Lido vs Rocket Pool vs Frax: The 2026 Rebuild

For most of 2026 the contest between Ethereum’s three best-known liquid staking protocols looked settled. Lido was the giant, Rocket Pool was the decentralization favorite, and Frax was the clever niche. That picture is now out of date. Over the past year each protocol tore up a large part of its own design and rebuilt it, and the single biggest change is happening this month. Lido has begun physically consolidating hundreds of thousands of validators, a migration set to shrink Ethereum’s total validator set from roughly 880,000 to about 628,000, close to a 29% cut in the number of attesting validators, according to Cryptonews, with the first consolidations scheduled for September 2026.

So the useful question right now is not only which token pays the most. It is what each of these protocols became after its rebuild, and which version fits the way you actually hold and use ETH. This piece compares Lido, Rocket Pool and Frax as they stand in September 2026: their size, their yields, how you get your ETH back, how decentralized each one really is, and the US regulatory frame around all of it.

A quick baseline. ETH trades near $2,509, about 49% below its August 2025 record, per CoinGecko. Around 43.0 million ETH, or 35.21% of supply, is staked, and the base validator reward has drifted down to 2.59%, according to validatorqueue.com. That last number matters more than any single protocol feature, because it is the raw material all three are packaging.

What a liquid staking token actually is

Staking ETH the classic way means locking 32 ETH into a validator and joining an entry queue that currently runs past 32 days for new deposits (validatorqueue). Your capital is productive but frozen. A liquid staking token solves that: you deposit ETH with a protocol, the protocol runs the validators, and you receive a transferable token that stands for your stake plus its accruing rewards. You can sell it, lend it, or post it as collateral without ever unstaking.

There are two ways that token tracks its rewards, and the difference runs through everything below. A rebasing token, like Lido’s stETH, keeps a price close to 1 ETH and grows the number of tokens in your wallet each day. An exchange-rate token, like Rocket Pool’s rETH, Frax’s sfrxETH, or Lido’s own wrapped wstETH, keeps the token count fixed and lets each token become worth steadily more ETH over time. That is why one rETH costs more than one ETH: it is not a broken peg, it is accumulated yield. Keep that distinction in mind, because it decides how each token behaves in DeFi and how it is taxed.

The scoreboard, and the gap that will not close

The first thing an honest comparison has to admit is the scale gap. Lido’s stETH carries a market value of about $24.24 billion across roughly 9.66 million tokens, per CoinGecko. Rocket Pool’s rETH sits near $929 million (CoinGecko), and Frax’s two ETH tokens together are worth around $262 million. In round terms, Lido’s staked ETH is worth about 26 times Rocket Pool’s and more than 90 times Frax’s. On the network’s own denominator, Lido stakes in the low-20s percent of all staked ETH, while Rocket Pool and Frax are each well under 1%.

The more revealing number is the direction. Lido’s share of all staked ETH slipped from 23.93% at the start of 2026 to 21.18% by the end of June, and it captured only 5.7% of the net new stake added in the first half of the year, per CryptoSlate. The network is growing; Lido is growing slower than the network. Most of the fresh stake is flowing to institutions and custodians that never touch Lido’s contracts. That is the backdrop against which Lido rebuilt itself.

One caveat on that table: wstETH is wrapped stETH, the same underlying Lido stake in a non-rebasing form, so its market cap is not additional stake on top of stETH. Figures are from CoinGecko on 11 September 2026 and move with the ETH price.

Lido after Core 2026: fewer validators, more scrutiny

Lido’s rebuild is called Core 2026, and it passed its final governance vote on 24 July 2026. Two things are happening at once. First, Lido is consolidating. Thanks to the higher validator ceiling introduced by the Pectra upgrade, it can migrate more than 265,000 existing validators onto compounding 0x02 credentials, lifting the share of ETH secured by compounding validators from 32.06% to 52.21% and cutting its own validator count by roughly a third (Cryptonews). Fewer validators doing the same work means lower infrastructure cost and lighter load on Ethereum’s consensus layer. The migration began in late July with about $16.5 billion in staked ETH, and the first live consolidations are set for this month, per CoinDesk.

Second, Lido is tightening who runs those validators. Its Curated Module v2 requires the roughly three dozen professional operators to post ETH bonds for the first time, turning reputation into money that is actually at stake. And on 1 September the DAO backed a new Validator Operations Standard, ValOS, a shared framework for documenting and independently reviewing how operators handle keys, infrastructure and incident response; Lido earmarked a small ecosystem grant, around $60,000, to subsidize the first assurance reviews, per The Crypto Times. ValOS does not change yields or requirements. It is a direct answer to a specific criticism: that a handful of infrastructure providers now run enormous amounts of stake.

That criticism is also why Lido runs dual governance, live since mid-2025, which lets stETH holders veto, or in the extreme exit ahead of, a DAO decision they oppose (Lido). Put together, Core 2026 is Lido trying to stay the default choice for size and liquidity while defusing the concentration argument that has trailed it for years. Whether it works is the open question, since its market share kept sliding through the rebuild.

Rocket Pool after Saturn: the permissionless bet at four ETH

Rocket Pool made the opposite bet, and in 2026 it doubled down. Its Saturn One upgrade, live since 18 February, cut the node-operator bond from 8 ETH to 4 ETH per validator, with the remaining 28 ETH supplied by the liquid stakers who hold rETH (Rocket Pool). A lower bond means more people can afford to run a Rocket Pool validator, which is the entire point: Rocket Pool is the only major liquid staking protocol with a fully permissionless operator set, around 2,000 independent operators rather than a curated list, according to Web3WAGMI.

Saturn also introduced megapools, which let a single contract act as the withdrawal address for many validators and sharply cut the gas cost of running them, and it reworked RPL, the protocol’s token. RPL is no longer required to launch a validator; instead, operators who stake RPL now earn a share of protocol revenue paid in ETH through a fee switch, rather than being diluted by RPL emissions (Crypto Briefing). Megapool validators can earn meaningfully more commission per bonded ETH than the old design allowed.

The cost of that decentralization shows up in two places. Size: rETH’s market cap is about $929 million (CoinGecko), a fraction of Lido’s. And yield: rETH has recently paid around 2%, a touch below stETH, partly because a smaller, more distributed operator set captures MEV less aggressively than Lido’s professional shops. RPL itself trades at $1.70 with a market cap under $40 million, per CoinGecko, a reminder that a governance token’s price and the protocol’s usefulness are only loosely related.

Frax after North Star: staking as one Lego brick

Frax never tried to win on size or on decentralization. Its Frax Ether product is deliberately a two-token machine. frxETH is meant to trade at roughly 1 ETH and pays no yield on its own; sfrxETH is an ERC-4626 vault that collects the staking rewards of all the underlying validators (Frax docs). Because holders who keep plain frxETH give up their share, the yield concentrates onto sfrxETH, which is why Frax’s staked token often shows the highest headline rate of the three. It is a design optimized for DeFi plumbing, not for being the biggest.

In 2026 Frax pushed further with frxETH v2, which turns validator operation into a lending market: node operators post ETH as collateral and borrow additional ETH from the pool to run validators, with the interest flowing back to sfrxETH holders (Frax docs). It also completed its North Star overhaul, which renamed the old FXS governance token to FRAX and made FRAX the gas token of Frax’s own Fraxtal network. The upshot is that Frax Ether is now one component of a wider stablecoin-and-rollup system run by founder Sam Kazemian, rather than a standalone staking brand.

Frax Ether is by far the smallest of the three, with protocol value locked around $94 million per DefiLlama and combined token market caps near $262 million (CoinGecko). For a certain kind of user that is perfectly fine: the appeal is composability and a clean yield-bearing vault, not scale or a large operator network.

The governance tokens: what LDO and RPL are actually worth

Each protocol has a token that is separate from its staked-ETH token, and their prices tell a story of their own. Lido’s LDO trades at $0.3767 for a market value of about $314 million, ranked #135, per CoinGecko. Set that against the roughly $24.24 billion of stETH that LDO holders govern and the mismatch is striking: the governance token is worth only about 1.3% of the staked ETH it steers, a ratio of roughly 77 to 1. That gap is exactly why dual governance exists. If control rode on LDO alone, it would be cheap to buy influence over a very large pool of other people’s ETH, so Lido handed stETH holders a veto.

LDO has had a rough cycle regardless, sitting about 95% below its August 2021 peak of $7.30 and only recovering off a June 2026 low of $0.2350 (CoinGecko). Rocket Pool’s RPL is smaller still, at $1.70 for a market cap under $40 million (CoinGecko), and after Saturn it is no longer required to run a validator; it is now an optional way to earn a share of protocol revenue in ETH. Frax took a third path entirely: its North Star overhaul folded staking into a wider system governed by the renamed FRAX token, so there is no standalone Frax Ether governance token to price. For a staker, the takeaway is simple: these tokens are bets on protocol growth, not on the yield of the staked-ETH token, and their volatility should not be confused with the relative stability of stETH, rETH or sfrxETH.

The yield question, and who really pays for it

Under the branding, all three protocols sell the same underlying thing: Ethereum’s validator reward, currently about 2.59% before fees (validatorqueue). That reward is built from newly issued ETH, users’ priority fees, and MEV, the extra value builders extract from ordering transactions; our explainer on how on-chain value is extracted and reclaimed covers that last layer. Whoever runs the validator keeps a cut, and you get the rest.

Net of fees, the three land close together. stETH pays roughly 2.3% after Lido’s 10% commission (split evenly between operators and the DAO). rETH has paid around 2% recently. sfrxETH is usually the highest, because it concentrates yield onto fewer tokens. None of these clears the roughly 3.8% an investor can earn on a three-month US Treasury bill today, which is the uncomfortable fact under the whole category: staked ETH is a thin, ETH-denominated coupon layered on top of ETH price risk, not a dollar yield. We dig into that gap in our look at real yield across the chains, and the broader question of who actually pays to secure a blockchain.

MEV is where the small yield differences actually come from. Because validators, or the builders they outsource block construction to, can earn extra by ordering transactions cleverly, a professional operator running sophisticated relays captures more of it than a hobbyist. Lido’s curated operators are tuned for this, which is part of why stETH tends to edge out rETH on realized yield even though the base reward is identical for everyone. It is also why the yield you actually receive drifts from month to month: issuance is steady, but tips and MEV rise and fall with on-chain activity. None of the three can promise a fixed rate, and any provider that does is, by the SEC’s own description, stepping outside the safe harbor.

Rebasing versus exchange-rate, and why it changes your taxes

The rebasing versus exchange-rate split is not a cosmetic detail. DeFi protocols generally prefer non-rebasing tokens, because a balance that changes every day breaks the accounting inside many lending and liquidity contracts. That is precisely why Lido ships wstETH, the wrapped, exchange-rate version of stETH, and why wstETH, not stETH, is what you usually see as collateral on Aave. Rocket Pool’s rETH and Frax’s sfrxETH are exchange-rate by design, so they slot into DeFi without a wrapper.

The same split changes how a US holder is taxed, and it is easy to miss. A rebasing token like stETH drips a stream of new tokens into your wallet, and staking rewards are generally treated as income at their fair market value when you gain control of them, which can mean a running series of taxable events. An exchange-rate token does not increase your token count; the gain shows up only when you sell, which tends to defer the tax into a single capital event. That is not tax advice, and the details depend on your situation, but two products with almost identical yield can produce very different tax paperwork. If you plan to hold for a long time and touch DeFi, the exchange-rate tokens are often the tidier choice.

The one-third problem, three different answers

The reason anyone worries about Lido’s size is a specific consensus concern. Ethereum researcher Danny Ryan laid out the thresholds in a widely cited note: a single staking entity past one-third of all stake can stall finality, past one-half can censor transactions, and past two-thirds can finalize an invalid chain (Danny Ryan). Vitalik Buterin has repeatedly named staking and liquid-staking concentration “one of the biggest risks to the Ethereum L1” in his roadmap writing (The Block).

Each protocol answers that differently, and the ranking is the mirror image of the size ranking. Lido, in the low-20s percent of all staked ETH, sits closest to the line and leans on dual governance, a curated-plus-permissionless operator mix, and now ValOS to argue it can be large without being a single point of control. Rocket Pool answers by construction: with around 2,000 permissionless operators and 4 ETH bonds, no single actor holds the keys. Frax mostly sidesteps the question by being small; at well under 1% of staked ETH it is not a systemic actor, and its model leans on a validator lending market rather than a broad operator base. If decentralization is your first concern, Rocket Pool wins that axis clearly, and Lido’s whole 2026 rebuild is an attempt to stop losing it.

Can you actually get your ETH back?

Liquidity is the feature you are really buying, so it is worth checking how each token converts back to ETH. There are always two doors: protocol redemption, where the token is burned and ETH is returned from the validator set, and the secondary market, where you simply sell. Since the Shapella upgrade of 2023, all three tokens can be redeemed through their protocols, so both doors exist for each.

The details differ. Lido’s withdrawal queue normally clears in a day to a few days, with a Bunker mode that slows things down in a mass-exit scenario, and stETH is liquid enough that most holders just sell instantly. Rocket Pool’s rETH is redeemed from a deposit pool; when that pool is full, redemption is quick, but if it runs dry you fall back on the secondary market, where a discount can appear. Frax’s frxETH and sfrxETH lean heavily on secondary-market liquidity, mostly on Curve, so a smooth exit depends on pool depth rather than a large redemption buffer.

The cautionary tale is stETH in June 2022, when it traded down to about 0.94 ETH as Celsius and Three Arrows Capital were forced to dump it into thin liquidity, per CoinDesk. There was no hack; stETH was simply not redeemable yet, so the only door was the market. That specific trap is closed now that redemptions exist, but the general lesson holds: a liquid staking token is only as liquid as its worst day.

Where the risk actually lives

The real risks are not the yield differences; they are the failure modes. There are four worth naming. Smart-contract bugs, because all three are complex, upgradeable code. Oracle and collateral risk, which bites when LSTs are used in lending markets: in March 2026 a stale Aave price oracle briefly undervalued wstETH and triggered roughly $27 million in avoidable liquidations, with zero bad debt but real losses for the accounts caught in it, per The Block. Slashing, where operator error can burn stake, though after Pectra the base penalty is small and the true danger is a correlated, many-validators-at-once failure. And concentration itself, which is a network risk that a single holder cannot diversify away.

The three protocols distribute those risks differently. Rocket Pool spreads operator risk the widest, so a single bad actor matters least, but its permissionless set is harder to audit centrally. Lido concentrates operator risk among vetted shops, and its 2026 answer is to bond and independently review them. Frax’s lending-market model adds a risk the others do not have, borrowed-ETH leverage sitting inside the validator set, in exchange for a cleaner yield structure. There is no free lunch here; each design trades one risk for another.

The US regulatory picture

For a US reader, 2026 is the year the legal cloud mostly lifted. On 5 August 2025 the SEC’s Division of Corporation Finance stated that liquid staking activities do not, by themselves, involve the offer and sale of securities within the meaning of the Securities Act, per the SEC statement and its companion release. Commissioner Hester Peirce described liquid staking as “a variant on the longstanding practice of depositing goods with an agent who performs a ministerial function in exchange for a receipt that evidences ownership of the goods.” The guardrail is spelled out too: a provider that guarantees a fixed return or exercises real discretion over how and when to stake falls outside that safe harbor.

That clarity, building on the SEC’s earlier protocol-staking statement, is why staked-ETH products have reached regulated wrappers; our review of the crypto ETF assembly line tracks those funds. It does not mean enforcement risk has vanished, as our piece on SEC crypto enforcement in 2026 lays out; it means the baseline activity of these three protocols is, for now, on the right side of the line. Europe’s MiCA, by contrast, still has no dedicated staking category, so this is a US-centric clarification rather than a global one.

The emission war hanging over all three

One 2026 development threatens all three at once, and it is worth watching. A draft proposal, EIP-8363 (the Tapered Issuance Burn), would gradually burn validator rewards down to zero net new issuance once about 60.25 million ETH, roughly half the supply, is staked, per CoinDesk. With about 43 million ETH staked today, that is not a distant target. Co-author Jérôme de Tychey warned there could be “more than 70 million ETH staked by January 2028 if nothing changes.”

The proposal did not make near-term inclusion and drew fierce pushback from DeFi founders, who argue it would gut staking-based strategies. But the intent is explicit: cap the staking ratio to fight the exact centralization that Lido’s size embodies. If some version eventually passes, the base reward all three repackage gets thinner, and the competition shifts even further toward fees, composability and risk management rather than headline yield. In other words, the case for picking a protocol on anything other than a fraction of a percent only gets stronger.

How to choose between them

Because net yields sit within a fraction of a percent of each other, the real choice is about trust model, liquidity and how you intend to use the token, not the APR. A simple way to frame it: pick Lido if your priorities are deep liquidity, the widest DeFi acceptance and institutional-grade tooling, and you are comfortable holding the market’s most concentrated position. Pick Rocket Pool if decentralization is the point and you accept smaller scale and slightly lower yield to get a permissionless operator set. Pick Frax if you want a clean, composable yield vault embedded in a broader DeFi and rollup system, and you do not need size.

Whichever you choose, treat the token as what it is: ETH exposure plus a thin yield plus a set of protocol-specific risks. Hold your own keys where you can, understand what you are signing, and remember that the biggest variable in your return is still the price of ETH, not which of these three logos is on the wrapper.

Frequently Asked Questions

Is Lido safe to use given how much of Ethereum it controls?

Lido has never been exploited at the protocol level and its stETH stays deeply liquid, but its size is a governance concern rather than a bug: it sits in the low-20s percent of all staked ETH, close to the one-third threshold that researchers flag. In 2026 it added dual governance, ETH-bonded operators and the ValOS operations standard to answer that concern. For an individual holder, the practical risks are smart-contract bugs and market liquidity, not an imminent collapse.

Which pays the highest yield, Lido, Rocket Pool or Frax?

They are closer than the branding suggests. All three repackage Ethereum’s base reward of about 2.59 percent, and after fees stETH pays roughly 2.3 percent, rETH around 2 percent, and Frax’s sfrxETH is usually the highest because it concentrates yield onto fewer tokens. The differences are small enough that trust model and how you use the token matter more than a fraction of a percent.

What is the difference between stETH, rETH and sfrxETH?

stETH is rebasing: your token balance grows daily while the price stays near 1 ETH. rETH and sfrxETH are exchange-rate tokens: the count stays fixed and each token becomes worth more ETH over time, which is why they trade above 1 ETH. stETH comes from Lido, rETH from Rocket Pool’s permissionless operators, and sfrxETH from Frax’s yield-concentrating vault.

Can I lose money staking with Lido, Rocket Pool or Frax?

Yes. The main ways are a smart-contract exploit, a validator slashing event caused by operator error, and selling your token at a discount during a liquidity crunch, as stETH briefly did in 2022. Since the Shapella upgrade all three tokens can be redeemed for ETH through the protocol, which makes a lasting discount less likely, but none of these products is risk-free.

Is liquid staking legal in the United States in 2026?

For now, yes, at the activity level. In August 2025 the SEC’s Division of Corporation Finance stated that liquid staking and the receipt tokens it produces are not by themselves securities offerings, as long as the provider stays administrative and does not guarantee returns. That is regulatory guidance, not a law, and it does not remove other risks, but it is why staked-ETH products have reached regulated fund wrappers.

By Yuki Tanaka, senior staking and DeFi correspondent at HOGE Wire.