Earn Ethereum: one underlying rate, four very different prices

Start with the number nobody can change

Ethereum issues a fixed amount of new ETH to validators, and that issuance is divided among everyone staking. As more ETH gets staked, each validator's share falls. That arithmetic has driven the base staking APR down from the 4% to 5% range seen in 2024 to roughly 3.2% to 3.8% by 2026. On top of that, MEV — value validators capture by ordering transactions within a block — adds roughly 0.3% to 0.5%.

So the total pie is somewhere around 3.5% to 4.3% gross, and every platform is drawing from it. No exchange has a better validator. No protocol has a special arrangement with Ethereum. When you compare two ETH staking products, you are comparing fees and liquidity, not yields.

This is genuinely useful to internalise, because it immediately exposes anything advertising a much higher ETH return as either quoting gross before commission, or bolting on a second mechanism — lending, restaking, incentives — with its own separate risks.

Four routes, one yield source

Solo staking. Run a validator with 32 ETH and keep the entire reward. You also accept uptime responsibility, hardware management and slashing exposure. For anyone not already running infrastructure, the operational burden outweighs the fee saving.

Exchange staking. Deposit ETH, click stake, receive rewards. Coinbase and Kraken both offer it. Simplest possible experience, and you pay for it — Coinbase takes roughly 25% to 35% of gross rewards, leaving about 3.5% net. Kraken's commission reaches 30% on some assets and its flexible option pays rewards on only part of the allocated balance.

Liquid staking. Deposit ETH with Lido or Rocket Pool and receive stETH or rETH. You keep a tradable, collateral-eligible token while the underlying ETH stays staked. Lido charges a flat 10% and pays 3.8% to 4.1%; Rocket Pool's more distributed operator set pays 3.5% to 3.9%. The best economics available to a non-institutional holder.

Lending ETH. Supplying ETH to Aave or a CeFi savings account is not staking at all — it pays borrower interest instead of consensus rewards, and typically pays less. There is rarely a good reason to lend ETH at 2% when staking it pays nearly double, unless you specifically need the borrowing facility a money market provides.

What each route pays after fees

Rates as of 16 September 2026. Read the terms column carefully — Kraken's headline is gross, and Nexo's is a lending rate rather than a staking rate despite appearing on the same shelf as one.

Liquid staking tokens: stETH, rETH, cbETH

The core idea is simple and genuinely useful. Your ETH is staked and earning, but instead of being locked away you hold a token representing that position. You can sell it, lend it, or post it as collateral. stETH in particular is accepted across most of DeFi, which means a staked position can simultaneously back a loan.

Lido's stETH is the largest by a wide margin and rebases — your balance grows as rewards accrue. It charges 10% and has the deepest liquidity of any liquid staking token. Its size is also the main criticism levelled at it: Lido's share of total staked ETH is large enough that it is an active governance concern within the Ethereum community.

Rocket Pool's rETH takes the opposite approach with a permissionless, distributed node operator set. It appreciates in value rather than rebasing, which some find cleaner for tax purposes. Lower yield by a small margin, meaningfully better decentralisation.

Coinbase's cbETH is the exchange-issued version, convenient if you already hold assets there and subject to Coinbase's commission rather than a protocol fee.

The risk that people underweight is that a liquid staking token is not ETH. During market stress it can trade below the value of the ETH it represents, because sellers want out faster than the exit queue allows. That discount has historically been modest and temporary. It has also been real, and if you are using stETH as loan collateral it is the thing that will hurt you.

Exit queues, and why liquidity is worth paying for

Unstaking ETH directly means joining an exit queue whose length depends on how many validators are leaving simultaneously. In calm conditions that is hours. In a rush for the exits it can run into weeks, and it is precisely during a rush that you most want out.

This is the practical case for liquid staking that has nothing to do with yield. Holding stETH, you do not queue — you sell into a deep secondary market and accept whatever small discount exists at that moment. Holding a staked balance at an exchange, you wait, unless that exchange maintains a liquidity buffer and chooses to front the withdrawal itself.

Whether that optionality is worth anything depends on whether you can imagine wanting your ETH quickly. For a genuine long-term holder, probably not. For anyone who might rebalance, it is worth more than the small yield difference between routes.

Restaking: extra yield, extra failure modes

Restaking lets you reuse staked ETH to help secure additional services, earning additional rewards from them. The pitch is capital efficiency: the same collateral secures two things and earns from both.

The cost is that you inherit a second set of slashing conditions. Your ETH can now be penalised for a failure in a system that is younger, less battle-tested and less economically significant than Ethereum itself. The additional yield has generally been modest, and the additional complexity has not.

We would not put a first ETH position into restaking. For a large holder who already understands the base layer thoroughly and wants to allocate a slice to a higher-risk sleeve, it is a reasonable thing to explore. As a default, it is complexity in search of a return.

What we would actually do with ETH

If you hold ETH long term and are comfortable with a self-custody wallet, a liquid staking token is the clear answer. You capture close to the full network reward, you keep the position usable, and you avoid the exit queue. Between Lido and Rocket Pool, the choice is genuinely between liquidity depth and decentralisation — both are defensible and the yield difference is small enough not to be the deciding factor.

If you are not comfortable with a wallet, exchange staking is fine. You will pay a third of your rewards for the convenience, which is expensive but not absurd given what it buys: recovery, support, and no possibility of losing everything to a signing mistake. Just go in knowing the commission, because most interfaces show you the gross rate.

What we would not do is lend ETH at 2% when staking the same coin pays nearly double from a structurally safer source. That is the one clearly dominated option on this page, and it appears on more platform menus than it should.