Rocket Pool Review

Rocket Pool scores 9.0 out of 10 in our staking platforms ratings, up from 8.0. That ties it with Marinade, and ties are listed alphabetically. The earlier review marked fees down because the commission rate was not stated in the source we could open. Rocket Pool's published documentation now states it in full: 14% of the rewards on pooled ETH, split 5% and 9%, plus a 0.05% deposit fee. The documentation also explains who absorbs validator losses, and in what order. Rocket Pool keeps the category's only 10 for decentralisation, because anyone can run a node without approval.

This is desk research, checked on 1 November 2026. Rocket Pool's documentation website blocked automated access, so the evidence comes from the same documentation in its official GitHub repository and from the protocol's code repository. We did not stake ETH or run a node.

How Rocket Pool scores in staking platforms

Five criteria, each worth 20%. Two marks rose.

Arithmetic: (10 + 9 + 10 + 7 + 9) / 5 = 45 / 5 = 9.0.

What changed: Fee transparency rose from 6 to 10 after the protocol FAQ in Rocket Pool's documentation repository was found to state the commission and its split. Risk disclosure rose from 6 to 7, because the same FAQ sets out the order in which losses fall and gives a worked example. The July review had only the code repository's README to go on.

Strongest criteria: decentralisation and fees

Decentralisation (10)

Rocket Pool's documentation describes node operation as open to anyone with the hardware and the bond. Since the Saturn 1 upgrade, each new validator needs 4 ETH from its operator and 28 ETH from the rETH pool. Before Saturn 1, the minimum was 8 ETH. The bond size is a parameter that the Protocol DAO can change by vote. No other service in this category lets any operator join without approval, which is why this is the only 10 for this criterion. One caveat sits outside this mark: the rETH exchange rate is reported by an invite-only Oracle DAO, and that is a governance dependency, not a limit on who can run validators.

Fee transparency (10)

The protocol FAQ says the total commission "is 14%, split into a 5% base commission share" paid to every operator and "a 9% voter share" paid to operators who stake RPL. The commission comes from rewards on the 28 ETH borrowed from the pool, and governance can adjust both shares. Staking through the protocol also costs a 0.05% deposit fee. rETH holders therefore earn the network's staking rewards less about 14%, and the full split is published, which is what this criterion rewards. Older minipools keep the commission written into their own contracts.

Middle and weaker criteria: custody, risk and exit

Custody & control (9)

A staker holds rETH in their own wallet, and the contracts that hold and move funds are open source under GPL-3.0. The staked ETH itself sits in protocol contracts and in validators run by independent operators. Saturn 1 added a delay between a passed upgrade and its execution, which the documentation says allows users "ample time to exit." Even so, the staked ETH is not under the user's own key, so the mark is 9.

Exit & liquidity (9)

There are two exit routes. Through the protocol, rETH can be burned for ETH, but only when the deposit pool holds enough ETH. Otherwise the app shows "Insufficient liquidity" and points the user to a market route instead. Saturn 1 keeps a buffer of up to 1% of rETH value inside the rETH contract to prioritize those burns. On a market, rETH can be sold through Uniswap or through CoW Swap, a third-party route the docs flag as outside the protocol's control. rETH can always be sold, but redemption at the protocol rate is not guaranteed, which keeps the mark at 9.

Risk disclosure (7)

The FAQ explains how a validator loss is handled. The operator's bonded ETH covers it first, and staked RPL covers it next. Only the remainder "is spread across the entire network through a reduction in the relative value of rETH." A worked example shows the arithmetic. The code was audited by Sigma Prime and ConsenSys Diligence before the 2021 launch and by Cantina, Bailsec and Sigma Prime for Saturn 1. There is no insurance beyond operator bonds and RPL collateral, no cap on what a staker might lose in a loss larger than those bonds, and no published incident history. That leaves this criterion clearly explained but with limited cover. One detail is out of date: the worked example still uses the pre-Saturn 8 and 16 ETH bonds.

What the score does not cover

• Running a node. Operator economics, RPL staking and hardware are outside a staker's rating.
• rETH market price. We did not measure how far rETH trades from its protocol rate.
• Price of staking. A 14% commission is higher than Lido's 10%. This rubric scores disclosure rather than price, so a reader comparing net yield should do that arithmetic separately. Our guide to staking returns after costs shows how.

Rocket Pool vs Marinade and Lido

Marinade also scores 9.0 and is listed first only because ties are alphabetical. Rocket Pool is ahead on decentralisation (10 to 9). Marinade is ahead on risk disclosure (8 to 7), because its validator bonds fund published, rule-based compensation for downtime. They tie on fees (10), custody (9) and exit (9). Lido (8.0) is Rocket Pool's direct Ethereum peer and also earns 9 for exit. It trails on decentralisation, because its curated operator set is gated, on fees, because part of its split is listed only as "Blended," and on custody (8 to 9). The Lido review has its full marks.

Who it suits

Rocket Pool's rating fits an ETH holder who wants a liquid token backed by a permissionless operator set, with losses absorbed first by the operators' own bonds. It fits less well someone who wants the lowest commission, guaranteed redemption at the protocol rate, or insurance beyond operator bonds. This rating is not a recommendation, and staking rewards and the rETH rate can both fall.