Restaking is the practice of taking ether that is already staked to secure Ethereum and pledging it again to secure other services, in return for extra rewards. EigenLayer, which launched in June 2023, built the market for it, and by 2024 it had become one of the largest protocols on Ethereum, DL News noted. Much of the money arrived through points programmes that promised future airdrops to anyone who deposited.

Two years later the points are gone and most of the money has followed. At AnkhLabs we think restaking is the clearest case study of this cycle for anyone planning incentives, because it shows exactly what points can buy and what they cannot. This article covers where restaking stands, what happened after the points era and what founders should take from it.

What is restaking, and how does it work?

When you stake ether, you lock it to help run Ethereum and earn a yield. Restaking lets the same ether also back other networks or services, which EigenLayer called actively validated services, or AVSs. If the operator running your stake misbehaves, part of it can be taken away as a penalty, a process called slashing. EigenLayer switched slashing on in April 2025, Blockhead reported.

Most users never restaked directly. They used liquid restaking protocols such as ether.fi, Renzo and Kelp, which took deposits, restaked them on EigenLayer and handed back a tradable token. Those tokens could then be used across DeFi, which is how one deposit came to carry several layers of reward and several layers of risk.

How big is restaking in 2026?

Far smaller than at its peak. These are DefiLlama's figures for the main protocols.

Source: DefiLlama. Dollar values move with the price of ether.

Ether itself traded at about $2,234 on 6 November 2026, about 55% below its all-time high of $4,946 on 24 August 2025, according to CoinGecko. That accounts for a large part of the fall in dollar terms. EigenCloud's deposits fell about 74% from their peak, though, so a good share of the ether left as well.

The tokens fell further. As of 6 November 2026, CoinGecko showed EIGEN about 97% below its December 2024 high, ETHFI about 93% down and REZ about 99% down.

The economics explain why. On 8 September 2026 the restaking category on DefiLlama held about $10.02 billion and had earned $99,977 in fees over the previous week, while liquid staking held $51.87 billion and earned $27.35 million, roughly 53 times more per dollar, according to CoinDesk figures reported by Blockhead. The five largest remaining liquid restaking tokens made a combined $953,350 in gross profit in the second quarter of 2026, down from $2.18 million three quarters earlier. Puffer, which raised $23 million, made $21,590 in the quarter.

What happened after the points era?

The warning signs were visible early. In March 2024 the ether.fi chief executive Mike Silagadze said "anybody that has the word AVS or restaking is just getting money shoved down their throat", as DL News recalled. After EigenLayer announced its airdrop and the backlash began, nearly 14,000 wallets triggered withdrawals from 29 April, The Defiant reported. By July 2024 EigenLayer's deposits had fallen $2.28 billion in 30 days to $15.1 billion, with Renzo down 45%, CoinDesk reported. Depositors were moving on to farm the next set of points, and restaking yields of about 3.4% looked thin next to rivals paying more than 10%.

Points programmes wound down through 2025, and slashing added a penalty without adding yield. In July 2025 Eigen Labs laid off engineers and pivoted toward EigenCloud, which it pitches as a way to verify the actions of AI agents. In December the Eigen Labs researcher known as Kydo wrote that restaking "did not turn into the 'everything layer' some people, including myself, once hoped. It also did not vanish." He said the target market had been too niche and that the hype had attracted weaker developers, according to DL News.

The decisive move came from one of its best-known partners. In August 2026 ether.fi removed restaking from weETH, its main token, and offered anyone who still wanted restaking a separate token built on the rival platform Symbiotic. Less than 1% of its assets were still restaked by then. "There were no meaningful yield opportunities in restaking and there was some perceived risk from stakers," Silagadze told CoinDesk, as quoted by Blockhead. He was more personal on X, quoted by The Defiant: "End of an era. Sad."

The risk was not theoretical. On 18 April 2026 an attacker drained about 116,500 of Kelp's restaking token rsETH, worth roughly $292 million, through Kelp's cross-chain bridge, The Block reported. The unbacked tokens were used to borrow real ether on Aave, which froze its rsETH markets, and the AAVE token fell about 10%. A token that bundled staking, restaking and bridging carried the risks of all three.

Where does restaking go from here?

Restaking has not disappeared. It has narrowed. EigenCloud still secures about $5.7 billion, and its listing on DefiLlama now describes Eigen Labs as "an AI research lab building tools for human and agent coordination", which is a long way from the 2024 pitch of extra yield on ether. Symbiotic is smaller but now hosts ether.fi's opt-in restaking token. The second-largest protocol that DefiLlama files under restaking is not on Ethereum at all. It is Babylon, which lets holders stake bitcoin to help secure other networks and held about $3.1 billion on DefiLlama as of 6 November 2026.

What survives is the version with a paying customer. Restaking makes sense where a network or service will pay real fees for borrowed security, and it struggles wherever the reward is mostly a token that the protocol itself prints. For anyone launching a product in this space, who pays for the security should be the first question the marketing answers.

What are the marketing lessons from restaking?

Restaking did not fail because the idea was foolish. It failed as a growth story because the marketing ran years ahead of the revenue. These are the lessons we would take into any incentive plan.

• Points rent deposits, they do not buy them. Every major exit followed a change in rewards. If your growth chart is mostly farmers, it will reverse when the programme ends. Our guide to crypto points programs covers how to design for that, and our piece on sybil detection explains how to see who is really there.

• Market the yield you can show. Deposits are a vanity number when fees are close to zero. Publish what users actually earn and where it comes from, and say so when the answer is "mostly our own token".

• Price in stacked risk. Each extra layer of reward added a layer of risk that most users did not read. Clear risk disclosure is slower to sell but cheaper than an incident. Our article on exchange security after the Bybit hack covers how trust is lost and rebuilt.

• Airdrops end campaigns, not communities. The day a token goes live is the day farmers decide whether to stay. Plan the months after it with the same care as the months before, as we argue in our airdrop marketing strategy guide.

• Pivot in public and with numbers. ether.fi explained its exit with a reason and moved into a card and borrowing business. Card fees rose from 17% of its monthly revenue in January to 46% in July, according to Blockhead. A pivot backed by figures is easier for a community to accept than a quiet rebrand.

What we'd tell a founder planning incentives

First, tie rewards to staying, not to arriving. Our Mythical Forest programme was a points-driven engagement programme powered by the $MYTH token, with a social farming system of referrals and rewards tied to sustained activity, and it onboarded 1M+ users with 10M+ campaign impressions. The design choice is the point. A reward that pays for week ten keeps people until week ten. More examples are in our case studies.

Second, write the end of the programme before you launch it. Decide what users get when points stop, what you will say on that day and which numbers you will report afterwards. Our strategic advisory work often starts with that exit plan.

Third, report retained deposits, not peak deposits. A peak flatters the campaign. The share still there 90 days after rewards end tells you whether you built a product.

If you are designing a points or rewards programme and want a second opinion before it goes live, talk to AnkhLabs.

The bottom line

Restaking is still running, with about $5.7 billion on EigenCloud, but the idea that it would become a new layer of the internet has faded, and its own builder now markets AI verification. The points era showed that incentives can gather tens of billions of dollars quickly. It also showed that money gathered for points leaves for points. Founders who remember the second half of that lesson will spend their incentive budgets very differently.

This article is for information only and is not investment advice. Figures are as of 6 November 2026 unless stated.