EigenLayer is a new protocol, or set of instructions, that sits atop the main Ethereum blockchain. It offers a way for those who have already invested tokens by means of staking—a blockchain feature that rewards those who lock up their tokens to secure the network—the ability to “restake” those tokens. The term is what it sounds like. By restaking, users take Ethereum tokens they have already invested and then invest them again. But instead of receiving a reward from the Ethereum network, as happens with staking, those who restake get paid by other projects piggy-backing of the blockchain’s existing security pool.

Restaking has existed for some time but EigenLayer is seeking to turbocharge the process by adding a lot more participants. The protocol aspires to be something akin to an Amazon Web Services for cryptoeconomic security. 

The implications of this are “profound,” a16z wrote in a blog post about the company, noting that EigenLayer has the potential to accelerate innovation on Ethereum “100x faster.” While this appears to be a promising development for the blockchain, EigenLayer has yet to be tested at scale. And while it could result in a burst of innovation for Ethereum, some point to a major potential drawback: A new concentration of risk that, in a worst case scenario, could harm both users and the broader blockchain eco-system. Should Ethereum users view EigenLayer with excitement, fear or both?

‘Opens up a wormhole’

The arrival of EigenLayer is a boon both to investors, who can now more easily earn an additional slice of yield, and for developers who can more quickly add a trust layer in the form of collateral. The issue, however, is that using the same pool of collateral to secure multiple aspects of the blockchain creates new vulnerabilities.

“It opens up a wormhole,” Rushi Manche, co-founder of Ethereum layer-2 Movement Labs, told Fortune. Manche raised a troubling hypothetical posed by EigenLayer critics: What happens if there’s a hack or a bug in the restaking smart contract? How does that affect the wider ecosystem which is also relying on this single deposit? 

At the same time, Manche noted that proponents say ”an entire new design space has opened up” for developers thanks to EigenLayer, allowing the same asset to be stretched much further, and achieve much more. From this view, the rewards far outweigh the risk. It’s also worth considering the drawbacks if tools like EigenLayer are not available.

Specifically, in the absence of restaking, decentralized applications built on top of Ethereum must create their own proof-of-stake token—a high bar of entry for any start-up. With restaking, protocols can dip into the main blockchain’s almost $95 billion of staked Ether for security. To revisit the cloud computing analogy: If hackers want to attack small businesses, they have to get past the likes of Google Cloud or AWS on which those busiensses rely. EigenLayer proponents make the same case: hackers need to get past a security system created by Ethereum, which has worth more than triple the market cap of Nike. 

But, what happens when the same collateral gets reused again and again? 

“EigenLayer reminds me of the 2008 recession: under collateralize, over leverage. That’ll go well,” one X user posted— just one of the industry onlookers who have drawn parallels with the financial crash, where banks rehypothecated subprime mortgage loans.

Austin Campbell, the former head of portfolio management at Paxos, told Fortune that the comparison isn’t trivial. “At the core of every levered system is the simple truth that the more leverage you use, the less stable your system is,” he said. 

There is a risk of “cascading failures” if one project fails, agrees Omid Malekan, an adjunct professor at Columbia Business School. But he and many experts were also quick to offer nuance to the 2008 parallels: restaking doesn’t involve rinse and repeat borrowing and lending. Instead, it’s taking a locked-up asset, and locking it up again. Instead, think of it more like reusing the security deposit on an apartment to secure further apartments, suggests Jack O’Holleran, CEO of Skale, an Ethereum Layer-2. 

But then, what if you trash both apartments?

One scenario of so-called apartment-trashing would be if those operating the restaking services—known as AVS for “Actively Validately Services”—choose to devalue the restaked tokens. “There could be this ‘house of cards’ effect, but backwards,” says O’Holleran. 

The debate is analogous to the criticism of liquid staking provider Lido, which currently accounts for 28% of all Ethereum being staked. In theory, Lido introduces a “centralized point of failure” which is “obviously concerning,” Tekin Salimi, founder of dao5 Capital and seed investor in EigenLater, told Fortune. But in his view, rather than barring this innovation, diversifying the liquid staking providers is the solution. 

Noting these concerns, EigenLayer says on its website that it includes a “veto committee” to contain slashing—the term used in the Ethereum world for destroying the collateral of validators who fail to fulfill their duties—  and to prevent contagion to the mainnet or other protocols. It will serve as a “doubly-trusted intermediary between AVS and staker.” So, slashing conditions fall to each AVS, instead of the overall protocol’s discretion. But what this will really look like in practice, is unclear. To that effect, slashing risks remain “the most real and yet the most undefined” of all those raised, thinks Vance Spencer, co-founder of crypto venture capital firm Framework Ventures.

‘Don’t overload the system’ Ethereum founder warns

EigenLayer has given rise to specific security concerns, but also to a more broader existential one: what fraction of staked tokens can be concentrated on a separate protocol on Ethereum, which holds up  decentralization as its paramount value? According to O’Holleran, empowering EigenLayer amounts to the Ethereum community giving considerable leverage and power to another system. Furthermore, copycat restaking services are also popping up, such as Karak, which announced a $48 million Series A raise last month. In addition, an unnamed but similar version is soon coming to market, sources close to the matter told Fortune. 

Vitalik Buterin, Ethereum’s founder, raised concerns about EigenLayer in a blog post last May. “Be wary of application-layer projects taking actions that risk increasing the “scope” of blockchain consensus to anything other than verifying the core Ethereum protocol rules,” he wrote. Buterin warns of the “slippery slope” of the underlying network’s “fragile” social consensus losing credibility. His message was clear: Don’t overload the network’s consensus.

One of the concentration risks of top concern to Buterin is overloading the system’s security. 

The strength of Ethereum’s security system is directly correlated to the economic value of the network—it would be prohibitively expensive for hackers to muster the resources to defeat the consensus system that protects  the network. “We agree that Ether is being used to secure the main net network, and that is the only purpose thereof,” says Campbell. But if you are using your staked tokens to secure Ethereum and other networks, then the surface of things that Ether is being used to secure grows. So, the value resting on any single point gets larger, explains Campbell, which may provide more attractive incentives for hackers.

The second concentration risk concerns decision-making. Let’s say 40% of Ethereum’s tokens eventually becomes restaked and EigenLayer’s smart contract experiences a hack or bug. It’s unclear when there would be consensus to rollback Ethereum to release it, explains Tom Schmidt, partner at crypto venture capital firm Dragonfly. “I don’t know where that threshold is,” he admits.