Restaking’s Sector Map for 2026: Who’s In, Who’s Out

Restaking turned three years old in 2026, and the sector no longer looks like the one Sreeram Kannan first pitched through EigenLayer’s whitepaper. What began as a single idea, let staked ETH secure more than just Ethereum, has splintered into at least five distinct businesses with different tokens, different collateral, and increasingly different definitions of what restaking even means. The freshest data point: Karak, one of the three protocols most commonly named alongside EigenLayer and Symbiotic, has rebranded to OpenGDP and quietly dropped restaking from its own homepage.

This piece is a field guide to where the sector actually stands: who still calls this restaking, who has walked away from the term, how much value is actually locked where, and what a spring 2026 bridge exploit taught the category about the difference between a slashing condition on paper and a loss in practice. For a ground up explanation of the mechanics themselves, our beginner’s guide to restaking is a better starting point; this one assumes you already know roughly what an AVS is and wants to know who is actually winning.

What Restaking Means, in Two Paragraphs

Restaking starts with ordinary Ethereum staking: lock 32 ETH, or a fraction of it through a pool, to help validate the chain and earn a base reward currently in the high 2% annual range once priority fees and MEV are included, according to Staking Rewards. Restaking takes that same staked ETH, or a liquid staking token representing it, and pledges it a second time to secure a different piece of infrastructure entirely, such as an oracle network, a data availability layer, a rollup sequencer, or a bridge. In EigenLayer’s original framing, these outside services are called Actively Validated Services, or AVS. In exchange for extra yield, a restaker accepts extra slashing conditions; if an AVS is compromised because a restaked operator misbehaved, part of that operator’s stake can be destroyed.

Liquid restaking tokens, or LRTs, exist to make this less illiquid. Deposit ETH into a protocol like ether.fi or Kelp DAO and receive a tradeable receipt token that can still be used elsewhere in DeFi while the underlying ETH is restaked. That reuse, sometimes called rehypothecation, is exactly what made the sector’s April 2026 stress test, covered below, worth taking seriously.

The Fresh Hook: Karak Just Left the Restaking Business

The clearest sign that restaking’s early land grab is over came this month, when Karak, previously grouped with EigenLayer and Symbiotic as one of the sector’s three biggest names, rebranded to OpenGDP. The new homepage describes itself as “the operating layer for real world economic execution,” and makes no mention of restaking, AVS, or shared security anywhere in its main content; those terms have been replaced with tokenization, stablecoin settlement, and programmable economic rights. A companion announcement, “Hello, GDP!” on the project’s own blog, frames the move as an expansion of scope rather than an admission of failure, though the practical effect is the same: one of restaking’s original challengers is no longer competing for restaked collateral.

Karak had raised $48 million in a Series A led by Lightspeed Venture Partners in December 2023, with Coinbase Ventures, Pantera Capital, Mubadala Capital, and Framework Ventures participating at a valuation above $1 billion, according to FinancialIT. By early 2026 its restaking TVL had fallen to roughly $100 million, a small fraction of EigenLayer’s, and no KAR or GDP token had ever reached a stable listing on a major exchange despite years of points campaigns. The old V1 staking, V2 staking, and K2 bridge products still exist and are still linked from OpenGDP’s own footer, and the project says its XP staking program will keep running for now, but the homepage, the marketing, and evidently the company’s own read of where the opportunity lies have all moved on to tokenizing real world assets instead.

EigenLayer and EigenCloud: Still the Biggest, But Not What It Was

EigenLayer itself rebranded first, to EigenCloud, positioning its restaked ETH as the security backbone for a broader verifiable cloud that includes EigenDA for data availability, EigenCompute, and EigenVerify. Slashing has been live on mainnet since April 2025, meaning the risk side of restaking is no longer theoretical for EigenLayer’s roughly 2,000 operators. On the token side, EIGEN trades around $0.23, down more than 95% from its December 2024 all time high of $5.65, with a market capitalization near $173 million on a circulating supply of about 741 million tokens, per CoinGecko.

Total value locked is genuinely difficult to pin to one number this year. Multiple trackers have shown EigenCloud’s TVL anywhere from the mid single digit billions to the mid teens of billions of dollars depending on the week and the methodology, down from a widely cited peak near $19 billion to $20 billion in 2024. DefiLlama and various aggregator sites regularly disagree with each other by a factor of two or three on the same day, which is itself a useful data point: even the sector’s largest player does not have a single, agreed upon measure of how much capital it is actually securing.

What has changed more concretely is the incentive design. In December 2025, EigenLayer’s community passed ELIP-012, which replaced the old token emission mechanism with an EmissionsController that directs new EIGEN toward “productive stake”, meaning capital actively securing a live, fee generating AVS, rather than idle deposits. A 20% fee on AVS rewards, plus all cloud service fees, now flow toward a buyback mechanism intended to reduce circulating supply over time, with inflation held fixed at a combined 8% (7% base plus 1% ecosystem discretionary) and the Protocol Council retaining sole authority over any change to that total rate. Eigen Labs has also leaned further into the cloud framing with outside capital: a16z put $70 million directly into EIGEN tokens in mid-2025 specifically to back the EigenCompute and EigenVerify launch, according to CoinDesk. Founder Sreeram Kannan described the ambition on the Bankless podcast: “Anything you can program on the cloud, you should be able to program on Eigencloud with crypto-grade verifiability. If you do it on this, you don’t need to trust anybody,” a pitch noticeably broader than restaking ETH to secure a single oracle, per Bankless.

Symbiotic: Two Years Live, Still No Token

Symbiotic takes a more permissive approach than EigenLayer: any ERC-20 token can serve as restaking collateral, not just ETH or ETH denominated liquid staking tokens, and slashing has been configurable since its mainnet launch in January 2025 rather than added later. The protocol has raised $34.8 million in total funding, a $5.8 million seed round co-led by Paradigm and Cyber Fund in June 2024, and a $29 million Series A led by Pantera Capital with Coinbase Ventures participating in April 2025, according to CoinDesk. Its TVL has been reported between roughly $1.6 billion and $1.7 billion in mid-2026, putting its market share in the mid single digits as a percentage of total restaked value, a distant second to EigenLayer but still a meaningfully live protocol with more than 50 integrated networks, dozens of operators, and over 100,000 users.

What Symbiotic still does not have, more than two years after its points program began, is a public token. FinanceFeeds reported in mid-2026 that “no supply, allocation, vesting schedule, or conversion mechanics have been announced publicly,” and warned that the delay is itself now a competitive risk, since “every month without a token announcement increases the probability that restakers will rotate capital” toward protocols like EigenLayer or Babylon that already pay out in a liquid asset. Symbiotic’s own answer, so far, has been to diversify rather than to launch: in June 2026 it shipped Liquid Lane, extending its vault infrastructure to real world assets including credit, insurance, and tokenized asset liquidity, already securing more than $550 million across dozens of applications. Co-founder Misha Putiatin framed the opportunity bluntly: “The RWA market has crossed $33 billion, but most of those assets still can’t be redeemed on demand.” It is a notably similar pivot to the one OpenGDP just made, just executed as an addition rather than a full rebrand.

Babylon: Restaking Bitcoin Instead of Ethereum

Not every restaking protocol touches Ethereum at all. Babylon lets Bitcoin holders lock BTC directly using Bitcoin’s own timelock scripting, with no wrapping and no bridging to another chain required. That locked Bitcoin then secures proof of stake networks that opt in as Bitcoin Supercharged Networks, borrowing Bitcoin’s security budget for chains that would otherwise have to bootstrap their own validator sets from scratch. It is a fundamentally different trust model from EigenLayer or Symbiotic, since the collateral never actually leaves Bitcoin’s own ledger.

Babylon’s genesis mainnet and its BABY token both launched in April 2025, with an airdrop of 600 million BABY, 6% of total supply, split between phase one stakers, staking reward participants, Pioneer Pass NFT holders, and open source contributors, according to The Block. By mid-2026, Babylon’s own dashboard showed roughly 56,800 BTC staked, worth about $5.6 billion, making it by that measure the largest Bitcoin staking system in existence, per Babylon’s live tracker. BABY itself trades far below what that backing might imply: around $0.013, with a market capitalization near $53 million against a circulating supply of roughly 4 billion tokens, a wide gap between the value secured and the value of the token meant to capture some of that activity, per CoinGecko.

The next milestone worth watching is a planned integration with Aave V4 that would let native BTC serve directly as lending collateral without a wrapped intermediary, though the timeline depends on Aave V4’s own launch and the security review of that connector. If it ships, it would be one of the clearest bridges yet between Bitcoin’s passive holder base and Ethereum’s DeFi liquidity, using restaking as the connective layer rather than a custodial wrapped BTC token.

SSV Network: Restaking Through Distributed Validators

SSV Network approaches restaking from the validator infrastructure side rather than the collateral side. Its core product is Distributed Validator Technology: a validator’s signing key is split via Shamir secret sharing across four or more independent, non-trusting operators, who reach consensus on every signature through IBFT and BLS threshold signatures. No single operator ever holds a complete key, which reduces the odds that one server outage or one compromised machine gets a validator slashed. SSV describes itself as the largest Ethereum DVT provider, citing more than 7 million ETH staked, over $16 billion in ETH secured, and more than 1,800 operators on its own dashboard, though independent trackers have shown meaningfully lower figures in any given week, so treat SSV’s own numbers as the ceiling of the range rather than a universally agreed figure.

SSV 2.0, branded Based Applications, extends this into restaking proper: validators opt in using participation keys rather than withdrawal keys, so the core 32 ETH principal is never slashable, while only additional, voluntarily delegated capital carries risk. CEO Alon Muroch called it the protocol’s “biggest, most ambitious project” that “will profoundly change the restaking market,” according to Cointelegraph. A companion tokenomics change, cSSV, launched in April 2026, letting stakers convert SSV into a liquid, ETH denominated rewards token rather than relying on emissions alone. Despite the redesign, SSV trades around $2.03, more than 96% below its March 2024 all time high of $65.82, with a market capitalization under $30 million, per CoinGecko. Concrete usage data backs up the adoption story independent of SSV’s own token price: Lido’s Simple DVT module, which runs on SSV and Obol middleware, reached roughly 9,500 validators across 261 operators by mid-2025 and had filled its full allocated stake share by the end of that year, according to Lido’s own published metrics, a real and growing production workload that has nothing to do with where SSV trades on a given day.

The Restaking Sector at a Glance

Putting these five side by side is a useful gut check on how differently the word “restaking” gets used across the sector in 2026. Only two of the five still have a liquid, exchange listed token, only one has walked away from the category entirely, and the TVL figures below should all be read as directional rather than precise, for the reasons explained above.

When Liquid Restaking Breaks: The Kelp DAO Lesson

Restaking’s biggest real world stress test so far had nothing to do with a validator getting slashed. On April 19, 2026, attackers exploited a cross-chain bridge used by Kelp DAO’s liquid restaking token, rsETH, minting 116,500 rsETH out of thin air rather than draining existing deposits, an increase of roughly 18% in the token’s supply overnight. That freshly minted rsETH, worth about $292 million, was deposited as collateral on Aave V3 and used to borrow wrapped ETH, leaving Aave with roughly $196 million in bad debt once the exploit was discovered, according to CoinDesk. Aave’s total value locked fell from roughly $26.4 billion to about $20 billion over that single weekend.

The mechanism is a useful case study for anyone reading our deep dive on bridge hacks: the underlying restaked ETH was never actually at risk, and no EigenLayer slashing condition was ever triggered. The failure sat entirely in the bridge Kelp used to represent rsETH across roughly 20 chains, and in Aave’s decision to accept that bridged representation as collateral without pricing in the bridge’s own risk. Aave founder Stani Kulechov personally pledged 5,000 ETH toward the recovery and rallied a coalition including Lido, ether.fi, and Consensys to help cover the shortfall, saying at the time: “Aave is my life’s work and we’re working nonstop to find the best possible outcome for users,” per CoinDesk. By June 2026, the roughly 117,000 exploited rsETH had been burned and the token’s backing restored, with Kelp migrating its cross-chain bridge from LayerZero’s OFT standard to Chainlink’s CCIP, now requiring four independent attestors and 64 block confirmations before a cross-chain mint can settle.

LayerZero publicly disputed responsibility, arguing the exploit was a consequence of Kelp’s own bridge configuration rather than a flaw in the base protocol, a disagreement that never fully resolved. Either way, the lesson for the sector was the same: a liquid restaking token’s risk profile is only as strong as every bridge, oracle, and integration built on top of it, not just the restaking contract itself.

A Timeline of How Restaking Got Here

The sector’s last three years, compressed into eight dates that trace a path from a single Ethereum-only idea to five genuinely different businesses:

Where Does Restaking Yield Actually Come From

Restaking yield is not one thing. Depending on the protocol and the moment, a restaker’s return can come from up to four different sources:

• The underlying ETH staking reward itself, currently in the high 2% annual range
• Points that may or may not eventually convert into a token
• Protocol token emissions funded by inflation rather than by revenue
• A genuine AVS fee split, in the smaller number of cases where an outside service is actually paying for security

For most of restaking’s history, points and emissions have done the heavy lifting, which is precisely why EIGEN, SSV, and BABY have all fallen more than 90% from their all time highs even as the protocols underneath them kept growing their usage. On-chain revenue data makes the gap visible: even with TVL still measured in the billions, measured protocol revenue for the largest restaking platforms remains a small fraction of that figure on trackers like DefiLlama, meaning most of what restakers actually earn today is still emissions and points rather than fees paid by AVS customers for real security demand. EigenLayer’s ELIP-012 redesign, described above, is the sector’s most direct attempt to fix this by routing rewards specifically toward AVS that are actually live and paying fees rather than toward idle deposits. SSV’s cSSV mirrors the same instinct on a smaller scale, converting network fees into ETH denominated payouts instead of relying purely on SSV emissions. Whether either change is enough to make the yield real, in the sense of being funded by genuine demand for security rather than by a token’s own inflation, is still an open question; we cover the mechanics of that question in more depth in our explainer on restaking yield after the points era.

Restaking vs Liquid Staking: Same Idea, Different Risk

It is worth being precise about the difference between liquid staking and liquid restaking, since the two get conflated constantly. A liquid staking token like Lido’s stETH or Rocket Pool’s rETH represents ETH staked to secure Ethereum itself and nothing else; the only slashing risk is ordinary validator misbehavior, which is rare and generally small. Our comparison of Lido, Rocket Pool, and Frax covers that market in detail. A liquid restaking token adds a second, stacked layer of risk on top: the same capital is now also exposed to whatever AVS it is securing, plus the smart contract risk of the restaking protocol itself, plus, as Kelp DAO showed, the risk of any bridge or wrapper used to move the receipt token around.

Regulators have started to draw a similar line. United States Securities and Exchange Commission staff guidance issued in May 2025 concluded that protocol staking, whether solo, delegated, or custodial, is not itself a securities transaction, and extended similar treatment to liquid staking tokens like stETH in a follow up statement three months later, according to the SEC’s own release. That guidance stops at staking and liquid staking; it does not extend to restaking or liquid restaking tokens, which remain in a genuine gray area. Nothing in current staff guidance says restaking is a security, but nothing says it is not, either, and the guidance itself is non-binding staff level interpretation that a future commission could revise.

The Vitalik Warning, Revisited

Restaking had its most prominent critic before EigenLayer’s mainnet even existed. In a May 2023 essay titled “Don’t overload Ethereum’s consensus,” Vitalik Buterin warned against turning Ethereum’s validator set into a general purpose trust layer for arbitrary outside applications. His core argument: “Any expansion of the ‘duties’ of Ethereum’s consensus increases the costs, complexities and risks of running a validator.” Pile enough unrelated obligations onto the same set of validators, he argued, and a failure in any one of them risks dragging Ethereum’s own social consensus into a dispute that was never really about Ethereum, since a large enough AVS failure could create pressure to bail it out rather than let it fail cleanly.

Kannan’s response at the time, given to CoinDesk, was that restaking was actually the more conservative design, not the riskier one: “Anything that restaking can do, already liquid staking can do,” he said, arguing that composability risk already existed once staked ETH could be used as DeFi collateral, and that restaking simply made the risk explicit and priced rather than hidden. Three years and one live slashing mechanism later, the Kelp DAO incident arguably validated a version of both arguments at once: the restaking contracts themselves held up exactly as designed, but the composability Kannan described, in the form of a bridged LRT accepted as lending collateral, is precisely where the actual loss occurred. Vitalik’s warning was about consensus layer overload specifically, and that exact scenario has still not been tested at scale; no EigenLayer AVS failure has yet been large enough to raise the too big to fail question he originally described.

Regulators Still Haven’t Weighed In On Restaking Specifically

Restaking sits in a specific blind spot in current U.S. policy. The SEC’s broader 2026 shift, a mix of dismissed enforcement cases, a March interpretive release stating that most crypto assets are not themselves securities, and a proposed rulemaking agenda aimed at giving token issuers clearer paths to compliance, has been friendlier to crypto generally than at any point since the agency’s creation. Our mapping of US crypto enforcement in 2026 covers that shift in full. But the specific staff statements that cleared staking and liquid staking never mentioned restaking, AVS, or slashing conditions tied to a third party service, and no comparable statement has filled that gap since.

That absence matters more for some of the protocols above than others. A custodial restaking-as-a-service product, where a platform stakes on a user’s behalf and takes a cut, looks structurally similar to the staking-as-a-service arrangement Kraken settled with the SEC for $30 million back in February 2023, before the current guidance existed; that settlement forced Kraken to shut down its US staking-as-a-service product entirely, a considerably harsher posture than the one now applied to solo and delegated staking, and a reminder that guidance can and does change with the composition of the commission. A user restaking their own ETH directly through a smart contract they interact with themselves looks much closer to the self-custodial staking the 2025 guidance already covers. Points programs, which most of the protocols above have run at some stage, sit in their own separate gray area entirely, since a points balance that later converts into a token can resemble, in substance, an unregistered promise of future value depending on how discretionary that conversion is. None of this is likely to produce enforcement action against restaking protocols in the current environment, but it does mean the legal footing under LRTs specifically remains less settled than the footing under stETH.

What the Karak Exit Actually Signals

Line up the moves and a pattern appears that is easy to miss looking at any single protocol in isolation. EigenLayer did not abandon restaking, but it renamed itself around cloud services and now measures some of its success in AI-adjacent compute fees rather than in restaked ETH alone. Symbiotic did not abandon restaking either, but its newest growth product, Liquid Lane, has nothing to do with AVS security and everything to do with real world assets. Karak went further and dropped restaking from its homepage entirely. Three different protocols, three different distances traveled, but all three moving in a similar direction: away from restaking ETH to secure an AVS as the sole pitch, and toward broader infrastructure or asset tokenization plays that do not depend on convincing new AVS builders to pay for shared security.

None of this means restaking as a category is dying. Babylon and SSV are both still growing their core restaking or restaking-adjacent products, measured in real BTC staked and real ETH secured rather than in points. But the sector’s founding thesis, that a wave of new AVS builders would show up and pay meaningfully for restaked security, has not played out on the timeline or at the scale the 2023 to 2024 hype cycle implied, and published protocol revenue figures for the biggest restaking players remain thin relative to the billions still nominally locked. The protocols with the deepest pockets are hedging by building something else next to restaking, rather than betting the company entirely on AVS demand materializing. Whichever way that founding thesis eventually resolves, the version of restaking being sold to users in 2026, verifiable cloud services, Bitcoin security exports, tokenized real-world assets, is a considerably wider basket of bets than the one being sold in 2023, and not every protocol in this piece will still describe itself as a restaking protocol by 2027.

Frequently Asked Questions

What is restaking in crypto?

Restaking is the practice of using already staked ETH, or a token that represents it, to secure additional services beyond Ethereum itself, such as oracles, bridges, or rollup sequencers, in exchange for extra yield and extra slashing risk. EigenLayer popularized the term starting in 2023, and the sector has since grown to include Ethereum-based competitors like Symbiotic and SSV Network, plus non-Ethereum variants like Babylon, which applies the same basic idea to Bitcoin.

Is restaking safe?

Restaking adds risk on top of ordinary staking rather than removing it. Beyond normal validator slashing, a restaker is exposed to the security of whatever outside service their stake secures, the smart contract risk of the restaking protocol itself, and, for liquid restaking tokens, the risk of any bridge or wrapper used to move that token between chains. The April 2026 Kelp DAO incident, where a bridge exploit rather than a restaking failure caused roughly $292 million in losses, shows that the biggest real world losses so far have come from the infrastructure around restaking rather than from its core slashing mechanism.

What is the difference between staking, liquid staking, and restaking?

Staking locks ETH to help secure Ethereum and earn a base reward, currently in the high 2% annual range. Liquid staking, through tokens like stETH or rETH, does the same thing but issues a tradeable receipt token so the staked capital is not fully illiquid. Restaking takes that staked ETH, or a liquid staking token, and pledges it a second time to secure a separate service beyond Ethereum, adding extra yield alongside extra slashing conditions tied to that outside service’s own security.

What happened to Karak crypto?

Karak, one of the earlier challengers to EigenLayer in the restaking sector, rebranded to OpenGDP in July 2026. Its messaging shifted away from restaking and shared security entirely, toward tokenizing real world assets and stablecoin settlement infrastructure. Its older staking and bridge products still exist and remain accessible, but the project’s own homepage and public positioning no longer describe it as a restaking protocol.

Which restaking protocol has the most TVL?

EigenLayer, now rebranded EigenCloud, remains the largest restaking protocol by total value locked, though the exact figure varies significantly by tracker and week, ranging from the mid single digit billions to the mid teens of billions of dollars in mid-2026. Symbiotic is a distant second at roughly $1.6 billion to $1.7 billion, followed by Bitcoin-focused Babylon at about $5.6 billion in staked BTC value and SSV Network, which reports more than $16 billion in ETH secured through its distributed validator network, a broader figure than pure restaking TVL alone.

Written by the HOGE Wire markets desk.