withdrawals that automatically return the excess balance to 0x01 validators holding more than 32 ETH. This reflects the resource cost of this withdrawal, which has been free until now, and encourages migration to 0x02 consolidated validators.

Why it is needed

Ethereum is pursuing a roadmap to speed up finality by reducing the number of validators that vote at once. That requires staking service providers (SSPs) to migrate from multiple 32 ETH 0x01 validators to consolidated 0x02 validators that compound their balance. But current 0x01 validators automatically get any balance above 32 ETH withdrawn with no fee, and their capital efficiency is already near ideal, so once you factor in the switching cost there is little reason to move. On top of that, these sweep withdrawals consume resources on both the consensus layer and the execution layer, and nobody pays for that cost. With the roadmap depending on 0x02 consolidation, giving 0x01 validators this kind of accounting advantage is not appropriate, and that is the starting point of this proposal. So it aims to preserve fairness and accelerate consolidation by reflecting the resources that sweep withdrawals consume as a small fee.

What changes

Today

After EIP-8062

Who is affected

In depth

How the fee is applied

Fee calculation is added to the existing for loop in process_withdrawals(). For each withdrawal w, if the validator does not have compounding credentials (has_compounding_withdrawal_credential is false) and the withdrawal amount differs from the full balance (it is not a full withdrawal), the fee is computed as fee = w.amount // WITHDRAWAL_FEE_FRACTION. Since WITHDRAWAL_FEE_FRACTION = 2000, this is 0.05% of the withdrawal amount. The balance is then reduced by w.amount + fee via decrease_balance, and the fee is burned at the consensus layer. The existing withdrawal logic was left unchanged to minimize complexity, at the cost of the balance dropping just below 32 ETH right after skimming. This already happens today when a penalty is applied right after skimming.

Basis for the 0.05% level: gas conversion

The fee approximates the workload a sweep imposes on the execution layer. The breakdown is: 500 gas for a cold touch of an account with no code (per EIP-2780), 1000 gas for one leaf write to the account trie (per EIP-2780), and 704 gas for EIP-4895 withdrawal data (index uint64, validator_index uint64, address bytes20, amount uint64, totaling 8+8+20+8 = 44 bytes, costed like calldata at 16 gas per byte), for a total of 2204 gas. Applying the trailing one-year average base fee of 4.53 gwei gives 2204 x 4.53 = 9,984 Gwei, close to the expected average fee of 11,750 Gwei. This calculation does not include the consensus layer's share of the work cost. EIP-7002 EL-triggered partial withdrawals for 0x02 validators also do not charge for CL work, but instead pay more on the EL side when triggering a withdrawal. Also, 2048 ETH validators are exempt from the sweep fee so as not to undermine consolidation.

Why a rate rather than a flat amount

The fee is deducted at the consensus layer as a rate on the withdrawal amount rather than a fixed amount. As consolidation progresses and the number of 0x01 validators drops significantly, sweeps will cycle through each validator more often and the amount withdrawn each time will be smaller. With a flat fee, the burden could grow extremely large as the frequency increases. A fee proportional to the withdrawal amount keeps the staker's total burden steady even as frequency rises.

Scale of the cost borne by stakers

Assuming a 3% consensus layer yield, the annual cost per validator is 32 x 0.03 / 2000 = 0.00048 ETH, about $2 at $4,167 per ETH. According to Table 1 in the source, 100 validators (about 0.01% of stake) cost $200/year, 10,000 validators (about 1%) cost $20,000, 50,000 validators (about 5%) cost $100,000, and 100,000 validators (about 10%) cost $200,000. For an SSP, the switching cost may exceed the annual fee. Some smart contracts that require 0x01 credentials cannot be changed, and where a full withdrawal and re-entry is needed there is also an opportunity cost. However, the source explains that if the protocol commits to phasing out 0x01 and can raise the fee as needed, the switch becomes something that must eventually happen anyway, so the real comparison is between the recurring fee and the capital cost of funding the switch.

The capital efficiency gap and EIP-8068

Another advantage of 0x01 is capital efficiency. Under the current hysteresis design, 0x02 compounding validators, whose balance grows gradually, end up holding 0.25 to 1.25 ETH more than their effective balance (EB), leaving an average of 0.75 ETH idle and excluded from reward calculations. By contrast, the idle balance of a 0x01 validator averages only 0.0235/2 = 0.01175 ETH. According to Figure 1 in the source, this 0.05% fee does not fully offset the yield loss experienced by 0x02 validators with balances below roughly 1,000 ETH. EIP-8068 is a proposal to eliminate this gap itself with a neutral EB design that raises the EB at a hysteresis threshold of +0.5. There is also the issue of 0x02 validators exploiting hysteresis by reaching an EB of 33 ETH and then partially withdrawing to hold only 32.75 ETH while keeping the 33 ETH EB, which can be addressed separately as an alternative specification of EIP-8068.

Alternative specification: a higher fee

If EIP-8068 is not included in the same hard fork, there is an alternative of raising the fee to compensate for 0x02's lower capital efficiency. The main alternative is setting WITHDRAWAL_FEE_FRACTION = 200, i.e. a 0.5% fee, which gives a 32 ETH 0x01 validator capital efficiency similar to a 128 ETH 0x02 validator. A stronger option is 100, i.e. 1%, under which consolidating into 0x02 is always at least neutral if you run two or more 0x01 validators. A weaker option is 400. At the 0.5% level, this adds $20 per validator per year, meaning an SSP running 5% of stake would face a loss of about $1 million per year, creating a much stronger incentive to consolidate.

Relationship to the fast finality roadmap

The source explains that fast finality is best achieved by shrinking the active validator set that votes each round. Two paths are envisioned. In the path called Orbit SSF, the active set rotates on a stake-weighted basis, so heavily consolidated validators are always active while smaller validators participate less often. In the other path, the number of validator seats is fixed and only the largest validators (e.g. the top 8,192) get seats. Both paths require stake consolidation for the desired stake weighting and optimal performance, and this EIP is a mechanism to pull that consolidation forward.

Current status

This EIP was created on October 28, 2025 and is in Draft status, so details including the fee level may still change. The Security Considerations section of the source states that, as far as the authors are aware, there are no security risks associated with the proposed fee. There has been no core developer call decision yet on whether to include it in a fork.

Original EIP · ← EIPs