Lido vs Rocket Pool vs Frax: The Staking Spread in 2026

Staking ETH now pays less than a Treasury bill

The first Friday of September 2026 handed Ethereum stakers a rates problem they could not ignore. A stronger-than-expected August payrolls report pushed traders toward the odds of another Federal Reserve move this month rather than a cut, and the short end of the Treasury curve stayed firm. The yield on the 3-month US Treasury bill sat at 3.86% on September 4, the safest dollar return an investor can buy. On the same day, staking ETH through Ethereum’s consensus layer paid a base rate near 2.7%. Read literally, cash out-yields the asset you take smart-contract, slashing and price risk to hold. We walked through what that macro turn did to crypto positioning in our September jobs coverage.

That single comparison reframes the whole Lido versus Rocket Pool versus Frax question. When staking paid 5% and Treasury bills paid almost nothing, the choice of provider was close to a rounding error, because any of them beat cash by a wide margin. In 2026 the cushion has gone. Ethereum staking pays a thin, ETH-denominated coupon, and the differences between the three largest liquid staking protocols, their fees, their token mechanics, their decentralization and their failure modes, matter more now precisely because there is so little spread left to give away. As we argued in a piece on real yield as a spread over T-bills, every crypto return eventually gets measured against the risk-free rate, and staking is no exception.

This article treats the three protocols as three ways to earn that spread. It prices what each one actually pays after fees, what risk sits behind the number, and who should pick which. Proof-of-work miners face the mirror image of the same math, the falling price of hashpower, which we set side by side with staking in hashprice versus staking yield. The through line is the same: when the price of security falls, the operators with the lowest costs and the cleanest risk survive.

What Lido, Rocket Pool and Frax actually are

All three solve the same problem. Native staking locks 32 ETH (or, since the Pectra upgrade, up to 2,048 ETH) inside a validator, where it sits illiquid behind an entry queue and an exit queue. Liquid staking hands you a tradable token that stands in for that staked position and its accruing rewards, so your capital keeps moving through lending markets, liquidity pools and collateral vaults while it earns. Where the three protocols part ways is how decentralized the validator set is, how the token tracks value, and how the yield is packaged.

Lido is the incumbent by a wide margin. You deposit ETH and receive stETH, a rebasing token whose balance grows in your wallet each day; most DeFi users hold the wrapped version, wstETH, which keeps a fixed balance and rises in price instead. Behind stETH is a curated set of professional node operators plus a permissionless Community Staking Module. At current prices stETH carries a market cap of roughly 23.8 billion dollars, dwarfing every rival.

Rocket Pool is the decentralization purist. Anyone can stake any amount for rETH, an exchange-rate token, or put up a reduced bond and run a validator for commission. Its validator set is spread across roughly 2,000 permissionless operators, with rETH worth about 912 million dollars. Frax is the money-lego of the group. It splits the job across two tokens: frxETH, a liquid unit pegged to ETH that pays no yield on its own, and sfrxETH, a vault that collects the staking reward for the holders who lock into it. Frax Ether is the smallest of the three, with frxETH near 151 million dollars and sfrxETH near 104 million dollars.

Where the base yield comes from

Before comparing providers, it helps to know what the underlying yield is and why it keeps falling. A validator earns from three sources: new ETH issuance for proposing and attesting to blocks, priority fees that users pay to get transactions included, and maximal extractable value (MEV) captured through the block-building market. Issuance is the largest and most predictable piece, and it is designed to shrink as more ETH is staked. The protocol scales rewards roughly inversely with the square root of the total staked, so every new validator dilutes the per-validator slice.

That mechanism is doing exactly what it was built to do. Roughly 34% of all ETH is now staked, an all-time high, which has pushed the base reward to a three-year low near 2.7%. Validators that run MEV-Boost add another 0.5% to 1% on top, so realistic all-in yields land around 3.3% to 3.8% in ETH terms. The Pectra upgrade (see EIP-7251) raised the maximum balance per validator from 32 to 2,048 ETH, letting large operators consolidate many validators into one and compound rewards automatically, a change that matters a great deal to Lido’s roadmap below. None of this changes the headline: the reward pool is fixed by protocol rules, and the more capital chases it, the thinner each stake’s cut becomes.

Every staking yield is a spread over T-bills

Here is the lens that makes the 2026 comparison honest. The 2.7% base is paid in ETH, not dollars. If you convert it, a staker earning 2.7% in ETH is earning less, in dollars, than the 3.86% a Treasury bill pays with no credit or duration risk. Staking is therefore not a dollar-yield trade. It is a way to make an ETH position productive, and it only comes out ahead if you were going to hold ETH anyway or if ETH appreciates enough to cover the gap. The staking reward is the spread you collect on top of that price bet.

Once you see the yield as a spread, the provider question changes shape. You are no longer asking who pays the biggest number; you are asking who lets you keep the most of a small number after fees, and who charges the least in hidden risk to do it. The table below walks through the arithmetic in ETH terms, then measures it against the dollar risk-free rate.

The negative bottom row is the uncomfortable truth of the moment. It does not mean staking is pointless; it means the reason to stake is ETH exposure, not carry. It also means the difference between a provider that nets you 2.7% and one that nets you 2.3% is no longer trivial, because 0.4% is a meaningful share of a spread this thin. And it puts a hard question to every headline APR: is the yield real cash flow, or is it padded with token emissions that dilute you elsewhere? We pulled that thread apart in a companion piece on telling cash flow from emissions, and it turns out to separate these three protocols more than any marketing page does.

Lido: the incumbent that became infrastructure

Lido wins on scale, liquidity and integration, and it carries the risks that come with all three. stETH is the most liquid liquid staking token on Earth, accepted as collateral across the largest lending markets and paired in the deepest pools. Lido takes a 10% cut of staking rewards, split evenly between node operators and the DAO treasury, which leaves stETH holders with the lion’s share of the gross yield. For an institution that wants stETH without touching a public smart contract directly, the token now mints and burns inside a federally chartered bank after Anchorage Digital integrated Lido in mid-2026, and treasury company SharpLink committed 200 million dollars of ETH through the protocol in August.

Nathan McCauley, co-founder and chief executive of Anchorage Digital, framed the shift bluntly, saying liquid staking has become “one of the most important building blocks for institutional participation in Ethereum.” That institutional pull is exactly why Lido’s size is also its liability. Its stETH represents roughly 23% of all staked ETH, down from a 32% peak in late 2023 but still large enough to sit near the consensus thresholds that make researchers nervous. A single entity approaching one third of stake can, in theory, delay finality; approaching one half enables censorship. Vitalik Buterin has repeatedly named staking centralization “one of the biggest risks to the Ethereum L1” in his outline of the Scourge research phase, warning that economic incentives push smaller validators into large pools over time.

Lido’s answer has two parts. The first is dual governance, live since July 2025, which lets stETH holders veto or delay DAO decisions they oppose: a signal from holders of more than 1% of stETH triggers a timelock, and a rage-quit path opens if opposition passes 10%, letting dissenters exit before a contested change takes effect (The Block). The second is the 2026 Core upgrade. In late July, Lido began moving 16.5 billion dollars of staked ETH onto Pectra-era high-capacity validators, a migration that is expected to cut Ethereum’s total validator count by about a third and reduce attestation messages by roughly 29% per epoch. As part of the same overhaul, all 34 of Lido’s curated operators move to Curated Module v2 and post locked ETH bonds for the first time, turning reputation into skin in the game. The first consolidations were slated to begin in September, with no action required from stakers.

Rocket Pool: the decentralization bet

Rocket Pool sells the opposite value proposition. Instead of a curated shortlist of professional operators, it spreads validation across roughly 2,000 permissionless node runners, anyone with the hardware and a bond can join. That design is the whole point: if you believe concentration is the risk that matters, rETH is the token that most directly funds an alternative. rETH is an exchange-rate token, so its quantity in your wallet never changes while each unit slowly buys more ETH; at current prices one rETH is worth about 1.17 ETH, and that ratio only climbs.

The economics changed materially with the Saturn 1 upgrade, live on mainnet on February 18, 2026. It cut the minimum node bond from 8 ETH to 4 ETH, halving the barrier to running a validator, and introduced megapools that let one operator manage many validators under a single contract. Just as important for the token holder, Saturn activated the long-debated RPL fee switch: a share of protocol revenue, around 9%, now flows to staked RPL holders paid in ETH rather than through token inflation, and RPL’s old inflationary rewards are being phased out (Rocket Pool). That is the emissions-to-cash-flow shift in miniature: RPL is being rebuilt as a claim on real ETH revenue instead of a dilution machine, exactly the distinction that separates durable yield from the kind that quietly costs you elsewhere.

Rocket Pool’s tradeoffs are size and mechanics. Its total value locked sits around 1.1 billion dollars, a fraction of Lido’s, so rETH liquidity is thinner and large redemptions can be slower. Node operators charge a commission (historically around 14%) on the ETH the pool supplies, which the protocol documents in its own operator FAQ. And the RPL token doubles as insurance collateral, which ties the health of the token to the health of the network in ways stETH holders never have to think about. You are buying decentralization, and paying for it in liquidity and complexity.

Frax: the two-token money-lego

Frax approaches staking as one component of a larger stablecoin and rollup system rather than a standalone product. Its liquid staking splits into two tokens on purpose. frxETH is a plain, liquid unit that tracks ETH and pays no staking yield by itself; it is designed to be deployed into liquidity pools and integrations where its usefulness comes from being ETH-like and mobile. sfrxETH is an ERC-4626 vault that collects the staking reward. Because frxETH holders forgo the yield in exchange for liquidity and incentives elsewhere, the entire staking reward is concentrated onto the smaller pool of sfrxETH tokens, which is why sfrxETH has historically posted a higher headline APR than single-token rivals. One sfrxETH is currently worth about 1.15 ETH.

Under the hood, Frax runs a permissionless node-operator lending market in its v2 design, where operators borrow ETH to validate at a variable rate and sfrxETH captures both the staking yield and the lending interest. The wider Frax system was reorganized around its North Star plan, which renamed the old FXS governance token to FRAX around the end of 2025 and made FRAX the gas token of the Fraxtal rollup, positioning the project, in founder Sam Kazemian’s framing, as a full stablecoin operating system rather than a single staking pool. The catch is the mirror image of Rocket Pool’s: Frax buys yield concentration and composability at the cost of size and simplicity. Frax Ether is the smallest of the three by an order of magnitude, and its two-token, AMO-driven machinery is more moving parts to understand and to trust.

The three protocols at a glance

The table below sets the structural differences side by side. Read it as a map of tradeoffs rather than a scoreboard, because the right pick depends entirely on which risk you are least willing to hold.

Prices and market caps here are drawn from CoinGecko on September 5, 2026, with ETH itself near 2,459 dollars and a market cap around 300 billion dollars. The governance tokens tell their own story of scale: LDO trades near 0.38 dollars for a market cap around 320 million dollars, while RPL sits near 1.69 dollars for roughly 39 million dollars. Both are a small fraction of the ETH their protocols secure, a mismatch that is itself part of the governance-risk conversation.

The net-spread reality: what each actually pays

Start from the same gross number, because all three validate the same Ethereum and earn the same underlying rewards. The differences come from the fee and from how the yield is packaged. Lido’s flat 10% haircut is the most transparent; on an all-in gross of, say, 3.4%, a stETH holder nets around 3.0% in ETH before accounting for the token’s own tracking. Rocket Pool’s rETH nets a similar figure after node commission, with the twist that some of the value now accrues to RPL holders through the fee switch rather than only to rETH. Frax’s sfrxETH can print a higher headline because it concentrates the reward of a much larger frxETH float onto a small vault, but that number is only available to the subset of users willing to hold the vault token and forgo frxETH’s liquidity role.

The practical takeaway is that in 2026 fees are not a footnote. When the net spread is around 2.4% to 2.7% in ETH and the dollar risk-free rate is 3.86%, a provider that quietly costs you an extra 0.3% is eroding more than a tenth of your entire spread. That is a very different world from 2021, when a 5%-plus yield made fee differences invisible. It is also why the emissions question matters so much here: a headline APR propped up by inflationary token rewards is not the same as one paid in ETH revenue, and a thin-spread environment punishes the difference. Rocket Pool’s move to pay RPL in ETH and Frax’s decision to concentrate real staking yield onto sfrxETH are both, in their way, responses to the same pressure that squeezes miners.

Pricing the risk behind each spread

A spread is only worth what remains after the risk behind it. Each protocol pays roughly the same net yield, so the honest comparison is which risk you are being compensated to hold. For Lido, the dominant risk is concentration. Its size makes stETH the deepest and most useful token in DeFi, but the same size keeps it near the consensus thresholds that worry Ethereum’s researchers, and it concentrates governance influence in a token, LDO, worth a tiny fraction of the stETH it steers. Dual governance and the Core consolidation are attempts to defuse that, but the structural fact remains: when you hold stETH, you are partly betting that Lido’s size never becomes Ethereum’s problem.

For Rocket Pool, the dominant risk is liquidity and mechanism. rETH redemptions draw from a deposit pool; when that pool runs dry, holders who want out have to sell on the secondary market, where a discount can open until arbitrage refills the pool. The RPL collateral system adds a second layer, tying node economics to a small-cap token. For Frax, the dominant risk is complexity and counterparty. The two-token model, the AMO strategies and the v2 lending market are powerful, but they are more surface area for something to break, and the smallest float of the three means the least room to absorb a shock. None of these is a reason to avoid a protocol outright; they are the reason the same headline yield can be worth more or less depending on which risk you can stomach.

Can you actually get your ETH back?

Every liquid staking token has two exit doors, and the difference between them is the difference between a good week and a very bad one. The first door is protocol redemption: you burn the token and the protocol returns the underlying ETH, subject to Ethereum’s validator exit queue. The second door is the secondary market: you sell the token to someone else at whatever price the pool will bear. In calm markets the two doors give the same result. In a panic, the redemption queue slows and the secondary price is where the real action happens.

Lido offers the deepest secondary liquidity of any of the three, plus a withdrawal queue that typically clears in a few days, which is why stETH is the token institutions reach for. Rocket Pool’s rETH depends on its deposit pool for instant redemption; when the pool empties, exits route to the secondary market and a discount can appear, a mechanic every rETH holder should understand before they need it. Frax’s frxETH has no direct protocol redemption for stakers in the usual sense, so its liquidity is effectively secondary-market only, backstopped by the protocol’s own AMO positions. The cautionary tale predates all three current designs: in the 2022 stress that followed Terra’s collapse, stETH fell to roughly 0.94 ETH as Celsius and Three Arrows Capital were forced to dump into thin Curve liquidity. There was no hack and no protocol failure; withdrawals simply did not exist yet, so the only door was the market, and the market gapped. Redeemable withdrawals arrived later with the Shapella upgrade, which is why a repeat would look milder, but the lesson holds: the peg you should watch is market price against net asset value, not the token against ETH.

The leverage that widens the spread

There is a reason so much staked ETH ends up as collateral rather than sitting idle: leverage is the only way to turn a thin spread into a fat one. The classic trade deposits wstETH into a lending market, borrows ETH against it, buys more wstETH and repeats, stacking the base yield several times over through the multiplier that high loan-to-value ratios allow. On markets that grant wstETH an efficiency mode with loan-to-value up to 95%, a 2.7% base can be levered into a high-single-digit or low-double-digit return, which is how a below-T-bill yield becomes attractive again on paper.

The danger is that leverage widens the downside just as neatly. If the borrow rate rises above the staking yield, the carry flips negative and the position bleeds. If the collateral’s oracle misprices the token, even briefly, liquidations can cascade through otherwise healthy accounts, as a stale-timestamp oracle glitch did in early 2026 before it was reimbursed. Looping is a legitimate strategy, but it converts a low-risk yield product into a leveraged bet on the spread between the staking rate and the ETH borrow rate, and it is the main reason a token as sober as stETH can still blow up a portfolio. If you loop, size it for the day the borrow rate spikes, not the day it does not.

Regulation: the SEC finally drew a line

For years the biggest unpriced risk in US liquid staking was legal, not technical. That eased on August 5, 2025, when the SEC’s Division of Corporation Finance issued a statement that certain liquid staking activities and the receipt tokens they produce “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 deposited crypto asset is not itself an investment contract (SEC). Commissioner Hester Peirce, backing the statement, 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,” echoing the title of her earlier note on protocol staking, Providing Security is not a ‘Security’. Chairman Paul Atkins called the move a significant step in clarifying activities that fall outside the agency’s jurisdiction.

Two caveats keep this from being a blank check. First, the relief is staff guidance, not a rule or a statute, and it applies only when the provider sticks to administrative and ministerial functions; a custodian that guarantees a fixed return or exercises discretion over how much and when to stake falls outside it. Second, tax has not gone anywhere. In the US, staking rewards are generally taxed as ordinary income at the moment you gain control of them, and the thin ETH-denominated spread we have been describing is a pre-tax number, so your actual take-home depends heavily on where you live and how your rewards are characterized, a point we mapped in crypto taxes by state. A rebasing token like stETH that pays you daily can create a running stream of taxable events, while an exchange-rate token like rETH or sfrxETH may defer the reckoning until you sell, a difference worth modeling before you pick.

Which one fits which staker

There is no universally best liquid staking token, only a best fit for a given holder and a given fear. The matrix below is the short version of everything above.

The honest verdict for 2026 is that the three protocols are closer on yield than their marketing suggests and further apart on risk than a headline APR can show. Lido gives you liquidity and infrastructure and asks you to carry concentration. Rocket Pool gives you decentralization and asks you to carry liquidity and mechanism risk. Frax gives you composability and concentrated yield and asks you to carry complexity and small size. All three pay a spread that, in dollars, currently sits under the risk-free rate, which means the decision starts with a prior question: do you want to own ETH at all? If the answer is yes, pick the risk you can live with and let the spread compound. If the answer is no, the safest yield in the room is still a Treasury bill.

Frequently Asked Questions

Is staking ETH still worth it if Treasury bills pay more?

In dollar terms, a 3-month Treasury bill near 3.86% out-yields Ethereum’s roughly 2.7% base staking rate, so staking is not a dollar-yield play in 2026. Staking pays an ETH-denominated coupon on top of your ETH price exposure, so it makes sense mainly if you intend to hold ETH anyway and want that position to compound. If your goal is safe dollar income with no price risk, cash currently beats staking.

Which is safest, Lido, Rocket Pool or Frax?

There is no single answer, because each carries a different dominant risk. Lido carries concentration risk from its share of all staked ETH and its governance weight, Rocket Pool carries liquidity risk through its deposit pool and a small RPL collateral market, and Frax carries complexity and counterparty risk in its two-token, AMO-driven design. Lido is the most liquid and battle-tested, Rocket Pool is the most decentralized, and Frax is the most composable and the smallest.

What is the difference between stETH, rETH and sfrxETH?

stETH rebases, meaning your wallet balance grows daily, and it wraps into wstETH for use in DeFi. rETH and sfrxETH are exchange-rate tokens whose quantity stays fixed while each token buys more ETH over time. That is why rETH trades near 2,875 dollars and sfrxETH near 2,821 dollars while stETH tracks ETH near 2,458 dollars; the higher price reflects accrued rewards, not a premium you are overpaying.

How much do Lido, Rocket Pool and Frax charge?

Lido takes 10% of staking rewards, split evenly between node operators and the DAO. Rocket Pool node operators charge a commission, historically around 14%, on the ETH the pool supplies, and after the Saturn 1 upgrade a share of protocol revenue also flows to staked RPL holders paid in ETH. Frax does not charge a flat headline fee; it routes the staking reward to sfrxETH holders and its AMO, which concentrates the yield onto fewer tokens.

Are liquid staking tokens securities in the United States?

In an August 5, 2025 statement, the SEC’s Division of Corporation Finance said certain liquid staking activities and staking receipt tokens do not involve the offer and sale of securities, provided the provider limits itself to administrative and ministerial functions. A custodian that guarantees a fixed return or exercises discretion over staking falls outside that view, and the statement is staff guidance rather than a rule or law, so it can change.

By Yuki Tanaka, HOGE Wire staff writer covering Ethereum staking, DeFi yield and market structure.