Lido vs Rocket Pool vs Frax: Getting Your ETH Back

The pitch for liquid staking is a promise about the exit. Lock 32 ETH into a validator yourself and the capital is committed to the network; getting it back means joining a withdrawal queue and waiting. Deposit into Lido, Rocket Pool or Frax instead and you receive a token that keeps moving. You can trade it, borrow against it, or redeem it for the underlying ETH, all while the stake beneath it earns rewards. The liquidity is the whole point. It is also the first thing that gets tested when markets turn.

On 14 October 2026 that test is easy to see. Ethereum’s validator exit queue holds more than 967,000 ETH, a wait of roughly 17 days to unstake at the protocol level, while the activation queue runs past 1.2 million ETH and 21 days, according to validatorqueue.com. ETH itself trades near $2,475, about 50% below its August 2025 record, per CoinGecko. When the base layer makes you wait two and a half weeks for your own stake, the token you hold in the meantime, and how reliably it converts back into ETH, stops being a footnote and becomes the product.

This is a comparison of Ethereum’s three best-known liquid staking protocols through that single lens: how you get your ETH back. Along the way it covers what each one is, how the receipts pay you, what the yield really is after fees, whether the governance tokens capture any of the value, and where the SEC has landed. The three began as rival answers to the same question. By late 2026 they are not even competing for the same thing.

What liquid staking is, and the problem it solves

Ethereum secures itself through proof of stake: validators post collateral and earn rewards for proposing and attesting to blocks. Running one yourself takes 32 ETH (about $79,000 at today’s price), a dedicated node that stays online around the clock, and a tolerance for slashing, the penalty for downtime or double-signing. Two barriers keep most holders out: the 32 ETH minimum and the lock-up once you are in.

Ethereum co-founder Vitalik Buterin has named the first problem directly, telling Decrypt in 2024 that “the #1 thing in becoming more home staking friendly is to reduce the 32 ETH requirement.” Liquid staking is the market’s workaround for both barriers at once. A protocol pools deposits of any size, runs the validators on depositors’ behalf, and issues a transferable token that represents the deposit plus its accruing rewards. You keep the yield; the token keeps your capital liquid. That token is the liquid staking token, or LST.

The pool is deep. Around 43.7 million ETH, roughly 35.8% of all supply, is staked, earning a base rate near 2.58% before tips and MEV, per validatorqueue.com. Liquid staking tokens are the single largest slice of that stake, and Lido’s stETH alone ranks among the ten most valuable crypto assets by market capitalization, according to CoinGecko. Three protocols dominate the non-custodial side of that market, and they take three very different routes in and, more importantly here, out.

“Liquid” means two different things

Before comparing the protocols, it helps to separate the two exits an LST actually offers, because they behave very differently under stress.

The first is the secondary market. You sell your stETH, rETH or sfrxETH to another buyer on a decentralized exchange such as Curve or Uniswap, and you walk away with ETH immediately. The catch is price: you get whatever the pool quotes, which in calm times is essentially one-for-one but in a panic can slip below the value of the underlying stake. Speed is guaranteed; price is not.

The second is primary redemption. You hand the token back to the protocol, it burns the receipt, and it returns the underlying ETH at full value. Here the price is guaranteed (you get the stake you are owed) but the speed is not, because the protocol may need to exit a validator to free the ETH, and validator exits run through the same network queue that currently takes about 17 days. Each protocol softens that wait with a buffer of pending deposits and reserves, but a buffer only lasts as long as deposits roughly match withdrawals.

In normal conditions the two exits converge: arbitrageurs keep the secondary price near redemption value, because anyone can buy a discounted LST and redeem it one-for-one. In a rush for the door they diverge, and the gap between them is the clearest measure of how liquid a liquid staking token really is. Put plainly, the secondary market sells you speed and charges you in price, while primary redemption gives you full price and charges you in time. Every protocol below is really a different answer to how it manages those two charges, so keep both in mind.

Rebasing vs reward-bearing: how your receipt pays you

LSTs deliver their yield in one of two ways, and the difference shapes both how you use them in DeFi and how the peg looks on a chart.

A rebasing token tracks ETH roughly one-for-one and pays you by increasing the number of tokens in your wallet each day. Lido’s stETH is the canonical example: hold 10 stETH and the balance ticks up as rewards land, while each unit stays close to the price of one ETH. That one-for-one target is why people watch stETH for a depeg.

A reward-bearing token keeps your balance fixed and lets each unit appreciate against ETH instead. Rocket Pool’s rETH, Frax’s sfrxETH and Lido’s wrapped wstETH all work this way. One wstETH is now worth about 1.2503 ETH, one rETH about 1.1747 ETH, and one sfrxETH roughly 1.17 ETH, per CoinGecko (wstETH, rETH, sfrxETH). That premium is not a mispricing; it is accumulated staking reward. A reward-bearing LST is supposed to cost more than ETH, and the gap widens every day it exists.

The practical upshot is twofold. In DeFi, many lending markets and automated market makers prefer reward-bearing tokens, because a balance that silently grows every block is awkward for a smart contract to account for, which is why wstETH rather than raw stETH shows up as collateral across so much of the ecosystem. In tax terms, rebasing tokens generate a steady stream of reward events in many jurisdictions, while reward-bearing tokens defer that into a single gain when you sell. For redemption, what matters is that both kinds ultimately claim the same underlying ETH; the exchange rate just tells you how much.

Lido: the deep end of the pool

Lido is the giant. Its stETH has a market value near $24 billion across about 9.68 million tokens, a top-ten crypto asset in its own right (CoinGecko). Lido still accounts for roughly a fifth to a quarter of all staked ETH, though its share has been sliding; CCN clocked a year-to-date low around 22.8%, down from a peak near 32% in late 2023 as custodial and institutional products took most of the new inflows.

On the exit, Lido has two advantages. First, its secondary market is the deepest in the category: the stETH and wstETH pools are among the largest in DeFi, so you can usually swap size into ETH with minimal slippage (StakingRewards). Second, primary withdrawals have been live since Ethereum’s 2023 Shapella upgrade, and Lido keeps a buffer of incoming deposits that lets many redemptions clear in days rather than the full validator-exit wait, as long as net flows stay balanced.

The big 2026 story is consolidation. Through its Curated Module v2 upgrade, Lido is migrating about 265,000 legacy 32-ETH validators into roughly 4,000 higher-capacity ones that hold up to 2,048 ETH each and auto-compound rewards, restructuring some $16 billion of staked ETH, per The Block. Lido’s own blog frames it as a leaner, more sustainable validator set, and asset manager Bitwise became the first institutional operator in the rollout, according to Crypto Briefing. The migration, targeted for completion by Q1 2027, roughly doubles the share of Lido stake held in compounding validators and is a big reason the network’s total validator count is falling even as staked ETH rises. Lido is also pushing beyond staking itself, floating a roughly $60 million plan to move into institutional stablecoin yield, The Defiant reported. The throughline is that the largest liquid staking protocol is deliberately becoming something broader than a single stETH pool, which is worth keeping in mind when you weigh how committed it will stay to defending that pool’s depth over the long run.

Rocket Pool: the permissionless holdout

Rocket Pool is the credibly neutral option. Anyone can become a node operator by posting a bond plus some RPL collateral, which makes its validator set permissionless rather than curated. Its rETH is reward-bearing and much smaller than stETH, with a market value around $0.9 billion across roughly 315,000 tokens at about 1.1747 ETH each (CoinGecko). ProtocolWatch counted on the order of 1,500 active node operators in mid-2026, still one of the most distributed operator bases in ETH staking. That distribution is the whole pitch: a permissionless operator set is harder to censor, capture or pressure than a curated one, so holders who care about credible neutrality accept smaller scale in exchange for not trusting a gatekeeper to decide who runs the validators.

The protocol’s 2026 upgrade, Saturn One, went live on 18 February and reworked the economics in operators’ favor. It cut the per-validator bond from 8 ETH to 4 ETH, introduced megapools that bundle many validators under one contract, added express and standard activation queues, and replaced the old commission with a revenue split that pays operators a base rate plus a bonus funded by staked RPL and paid in ETH, per The Defiant. Crucially, holding RPL is now optional for launching validators.

On the exit, rETH behaves like Lido’s with less depth behind it. You can burn rETH for ETH straight from the protocol’s deposit pool when that pool holds enough idle ETH; when it does not, the redemption waits on an operator exiting a minipool or megapool through the same network queue. There is no instant-withdrawal guarantee, and the secondary market for rETH is thinner than stETH’s, so a large holder trying to leave in a hurry will feel the difference more sharply here than at Lido.

Frax: the contender that changed the subject

Frax is the smallest of the three in staking terms, and that is partly by choice. Its model splits the receipt in two: frxETH is a liquid token that carries no yield, and sfrxETH is an ERC-4626 vault that concentrates all the staking reward into fewer hands. Hold frxETH for maximum liquidity and DeFi flexibility, or stake it into sfrxETH to actually earn. On the way out you unwind the same path in reverse: redeem sfrxETH back to frxETH through the vault, then swap frxETH for ETH on a market or redeem it directly, which means Frax users face two conversion steps and, because the pools are small, more slippage than a stETH holder would. Today frxETH is worth about $151 million and sfrxETH about $108 million, per CoinGecko (frxETH, sfrxETH), a fraction of Lido’s or Rocket Pool’s footprint.

The reason Frax reads differently from the other two is that it has changed the subject. Through its North Star overhaul in 2026, Frax renamed its FXS governance token to FRAX and made it the native gas token of its own Fraxtal blockchain, replacing frxETH in that role, as PANews reported. Founder Sam Kazemian framed the rebrand as reflecting what Frax has become, “a scalable stablecoin operating system, not just a single protocol,” built around its GENIUS Act-compliant frxUSD stablecoin. Liquid staking is no longer the company’s headline act; it is a supporting product inside a stablecoin ecosystem. For readers weighing where that side of the market is heading, our look at stablecoin AMM design in 2026 covers the curve mechanics that keep those assets near their peg.

That does not mean the staking product is standing still. Frax has slated a capped mainnet launch of its frxETH v2, which adds a lending market between the protocol and its node operators, for around mid-October 2026, per its own project updates. But the strategic center of gravity has clearly moved, and that matters for anyone counting on Frax to defend a deep two-sided market for frxETH the way Lido defends stETH.

Architecture and exit mechanics, side by side

The table below sets the three protocols against each other on the features that decide how, and how fast, you get your ETH back. Figures are as of 14 October 2026.

How fast can you really get your ETH back?

Start with the hard constraint no protocol can engineer away: the network’s own exit queue. Ethereum processes validator exits at a fixed pace, and right now the queue holds about 967,000 ETH, a wait of roughly 16 days and 19 hours, while the activation queue is even longer at more than 1.2 million ETH, per validatorqueue.com. Any redemption that forces a validator to leave inherits that wait.

The LSTs soften it with buffers. Lido, Rocket Pool and Frax all hold a reserve of ETH from new deposits and reward income; when you redeem, they pay you out of that reserve first and only queue a validator exit if the reserve runs dry. In balanced conditions, where deposits roughly match withdrawals, that buffer is enough to turn a two-and-a-half-week wait into a few days. The deeper the protocol and the steadier its inflows, the more reliably the buffer holds, which is exactly why Lido’s exit tends to feel faster than Rocket Pool’s or Frax’s.

The buffer breaks precisely when you most want out. If a shock sends everyone to the exits at once, deposits stop, the reserve empties, and redemptions drop back onto the raw queue. At that point your only fast exit is the secondary market, and selling into a one-sided market means accepting a discount. That is the whole trade in one sentence: an LST converts the network’s time risk into price risk, and in a crisis you choose which one to eat. Staked-ether ETFs, which have multiplied in 2026, sit on top of the same plumbing and share the same queue, so the institutional wrapper does not escape the constraint either.

The scale of today’s queue makes the point concrete. Ethereum caps how much stake can enter or leave in each short slot, so the exit line does not clear faster just because more people join it; it drains at a fixed rate, which is why about 967,000 ETH translates into more than two weeks rather than a few hours. A holder with a few ETH can usually redeem out of an LST buffer and never see the queue at all. A treasury or an ETF trying to unwind tens of thousands of ETH in a falling market is the one that finds out, in real time, whether the buffer is deep enough or whether it is selling stETH into Curve at a discount. The protocols cannot change that arithmetic; they only decide who reaches it first.

This is why depth matters more than a headline yield difference of a few basis points. The protocol most likely to honor a one-for-one exit in a stampede is the one with the largest buffer and the deepest secondary pool, and on both counts that is Lido, with Rocket Pool and Frax progressively more exposed to the queue.

When the receipt trades below the asset

The textbook case is June 2022. Before Ethereum allowed any staking withdrawals, the only exit was the secondary market, so when Terra’s collapse and the failures of Celsius and Three Arrows Capital forced mass selling, stETH slid to about 0.94 ETH, a discount of roughly 6%, with a Three Arrows wallet reported converting tens of millions of dollars of stETH at a 5.6% to 5.9% discount, CoinDesk noted at the time. There was no redemption backstop, so arbitrage could not close the gap and the discount simply widened.

Withdrawals changed that. Since Shapella, anyone can buy a discounted stETH and redeem it for a full ETH, which drags the secondary price back toward one-for-one and is why stETH now trades within a fraction of a percent of ETH in normal conditions (StakingRewards). The arbitrage is not instant, though, because the redeemer still has to sit in the exit queue to collect, so in a fast crash a discount can open for days before it closes, long enough to punish anyone forced to sell at the bottom. The redemption queue is the peg’s anchor and its speed limit at the same time; without it, an LST is just an IOU trading on sentiment.

Peg risk has not vanished, it has moved. Because lending markets price LSTs through oracles, a glitch in that pricing can do damage even when the token itself is fine. In March 2026 a capped-oracle update on Aave briefly undervalued wstETH by about 2.85%, triggering roughly $27 million of liquidations across 34 positions, though the protocol took no bad debt, CoinDesk reported. The deeper and more widely integrated an LST, the more of this plumbing leans on it, which is also why LSTs used as cross-chain collateral carry the risks we mapped in our look at bridge security in 2026. For reward-bearing tokens like rETH and sfrxETH, remember the premium is yield, not a peg; the real question is how far their thinner pools would gap in a genuine rush.

Yield, fees, and what you keep

Exit speed is one axis; the actual return is another, and here the three cluster closely because they all draw from the same base rate. Ethereum’s consensus layer pays about 2.58% before tips and MEV (validatorqueue.com); each protocol then skims a fee and passes the rest to holders. The table shows the approximate picture.

Net rates move with network conditions and are approximate; MEV and priority tips can add roughly half a point to a point on top (StakingRewards). Frax’s number looks higher mainly because sfrxETH concentrates the yield that non-staking frxETH holders give up into the vault, not because its validators outperform.

Put any of those next to cash and the comparison is unflattering. The 3-month US Treasury bill yielded 4.06% on 9 October 2026, with the 1-year at 4.47%, per the Federal Reserve. In plain dollars, ETH staking does not clear the risk-free rate; you are accepting a lower nominal yield in exchange for exposure to ETH, not for income. That gap also caps how hard the protocols can compete on yield: with only about 2.58% of base reward to split, the fee contest between them is a fight over tenths of a percent, which is a big part of why their real differentiation has shifted to exactly the liquidity and exit questions this piece is built around.

The yield is thin; the position is still ETH

That last point is the one most easily lost in a protocol-versus-protocol fee comparison. A spread of 20 or 30 basis points between Lido and Rocket Pool is real, but it is a rounding error next to ETH’s price swings. ETH is down about 50% from its 2025 high (CoinGecko); a 2.3% annual yield does not begin to offset a move of that size in either direction.

So the honest framing is that staking through any of these protocols is a leveraged bet on ETH with a small coupon attached, not a fixed-income product. The decision that dominates your outcome is whether you want to hold ETH at all; the choice of LST only decides how efficiently you hold it and how cleanly you can leave. Staking does tilt the odds slightly versus simply holding spot ETH, since you collect the coupon either way, but it layers on smart-contract and exit risk that spot ETH does not carry, and in the worst moments those risks and the price drawdown tend to arrive together. Readers trying to form a view on the underlying asset can start with our survey of 2026 price targets and the models behind them, which brings the same scrutiny to crypto valuation that this piece brings to staking mechanics.

Governance tokens: LDO, RPL and FRAX

Each protocol has a token that is separate from the staking receipt, and each has been a poor proxy for the protocol’s usage. Lido’s LDO trades near $0.41 for a market value around $340 million, about 94% below its 2021 high; Rocket Pool’s RPL sits near $1.69 for roughly $39 million, down about 97% from its peak, per CoinGecko (LDO, RPL). Frax’s FRAX is the former FXS, now repurposed as Fraxtal’s gas token rather than a staking-governance asset.

The gap between usage and token price comes down to value capture. LDO is a governance vote over a treasury, not a claim on stETH’s fee stream, so billions in staked ETH do not translate into cash flow for the token. RPL has a more direct role, acting as operator collateral that backstops the protocol, and Saturn’s revenue split now routes a slice of protocol income to staked RPL, paid in ETH rather than inflation. FRAX has simply been redefined around the stablecoin and chain business, so its price now tracks that ambition rather than frxETH flows. The question of whether an infrastructure token can actually capture the value of the network it secures is not unique to staking; we worked through the same tension for compute in Akash’s case, and the lesson transfers: usage and token value are different variables, and here they have diverged hard. For a holder the takeaway is simple: owning the receipt (stETH, rETH or sfrxETH) is a claim on staked ETH, while owning the governance token is a separate, and so far much worse, bet on the business built around it.

Decentralization, concentration and the SEC

Lido’s scale is also its biggest controversy. When a single protocol controls more than a fifth of all staked ETH, Ethereum researchers worry about the thresholds at which one actor could degrade the network: roughly a third of stake can delay finality, and two-thirds could finalize a bad chain. Vitalik Buterin has repeatedly flagged the concentration of staking in a single dominant provider as a systemic risk for Ethereum, a concern The Block has tracked closely. Lido’s answer is dual governance, a mechanism that lets stETH holders delay LDO decisions they object to and, in the extreme, escape them entirely by exiting to ETH before a contested change takes effect, which hands even non-token-holding stakers a brake on the DAO. Its validator set, while curated rather than open, is spread across many professional operators now consolidating into a leaner, larger group. It is worth noting the network’s total validator count has fallen to about 837,000 from roughly 888,000 a few weeks earlier (validatorqueue.com), a direct result of the consolidation trend Lido is leading; fewer, larger validators is efficient, but it is a different decentralization profile than the one Ethereum started with.

For a US reader the regulatory picture is, for once, relatively settled. In an August 2025 statement, the SEC’s Division of Corporation Finance said that “Liquid Staking Activities…do not involve the offer and sale of securities within the meaning of Section 2(a)(1) of the Securities Act of 1933,” provided the provider stays administrative or ministerial. The guardrails are specific: the provider must not decide “whether, when, or how much” of a depositor’s assets to stake, and must not “guarantee or otherwise set the amount” of the rewards. A protocol that merely takes custody and routes deposits to node operators is fine; one that exercises discretion or promises a fixed return falls outside the safe harbor. That line, custody and routing versus discretion and guarantees, is the one every staking product now draws itself around, and it is why the exit-risk details above also shape the legal picture, as CryptoSlate has noted.

Which one fits you?

If your priority is the cleanest possible exit, the most DeFi integrations and the deepest secondary market, Lido is the default, with the caveat that you are adding to the concentration Ethereum researchers worry about and trusting a curated set that is consolidating toward institutional operators. If credible neutrality matters more to you than depth, Rocket Pool’s permissionless operator base is the strongest in the category, and you accept a thinner market and a slower worst-case exit in return. If you are already building inside Frax’s stablecoin and Fraxtal ecosystem, its two-token model gives you a clean split between a liquid leg (frxETH) and a yield leg (sfrxETH), but you are making a bet on Frax’s broader direction as much as on its staking product.

Whichever receipt you hold, two risks sit underneath all three: the ETH price itself, which dwarfs any yield difference, and the security of your own keys, because a liquid staking token in a compromised wallet is as gone as any other asset. If you are holding size, pair any of these protocols with serious self-custody; our hardware wallet comparison covers the tradeoffs. No protocol removes ETH’s volatility, and none can shortcut the exit queue when everyone heads for it at once; they only decide how gracefully you wait.

Frequently Asked Questions

Can I unstake from Lido, Rocket Pool or Frax instantly?

Not always at full value. You have two exits: sell the token on a market like Curve for immediate ETH at whatever price the pool quotes, or redeem it through the protocol for the underlying ETH one-for-one. Redemption can be quick when the protocol’s buffer holds enough ETH, but if many people withdraw at once it falls back on Ethereum’s validator exit queue, which was about 17 days on 14 October 2026.

Why does rETH cost more than stETH and ETH?

Because rETH is a reward-bearing token. Instead of increasing your balance, it lets each unit appreciate against ETH, so one rETH is now worth about 1.1747 ETH. That premium is accumulated staking reward, not a mispricing. Lido’s stETH rebases instead, staying near one-for-one while your token count grows.

Is stETH safe after the 2022 depeg?

The 2022 discount happened because Ethereum did not yet allow withdrawals, so stETH could only be sold, not redeemed. Since the 2023 Shapella upgrade, holders can redeem stETH for ETH one-for-one, which anchors the price and is why it now trades within a fraction of a percent of ETH. The residual risks are smart-contract bugs, oracle glitches and a crisis deep enough to drain the withdrawal buffer.

Which has the best yield: Lido, Rocket Pool or Frax?

They are close, because all three draw on the same base rate near 2.58%. After fees, holders net roughly 2.0% to 2.8% depending on the protocol and conditions, with Frax’s sfrxETH often looking highest because it concentrates yield into the vault. The differences are small next to ETH’s price moves, and none of them clears the roughly 4% US Treasury bill rate in dollar terms.

Are liquid staking tokens securities in the US?

In an August 2025 statement, the SEC’s Division of Corporation Finance said liquid staking activities do not involve the offer and sale of securities, as long as the provider stays administrative or ministerial: it must not decide whether, when or how much to stake, and must not guarantee the rewards. Providers that exercise discretion or promise a fixed return fall outside that safe harbor.

By Yuki Tanaka, HOGE Wire mining and staking desk.