Lido vs Rocket Pool vs Frax: The Bond Behind Your Yield

Buying stETH, rETH, or sfrxETH feels like moving ether into a savings account. You hand over ETH, a token comes back, and the balance grows on its own. What that smooth experience hides is that you no longer hold ether at all. You hold a claim on validators that other people operate, secured by rules you did not write and by capital you cannot see on your own screen. The yield is real. So is the dependency.

The three largest Ethereum liquid staking protocols, Lido, Rocket Pool, and Frax, answer the same uncomfortable question in three different ways: when the people running your validators fail, drop offline, double-sign, or get slashed, whose money absorbs the loss first, and how much of their own capital is on the line? For most of liquid staking history that question sat buried under marketing about annual percentage rates and composability. In 2026 it moved to the center, because all three protocols spent the year rebuilding their validator back-ends around posted bonds, collateral, and, in Frax’s case, an outright lending market where the validator is the borrower.

This comparison works from that axis outward. We start with what a liquid staking token actually is, then look hard at the collateral behind each receipt, and only then turn to the things most guides lead with: yield after fees, decentralization, liquidity, the governance tokens, and what United States regulators now say about the whole arrangement. All figures are current as of 27 October 2026, and every number carries a source you can check.

What a liquid staking token actually is

Ethereum secures itself through proof of stake. To run a validator directly you need 32 ETH, technical skill, and a machine that stays online around the clock; mess up the last part and the network docks your balance. That barrier is the reason liquid staking exists. Instead of running a validator, you deposit ETH into a protocol, the protocol stakes it through validators run by professionals or by a permissionless operator set, and you receive a token that represents your deposit plus its accruing rewards. The token is transferable and usable across decentralized finance, which is the liquid part. Ethereum itself still shows a long activation line for new stake: the entry queue held about 1.3 million ETH on 27 October, a wait of roughly 23 days, according to validatorqueue.com.

There are two ways the token can pay you, and the difference matters more than most newcomers expect. A rebasing token like Lido’s stETH keeps a price near one ETH and grows your balance: hold 10 stETH today, hold a little more tomorrow. A reward-bearing token like Rocket Pool’s rETH, Lido’s wstETH, or Frax’s sfrxETH keeps your balance fixed and lets the exchange rate climb: one rETH was worth about 1.1783 ETH on 27 October per CoinGecko, and that number only rises as rewards compound. The mechanical difference drives a tax difference in the United States, where a growing balance can look like income received on a rolling basis while an appreciating exchange rate tends to defer the event until you sell. We will return to that, and it is covered in depth in our guide to crypto staking taxes in 2026.

The receipt is also reusable. The same stETH or rETH can be posted as collateral to borrow against, dropped into a liquidity pool, or re-pledged into restaking systems that rent Ethereum’s trust to other networks, a layered dependency we unpack in our explainer on restaking. Every one of those uses stacks a new counterparty on top of the validator risk you started with.

Three philosophies, in one sentence each

Lido optimizes for scale and liquidity: a huge pool of ETH handed to a vetted operator set, now opening to permissionless operators, with stETH the deepest and most widely integrated receipt in crypto. Rocket Pool optimizes for trust minimization: a permissionless operator set where every operator posts real ETH as first-loss capital, so rETH is backed by people with their own money at stake. Frax optimizes for capital efficiency and treats staking as one module inside a stablecoin system, culminating in a 2026 redesign that turns validator funding into an explicit on-chain loan.

Those are not marketing slogans; they fall straight out of how each protocol capitalizes the people who run its validators. That is where the real comparison lives, so that is where we go next.

Lido: the giant that finally made its operators post collateral

Lido is the market by any measure that counts. Its total value locked was about 22.4 billion dollars on 26 October, roughly 42.5 percent of the entire liquid staking category tracked by DefiLlama, which itself totals around 52.8 billion dollars across more than 260 protocols. stETH alone carried a market capitalization near 22.57 billion dollars at a price of 2,331.85 dollars, ranked tenth among all crypto assets per CoinGecko; its wrapped sibling wstETH added another 10.78 billion dollars at 2,908.28 dollars, each unit worth 1.2510 ETH.

For most of Lido’s history, its validators were run by a curated set of professional operators who posted no bond at all. They were chosen for reputation and track record, and if one of them was slashed, the penalty did not come out of the operator’s pocket; it flowed straight into the stETH rebase, trimming every holder’s balance by a sliver, with the Lido DAO treasury as a discretionary backstop it was never contractually obliged to open. That design scaled beautifully and concentrated an uncomfortable amount of Ethereum behind a permissioned list.

In 2026 Lido inverted the model in two ways. First, the Curated Module v2 now requires even its brand-name operators to post locked ETH bonds for the first time, part of the broader consolidation described by The Block. Second, the Community Staking Module lets anyone run a Lido validator permissionlessly by posting a bond, 2.4 ETH for a first validator and 1.3 ETH for each additional key, while Lido supplies the 32 ETH of stake. A larger 0x02 variant launched in late October with a 32 ETH operator bond for validators that can hold up to 2,048 ETH, reported by CryptoSlate. In the Community Staking Module, that bond is burned to cover penalties before any loss reaches stETH holders. The giant finally gave its savers a first-loss layer.

The consolidation underneath is dramatic in its own right. Using the 2,048-ETH validators that Ethereum’s Pectra upgrade enabled in May 2025, Lido is collapsing roughly 265,000 legacy validators into about 4,000 larger ones. Network-wide, staked ETH running on these 0x02 validators crossed 40 percent for the first time in late October, with 1.8 million ETH migrated on Lido Core in a single month, per Crypto Briefing. The validator count has already begun to fall: validatorqueue.com showed 817,035 active validators on 27 October, down from roughly 888,000 in late September, a visible fingerprint of the cleanup.

Lido also carries the most developed safety valve on governance. Its dual governance system, live since mid-2025, lets stETH holders who disagree with an LDO vote escrow their tokens to delay a proposal once they cross about one percent of supply, and trigger a rage-quit escape hatch past roughly ten percent. It is a direct answer to the charge that token-holder governance can override the savers whose ETH is actually at risk.

Rocket Pool: the 4 ETH bond as first-loss capital

Rocket Pool is a fraction of Lido’s size and proud of what that fraction buys. Its rETH carried a market capitalization near 846 million dollars across about 309,000 tokens on 27 October, per CoinGecko, and the network ran on roughly 1,500 independent node operators by mid-2026. There is no application form and no approved list. Anyone can run a validator, and that permissionless operator set is the whole point.

What keeps that openness honest is money. Every Rocket Pool operator posts a bond that sits in first-loss position ahead of rETH holders. The Saturn One upgrade, live on 18 February 2026, cut that bond from 8 ETH to 4 ETH per validator and introduced megapools: an operator puts up 4 ETH, the protocol tops it up with 28 ETH from the rETH deposit pool, and the pair runs a 32 ETH validator. If that validator is slashed, the operator’s 4 ETH is burned before rETH takes a scratch. That is why rETH has tracked its exchange rate for years with almost none of the drama liquid staking tokens are supposed to invite.

Saturn One also rewired the economics of RPL, the protocol’s secondary token. Under the new Universal Adjustable Revenue Split, every operator earns a base commission of about 5 percent, and operators who also stake RPL earn a bonus share of protocol revenue, up to around 9 percent, paid in ETH rather than in freshly minted RPL. The change flips RPL from an inflationary incentive into a claim on real cash flow, and the remaining RPL inflation is being phased out later in 2026, as The Defiant details. RPL itself tells the story of a token that lost its old job: it traded at 1.47 dollars on 27 October, down about 97.6 percent from its 2021 peak of 61.90 dollars per CoinGecko.

Frax: when the validator becomes a borrower

Frax approaches staking as an engineering problem about capital efficiency rather than as a product in its own right. Its first design split the job across two tokens: frxETH, a liquid token that deliberately pays no yield, and sfrxETH, an ERC-4626 vault that concentrates all of the staking yield into the holders who deposit into it. The trick is that frxETH holders who never stake into the vault forfeit their share, so sfrxETH can print an effective yield above the raw network rate. On 27 October frxETH traded at 2,328.69 dollars with a market cap near 140 million dollars, while sfrxETH sat at 2,733.61 dollars, about 1.1759 ETH, per CoinGecko. Both are an order of magnitude smaller than rETH, let alone stETH.

The 2026 story is frxETH v2, which reframes staking as a lending market and shipped to a capped mainnet in mid-October. Instead of handing stake to operators, Frax lets a node operator deposit ETH into a ValidatorPool contract under the operator’s own control and borrow additional ETH from the protocol to fund validators. The collateral is not RPL or a cash bond; it is an escrowed exit message, a pre-signed instruction the protocol can broadcast to force the validator out of Ethereum if the loan turns unhealthy. An off-chain health bot watches every position and triggers that exit as a liquidation, and the interest rate rises automatically when ETH gets scarce, according to the Frax documentation. Leverage runs from roughly 4x for anonymous operators to as much as 8x for whitelisted community operators approved by governance. The initial mainnet is capped at about ten validators, mostly inside the Frax ecosystem, with a wider opening planned if the system stays stable for two to three weeks.

Why so small, when Lido counts its ETH in the millions? Because staking is no longer the main event at Frax. Under the North Star rebrand the project renamed its FXS token to FRAX and made FRAX the gas token of its Fraxtal network, repositioning the whole operation as, in founder Sam Kazemian’s words, “a scalable stablecoin operating system, not just a single protocol,” as documented by 4pillars research. Liquid staking became a supporting act for a stablecoin business. frxETH v2 matters here less for its size than for what it makes visible: the thing a liquid staking token has quietly been all along is a loan against validators.

The architecture, side by side

The table below strips each protocol down to the mechanics that decide your risk and your yield. Notice how the operator-capital column, the one most guides skip, maps almost perfectly onto the trust model each protocol advertises.

Whose money is the first-loss layer?

Here is the question to carry into any liquid staking decision: a validator you rely on gets slashed, say for double-signing after a botched migration. Where does the loss land, and in what order? The three protocols now give three genuinely different answers.

At Rocket Pool the order is cleanest. The operator’s 4 ETH bond burns first, then any RPL they staked, and only a penalty large enough to exhaust both would reach rETH. Because a solo slashing event after Pectra is small, a single operator failure is effectively walled off from rETH holders. At Frax v2 the structure is similar in spirit: the operator’s deposited ETH is the first-loss slice, and the escrowed exit message lets the protocol pull the validator and recover before the shortfall spreads, with a liquidation bot standing by. At Lido the answer now depends on which module your stake sits in. Community Staking Module and Curated Module v2 validators carry operator bonds that absorb the first hit; older curated stake without a fresh bond still socializes the penalty across the stETH rebase, which is why the 2026 bond rollout is more than cosmetic.

Skin in the game per validator is therefore a real, comparable number. Rocket Pool asks the most of its operators relative to the stake they control, Frax formalizes the same idea as a loan-to-value ratio, and Lido, historically the thinnest on this metric, spent 2026 catching up. None of this touches the other first-loss question, the smart contract itself. Every one of these protocols is a stack of code holding billions of dollars, and a bug in the staking contract, the withdrawal logic, or an oracle can vaporize funds no bond was ever sized to cover. That is why audit depth belongs in the comparison, a discipline we profile through the toolchain built by firms like Trail of Bits. A real 2026 reminder: on 10 March an oracle pricing glitch on Aave briefly undervalued wstETH by about 2.85 percent and triggered roughly 27 million dollars of liquidations across 34 positions, though with no bad debt, as CoinDesk reported. The code around the receipt can hurt you even when the validators behave.

The yield, after everyone takes their cut

Staking yield in late 2026 is thin and getting thinner as more ETH competes for the same issuance. validatorqueue.com put the base consensus APR at about 2.47 percent on 27 October, with 44.1 million ETH staked, 36.12 percent of supply. On top of that sits a variable slice of priority fees and MEV, which lifts an efficient validator toward the 3 percent range before fees. Then the protocol takes its cut, and what reaches your receipt is smaller still.

Lido takes 10 percent of staking rewards, split between its operators and the DAO treasury, leaving stETH holders a net yield in the low 2 percent range. Rocket Pool historically took a node-operator commission near 14 percent, now restructured under UARS into a roughly 5 percent base plus the RPL-backed bonus, which leaves rETH holders fractionally below stETH on net. Frax’s sfrxETH can print a touch higher because of the two-token forfeiture trick, not because its validators earn more. These are differences of tenths of a percent, which is the real headline: when the base rate is barely above 2 percent, fees, tax treatment, and risk matter more than the advertised APR.

The United States tax angle sharpens the point. A rebasing token hands you new units continuously, and new units received tend to be income at their fair market value when they land, generating a running stream of taxable events even if you never sell. A reward-bearing token hands you nothing new; your single position simply appreciates, and the tax event tends to wait for disposal. For a US holder in a taxable account, that can make wstETH, rETH, or sfrxETH materially simpler than raw stETH, a nuance our staking tax guide works through, with the caveat that wrapping itself can be a taxable disposal. None of this is tax advice; it is a reason the token-mechanics column above is not a footnote.

Decentralization and the one-third problem

Ethereum has hard thresholds baked into its consensus. A single entity or coordinated bloc controlling more than one third of stake can stall finality; past one half it can censor; past two thirds it can finalize a dishonest chain. Lido sits close enough to the first line to make people nervous. Its share of all staked ETH ended recent months near 21 percent, down from about 24 percent in January, according to CCN, but within the liquid staking segment specifically stETH is still around 62 percent. Back in 2023 the Lido DAO voted down a proposal to self-limit its share, a decision critics have never let go.

Vitalik Buterin has long treated staking concentration as one of the more serious risks to Ethereum, warning about the pull of a single dominant staking token in commentary summarized by The Block. His preferred cure runs straight through the collateral theme of this piece: make it cheaper to run a validator so more people do. “The #1 thing in becoming more home staking friendly is to reduce the 32 ETH requirement,” he told Decrypt. That is precisely what lower bonds accomplish: Rocket Pool’s 4 ETH and the Lido Community Staking Module’s 2.4 ETH bond both widen the operator set, and Lido estimates its module gives a community operator a capital multiplier of roughly 2.2 times versus solo staking, per lido.fi. More operators with real bonds is a better decentralization story than a short approved list, which is why the 2026 shift toward permissionless, bonded operators matters beyond any single protocol’s marketing.

Consolidation complicates the picture. Packing stake into fewer, larger 0x02 validators is good for Ethereum’s bandwidth and for operator economics, but it also means fewer distinct validators stand between the network and the one-third line. That is a trade Ethereum chose deliberately; it is worth knowing you are part of it when you hold any of these tokens.

Liquidity, the peg, and the two meanings of liquid

The word liquid hides two very different exits. One is selling your receipt on the open market, which is instant but pays whatever the market offers, sometimes below fair value. The other is redeeming through the protocol for the full underlying ETH, which pays full value but moves at the pace of Ethereum’s exit queue. stETH enjoys the deepest secondary market in the sector, anchored by a vast Curve pool whose design we cover in our look at AMM design in 2026; rETH and the Frax tokens trade thinner, so a large holder rushing the door pays more slippage.

The peg is not guaranteed. In June 2022, before Ethereum even allowed withdrawals, stETH slid to about 0.94 ETH on the secondary market as Celsius and Three Arrows Capital unwound huge positions into shallow liquidity, an episode CoinDesk and Nansen dissected at the time. Redemptions now exist, which is the structural fix, but they are still rate-limited by the validator exit queue. That queue swung violently in early October 2026 when a security scare pushed thousands of validators to exit at once; by 27 October it had drained back to a near-empty 798 ETH and a wait of barely a day, per validatorqueue.com. The lesson holds: in calm markets the receipt trades at par, and in a panic the discount and the queue both reappear exactly when you want out.

LDO, RPL, FRAX: what the tokens actually capture

Do not confuse the receipt token, which holds your ETH, with the governance token, which holds a bet on the business. All three governance tokens have had a brutal few years, and each now captures value differently.

LDO is a pure governance claim on the largest pool in staking, which sounds valuable until you remember it earns nothing directly and now operates under a dual-governance regime that lets stETH holders overrule it. It traded at 0.3686 dollars on 27 October, down about 94.9 percent from its 2021 high, per CoinGecko. RPL spent years as an inflationary collateral token and has been refashioned into something closer to a revenue share, but the market has not rewarded the pivot yet: 1.47 dollars, down about 97.6 percent from its peak. FRAX, formerly FXS, walked away from the staking narrative entirely to become the center of a stablecoin system, which is either a shrewd escape from a low-margin business or an admission that liquid staking alone could not carry a token; reasonable people read it both ways.

What the SEC actually said about liquid staking

United States regulation of liquid staking turned a sharp corner in 2025. On 5 August the Securities and Exchange Commission’s Division of Corporation Finance published a staff statement declaring that, in its view, “Liquid Staking Activities…do not involve the offer and sale of securities,” a position you can read in full on sec.gov. For a sector that spent years bracing for an enforcement case, that sentence was enormous.

The guardrails are the interesting part, because they read like a description of the collateral models above. The relief applies only where the provider stays, in the staff’s words, “administrative or ministerial in nature” rather than “entrepreneurial or managerial.” The provider cannot “decide whether, when, or how much of a Depositor’s Covered Crypto Assets to stake,” and cannot “guarantee or otherwise set the amount of rewards owed to the Depositors.” In plain terms, a protocol that merely routes your ETH to validators and passes through whatever the network pays looks like a receipt for deposited goods; one that promises a fixed yield or exercises discretion over your stake starts to look like a security. That test maps neatly onto the three designs here, and it is the same control-and-chokepoint logic we trace through our guide to DeFi compliance in 2026. It also helped clear the runway for the spot staking ETFs that US issuers brought to market earlier in 2026.

The 2026 convergence: everyone became a lender

Step back from the logos and a single pattern defines the year. Each protocol turned its node operator into a borrower who posts collateral. Frax made it literal, with a lending market, an interest rate, and a liquidation bot. Rocket Pool has effectively run a credit line for years: the operator posts 4 ETH and borrows the other 28 from the rETH pool, with the bond as the lender’s protection. Lido arrived last and from the opposite direction, bolting bonds onto a model that never had them, so that a Community Staking Module operator now borrows Lido stake against a posted deposit much as a Rocket Pool operator borrows from the deposit pool. Three philosophies, converging on one balance-sheet shape.

The convergence is good for savers, mostly. First-loss capital sitting under your receipt is strictly better than a reputation and a discretionary treasury. But a credit market carries its own failure modes: cascading liquidations if ETH gaps down against borrowed positions, interest-rate spikes that squeeze operators out exactly when the network is stressed, and the same reflexive dynamics that have bitten every other corner of decentralized finance. The receipt that felt like a savings account is, at its base, a leveraged claim on validators funded by a loan. That is not a reason to avoid it. It is a reason to know which loan you are holding.

How to choose among the three

There is no single winner, only fits. The honest framing is to match the protocol to what you actually value, then accept the trade that comes with it.

• You want maximum liquidity and the widest DeFi integration: Lido’s stETH or wstETH. You accept that you are adding to the most concentrated pool in staking, mitigated somewhat by dual governance and the new operator bonds.
• You care most about trust minimization and first-loss backing: Rocket Pool’s rETH. You accept a thinner secondary market and a smaller ecosystem in exchange for permissionless operators with real money at stake.
• You want a reward-bearing token and a tidy US tax surface: wstETH, rETH, or sfrxETH all defer the taxable event to disposal, unlike rebasing stETH, though the act of wrapping can itself be a disposal.
• You are chasing a marginal yield edge: sfrxETH, with eyes open to its small size, thin liquidity, and the fact that frxETH v2 is still a capped, young system.
• You would rather not hold any single protocol’s risk: split across two receipts, or consider a regulated staking ETF now that the SEC has blessed the activity, accepting a management fee in exchange for not touching a private key.

Whatever you choose, hold on to the one fact this whole comparison turns on: you are not holding ETH. You are holding someone’s promise, backed by someone’s collateral, to give you ETH back. In 2026 that promise finally comes with a bond behind it. Read the bond.

Frequently Asked Questions

Is Lido, Rocket Pool, or Frax the best for staking ETH in 2026?

There is no single best; it depends on your priorities. Lido offers the deepest liquidity and widest DeFi integration but the most concentration, Rocket Pool offers permissionless operators with the most first-loss collateral behind rETH, and Frax offers a capital-efficient reward-bearing sfrxETH inside a stablecoin system but at a far smaller size. Match the protocol to whether you value liquidity, trust minimization, or a tidy tax surface.

Is liquid staking a security under US law?

On 5 August 2025 the SEC Division of Corporation Finance stated that liquid staking activities do not involve the offer and sale of securities, provided the staking provider stays administrative or ministerial and does not decide how much of your assets to stake or guarantee the rewards. A protocol that promises a fixed yield or exercises discretion over your stake could fall outside that relief.

What is the difference between stETH, rETH, and sfrxETH?

stETH is rebasing: its price stays near one ETH and your balance grows daily. rETH and sfrxETH are reward-bearing: your balance stays fixed and the exchange rate climbs, about 1.1783 ETH per rETH and 1.1759 ETH per sfrxETH on 27 October 2026. The reward-bearing design usually defers US taxes until you sell, while a rebasing balance can create taxable income as rewards arrive.

What staking yield do Lido, Rocket Pool, and Frax pay in 2026?

All three ride the same base Ethereum staking rate, about 2.47 percent in late October 2026, plus a variable slice of MEV. After fees, stETH and sfrxETH net in the low-2-percent range and rETH sits fractionally below, so the differences are tenths of a percent. The thin spread is why fees, taxes, and risk matter more than the headline APR.

Is my ETH safe in a liquid staking protocol?

Your ETH faces three stacked risks: validator slashing, now cushioned by operator bonds at all three protocols; smart-contract or oracle failure, as when a March 2026 Aave glitch liquidated about 27 million dollars of wstETH positions; and secondary-market depeg risk in a panic, as when stETH fell to about 0.94 ETH in 2022. Redemptions reduce but do not remove the peg risk, because the validator exit queue rate-limits withdrawals.

By Yuki Tanaka, Mining and Staking desk, HOGE Wire.