Actively Validated Services
Actively Validated Services
How EigenLayer is unshackling crypto entrepreneurship
Hey there,
NFTs, DAOs, and DeFi have each been fundamental primitives in the gradual disruption of finance. Shared security is another key component in this evolution. EigenLayer has emerged as a key player in this ecosystem, and their token is set to become transferable on the 30th of September.
We see this as the “AWS moment” for Web3—a pivotal point where the cost of crypto-economic security will decline, similar to how server costs fell in the mid-2000s. Today’s piece offers a breakdown of how AVS’ work and the reasoning behind our beliefs.
This is one of many sponsored pieces we will be releasing in collaboration with EigenLayer in the coming weeks. As always, if you’re a founder working on building an AVS, reach out using the form below.
Joel
Civilisation advances by extending the number of important operations which we can perform without thinking of them.
Alfred North Whitehead
In late 2002, eight people attended a tech conference hosted by Amazon at its headquarters in the old Pacific Medical Center in Seattle, Washington. Despite this unremarkable attendance, the meeting marked a turning point in the fate of the company, the economics of start-ups, and, arguably, the very course of capitalism. It was the day Amazon launched the first version of Amazon Web Services (AWS).
Here is a non-exhaustive list of things you needed to start an internet company in the 1990s and early 2000s: physical servers, networking equipment, data storage, software licences for databases and operating systems, a secure facility to house hardware, a team of system administrators and network engineers, and robust disaster recovery and backup solutions. All of this cost at least $250,000 and took several months to a year to set up.
Astonishingly, these infrastructure expenses had virtually zero correlation with a company’s unique product or service. You could be building a pet store or a social media network, and you would have to go through the same process from the ground up. By some estimates, 70% of engineering hours were spent building and maintaining data centres, and only 30% were spent on the actual business.
By introducing cloud computing, AWS fundamentally altered the economics of running a start-up. It eliminated the need for upfront investment of time, effort, money, and personnel through a flexible pay-as-you-go model. Transforming infrastructure from a capital investment into an operational one enabled small teams with revolutionary ideas to launch rapidly and validate their thesis. Many of these ended up becoming the Stripes and Airbnbs of the world.
Around the same time, an anonymous programmer known as Satoshi Nakamoto was altering the fabric of capitalism in a different way. By figuring out how to get computers distributed across the globe to agree on a common truth without having to trust each other, Satoshi solved a problem that had perplexed computer scientists for decades. It was a pivotal, zero-to-one innovation in the history of technology.
While Satoshi’s Bitcoin primarily used this trustless distributed system to maintain a payments ledger, Vitalik Buterin created Ethereum, expanding it to support any general-purpose computation. Over time, other use cases for this system started emerging—from decentralised storage networks like Filecoin to oracle networks such as Chainlink that securely provide real-world data to blockchains.
However, setting up such a decentralised network from scratch is similar to starting an internet company pre-AWS—costly, resource-intensive, and often tangential to the network’s core problem. And given that many of these networks handle real money right from inception, the consequences of errors are catastrophic.
When a problem starts affecting a sufficiently large number of people, solutions emerge. Amazon made it easy to start an internet company, and now the EigenLayer team is doing the same for those seeking to build trusted distributed computer networks. Each network built on EigenLayer is termed an Actively Validated Service (AVS), the topic of this article.
But before getting into what an AVS is, let’s first understand why it can be so difficult to bootstrap a distributed network in the first place.
The Challenge
Let’s revisit the problem—you have a global network of computers, each operating independently, and you need to establish consensus on a shared truth without them trusting each other. This truth could be anything— the balance of a token in an account, the stock price of NVIDIA, the result of a complex computation, or the availability of a file on the network.
Nodes in these networks could have an inherent incentive to manipulate the truth, such as falsely displaying a token balance as higher than it actually is. However, as long as most of the network agrees on the actual truth, the malicious actor can be ignored. The situation gets precarious when a majority of the network agrees on a state that deviates from the truth. In such a case, the network is compromised.
Satoshi ingeniously combined concepts from cryptography and game theory and created Bitcoin's proof-of-work system to solve this problem. Most networks today use a variation of PoW, proof-of-stake (PoS), with four key elements:
Cryptography: to thwart impersonation and ensure data integrity and authenticity across the network.
The Carrot: Genuine participants (validators) receive financial incentives through transaction fees from users and newly minted tokens from the network.
The Stick: Malicious actors face economic penalties. Validators must deposit a stake of the network's native token to participate. If they act maliciously, their stake may be destroyed (slashed).
Strength in distribution: A larger number of validators with a well-distributed stake makes the network more resilient to attacks.
PoS networks allow regular users, who may not directly validate the network, to delegate their stake to validators and earn a portion of the rewards validators receive for their role. However, this approach also exposes users to the risk of their stake being slashed if the chosen validator acts maliciously.
Protocols on certain blockchains, such as Ethereum and Solana, offer stakers a liquid token in exchange for staking the native token (Lido, for instance, provides Ethereum stakers with the stETH token). This derivative asset is termed a Liquid Staking Token (LST). I have explained the concepts of staking and LSTs in a previous article.
Given this context, put yourself in the shoes of a team who wants to build a PoS network from scratch.
You will have to start by looking for a set of validators—actors who have both the technical expertise and hardware to become part of your network. Where might you find such individuals? Likely on Discord and X. However, to really capture their attention amidst a sea of competing projects, you will need to either execute a marketing masterclass or secure substantial VC funding.
Once you have their attention, you have to convince them to become part of your network. Now, this is easier said than done. Remember, validators must either stake their own capital as collateral or exert the effort to attract stake from others. Given that your network is in its infancy, your token is probably not very valuable. Why would validators risk acquiring a token that could plummet in value at any moment, especially when they are likely already exposed to the volatility of assets from other networks they validate?
Your best bet would be to sweeten the carrot: offer validators (and stakers) superior rewards to compensate for the increased risk. This explains the high Annual Percentage Yield (APY) for staking in nascent networks. However, there is a caveat: high emissions function as an indirect expense on the overall network, potentially diluting the value of the token.
Even if you navigate these challenges, there is a strong likelihood that your project's initial stages will have fewer validators than ideal. This scarcity of validators reduces security, exposing your network to the risk of a majority attack.
Beyond these fundamental issues, you also need to plan or account for other factors such as the geographic diversity of validators, creating secure and audited client software, and, depending on your project’s specifics, infrastructure elements, including data availability, transaction ordering, finalisation services, and block proposing.
Much like internet startups pre-AWS, these steps are time and resource intensive, and none of them directly correlates with the core problem your network tries to solve.
Security-as-a-Service
Recently, I explored how the internet spawned a new generation of businesses (platforms) that generated value by efficiently connecting supply and demand. In the scenario we just discussed, we have a set of validators—the supply—who want to make money providing technical services but mitigate the financial risks of doing so. The demand side comprises aspiring blockchain protocols in search of trusted, reliable validators to secure their networks.
EigenLayer has emerged as a platform that bridges this gap, connecting validators (termed “operators”) with networks looking for their services (termed “Actively Validated Services” or AVSs).
Now, put yourself back in the shoes of a new protocol developer.
First, EigenLayer provides a pool of trusted validators (termed “operators”) who have committed to validating multiple services, including nascent ones. This addresses your initial challenge: Where do you source reliable validators?
Second, and perhaps EigenLayer’s biggest unlock, is decoupling the carrot from the stick. Operators do not need to stake your native token to secure your network. EigenLayer requires them to deposit (or attract stakes of) established assets such as ETH and LSTs. In the event of malicious behaviour, these assets would be slashed.
This decoupling means that stakers and operators can avoid the risk of holding additional, newer tokens. Instead, they can earn supplementary returns on established assets they're already comfortable holding. (Saurabh's analysis of intersubjectivity explains how EigenLayer enhances capital efficiency).
From the protocol's perspective, this model eliminates the need to generate token emissions (and potentially inflate your token) to compensate validators. Instead, you can benefit from the far more robust security assurances backed by ETH as collateral. In fact, this flexibility even allows you to not issue a token at all if you do not want to!
Third, you get to curate your operator set based on your specific product’s security needs. Before integrating them into your network, you can weigh factors such as a validator's technical capabilities, secured stake, geographic location, and track record in securing other networks. Compared to the daunting task of building a network from scratch, this level of selectivity is not short of a luxury.
While one security risk diminishes, another emerges—the reliance on EigenLayer itself. However, EigenLayer is not a separate blockchain but a set of smart contracts deployed on Ethereum. Ethereum boasts over 6,000 nodes and $86 billion in capital backing it. While smart contract risk persists, Ethereum itself is as safe as a blockchain can get.
You might be wondering: what about the carrot? How do the economics of building on EigenLayer work? Protocols can reward operators and delegators in any ERC-20 token. In practice, this gives an AVS two options:
Distributing rewards in established tokens like ETH or stablecoins. In this case, the relationship between operators and the AVS is transactional—operators provide a service, and the AVS pays them in a widely accepted currency. EigenDA, the first AVS, kicked off operator rewards by distributing ETH to operators and delegators.
Distributing rewards in their own token. This more closely resembled the economics of a traditional crypto network. While this model gives the AVS flexibility to pay for security via token emissions (rather than the direct expense of ETH/stables), they also have to convince operators that their token will hold value. If they do not do this, it will be difficult to attract operators who do not end up selling the AVS token as soon as they receive it.
Initially, 10% of AVS rewards will go to operators and the rest to stakers. This parameter will be made flexible in the future. Further, to “strengthen incentive alignment”, EigenLayer plans to distribute an amount equivalent to 4% of the initial $EIGEN supply as incentives to delegators and operators to participate in the network.
EigenLayer’s compelling value proposition has attracted a sprawling variety of projects seeking deployment as AVSs. The roster includes the usual suspects—projects with established needs for operator services such as rollups, data availability services, bridges, oracle networks, and sequencing layers.
However, given that operators can theoretically support any type of computation (not limited to state transitions), we are witnessing the emergence of many innovative and experimental projects leveraging EigenLayer. These include DePin networks, AI inference engines, zero-knowledge proof coprocessors, privacy-oriented protocols (including TEE, FHE, MPC), zkTLS networks, and even a policy engine for smart contracts.
The Great Unshackling
Earlier, I made a somewhat bold claim that AWS changed the nature of capitalism. Allow me to back it up.
Pre-AWS, the high capital required to launch a company meant that founders either self-funded or secured investment from external sources (friends, family, venture capital). This financial barrier effectively excluded most of the global population from internet entrepreneurship, relegating it to an endeavour exclusively for the wealthy or those in privileged geographies.
By dismantling these constraints, AWS not only streamlined the process for existing entrepreneurs—a relatively small cohort—but also unleashed the creativity and imagination of many who previously deemed starting a company beyond their reach. This democratisation catalysed a surge in start-up experiments. While many failed, the ones who succeeded drove an unprecedented rise in economic productivity and human convenience.
From an individual entrepreneur’s point of view, cloud computing opened up a range of options—from attempting to build the next billion-dollar enterprise to taking someone like Pieter Levels’ path and making millions a year as a solo developer. Or anything in between.
We are excited about EigenLayer and AVSs because they unlock similar opportunities for trustless distributed networks. Have an idea that requires more than one computer to operate without mutual trust? You can now swiftly implement it with an AVS.
From governance chains to zkTLS networks, we are witnessing various experiments that might have been unfeasible previously. As more entrepreneurs recognise the dramatic reduction in human and financial capital required to establish such systems, we anticipate an explosion of more experiments.
Most will fail.
Some will chart the course of this industry’s future.
Playing around with ChatGPT’s new voice mode,
Shlok Khemani