ERC-4337: Wallets Are Getting a Whole Lot More Interesting

The Important Stuff

• ERC-4337 is a new standard built on top of the Ethereum Virtual Machine ("EVM") that enhances smart contract wallets on Ethereum-based blockchains.
• The introduction of ERC-4337 is expected to popularize a new type of wallet experience called embedded wallets, which are smart contract wallets built on a set of APIs that individual applications can use to customize the wallet experience for their users.
• Some of the less obvious implications of ERC-4337 include handing over more control over private keys to third party entities, changing the current MEV landscape, and potentially, helping applications with user conversion and retention.

Why Is ERC-4337 Needed?

ERC-4337 is a new standard built on top of the EVM that will likely have profound implications for wallets and the wider crypto industry.

Wallets are a fundamental component of blockchains as they allow users to move assets on-chain and interact with web3 applications. There are currently two different wallet types on Ethereum and most other blockchains — Externally Owned Accounts (“EOAs”) and Smart Contract ("SC") accounts.

EOAs is a piece of technical jargon that refers to accounts controlled by private keys that can't be changed. Wallets powered by EOAs include products that you’re probably familiar with, including MetaMask, Phantom, and the recently launched Uniswap Wallet.

On the other hand, SC accounts have more expressive and programmable rules as they're controlled by smart contract logic. If you’re using a wallet that supports SC accounts you can define your own set of conditions in which transactions are valid, such as giving multiple entities the ability to authorize transactions instead of relying on a single private key. Argent and Safe are examples of popular SC wallets.

With that background out of the way, let’s talk a bit more about ERC-4337. 

Existing SC wallets still rely on EOAs to initiate transactions and pay for gas. While certain SC wallet providers built their own custom infrastructure to deal with this, it required a lot of effort and resulted in fragmented implementations. 

The idea behind ERC-4337 is simple: level the playing field for developers building SC wallets by standardizing some of the off-chain and on-chain infrastructure required by SC wallets to operate. In short, ERC-4337 will make it easier for developers to build SC wallets.

Initially, we weren’t planning to write a piece on ERC-4337. But after speaking to wallets teams, embedded wallet SDK providers, and people in the MEV space, we got the sense that many people in the industry are still a bit confused about the first and second-order effects of ERC-4337. 

This piece is an attempt at thinking through some of the immediate market implications of ERC-4337, in addition to some of its less obvious potential. While we still have more questions than answers, this is an attempt to think out loud about what’s a pretty wonky topic.

Immediate Implications of ERC-4337

A New Kind of Wallet

The wallets we use today, whether EOAs like Metamask or SC wallets like Argent, are all personal wallets. In a nutshell, they’re frontends built for users to interact with blockchains.

The downside to personal wallets is they're designed in a way that makes it hard for application developers to hide the technical details of how they work from users. As a result, users are forced to learn how personal wallets, and in turn blockchains, work before they can even use an application. When people say “for most people crypto is hard to use,” this is a big reason for it.  

For example, let’s take a user attempting to play a blockchain game for the first time — they’ll have to create a personal wallet, connect that wallet to the game, learn that they need tokens to pay for transaction fees, find a way to onramp fiat to buy those tokens, and play the game while having their wallet extension pop-up constantly asking them to “confirm this transaction" for any action they take. Talk about bad UX! 

As a solution to this crappy UX, embedded wallets are emerging to remove some of the points of friction that exist with personal wallets. Unlike personal wallets, embedded wallets are invisible to users as they come with a set of APIs rather than a dedicated frontend.

Application developers will be able to use these APIs to integrate wallet functionality directly into their user interfaces, giving them more creative freedom to experiment and customize the wallet experience for their applications. 

Going back to our earlier example, crypto game developers will be able to let users spin up a pre-funded custom-branded wallet so they can go from having no wallet and no money to playing a blockchain-based game as fast as it takes to create an email account. 

Under the hood, this would be powered by session keys (to pre-approve transactions users would normally need to make) and multicalls (to batch several pre-approved transactions into one bundled transaction).

Connecting Personal and Embedded Wallets

We suspect personal and embedded wallets may be quite complementary. Their relationship could resemble the hub and spoke model — embedded wallets will allow users to have an account on each app (spokes), while personal wallets (hubs) will unify all these accounts into a safer account management and identity layer. The combination of both will allow apps to customize their wallet UX without requiring users to have dozens of unconnected wallets (since the embedded wallets will be able to settle to your personal wallet).

We also suspect it’s unlikely that app developers will want to build embedded wallet middleware in-house as it would require a lot of work. Instead, they will likely outsource this to specialized embedded wallet connectors and SDK providers such as Biconomy and 0xpass which, in turn, could leverage specialized infrastructure that facilitates the use of ERC-4337 like Stackup. 

We see embedded wallet middleware companies providing prepackaged APIs that applications can use to embed wallets into their apps, in addition to making them compatible with existing personal wallets. We’re also seeing existing personal wallet providers build their own embedded Wallet-as-a-Service products (e.g., Coinbase, MetaMask, and Safe announced their own embedded wallet SDKs).

Below is a graphic that shows how we're currently thinking about the relationship between personal wallets, embedded wallet middleware, and embedded wallets.

Personal Wallets vs Applications with Embedded Wallets

The wallet wars may come sooner than you think, and one battleground we see emerging is applications using embedded wallets going against personal wallets.

To make this more concrete, applications like OpenSea may want to compete with a personal wallet like MetaMask to be the primary touch point for users by offering a superior account management experience with customized embedded wallets. 

On the flipside, personal wallets like MetaMask may want to keep users on their platforms by giving users access to applications like OpenSea through them.

What all of this means is that certain applications may be motivated to sidestep personal wallets by giving their users customized embedded wallets with a self-custody experience that rivals what personal wallets provide; in doing so, application teams will be able to own the user relationship in a much more direct way. 

All of the above is speculative, but if you follow the motivations of each player, we can’t help but think this is where things are directionally headed. After all, right now, popular CEX, DEX, and NFT marketplaces are the main onboarding funnels for users getting into crypto — why give away your close relationship with the user to personal wallets when you can own users more holistically?  

Less Obvious Potential for ERC-4337

If the above section was speculative, this section is doubly so — here, we’ll be trying to walk a hundred steps ahead while blindfolded. What are some of the less obvious things ERC-4337 can bring?

Outsourcing Security to Third Parties

To abstract key management away from users, third party entities such as existing web2 companies may have to handle some keys on behalf of users.

For example, users may need to give social login providers like Google certain permissions over their SC account. The same goes for providers of multi-factor authentication and social recovery. This is great for adoption because end-users will feel safer knowing their accounts can be backed up easily, giving them a comparable experience to what they're used to in web2 today. 

That’s not to say ERC-4337 will get rid of self-custody, because it won't. Users will ultimately retain control over their accounts as they'll likely be able to revoke the permissions granted to third party services at any time. Zero knowledge tech could even allow users to use third party services to control their accounts without having to hand them any keys. 

So even if ERC-4337 could mean ceding some control to third parties to improve the self-custody experience for users, they'll still be able to retain full control over their accounts with the added benefit of being able to tailor their accounts' security set-up to meet their needs, with more granular account permissioning and control.

MEV Capture

We expect ERC-4337 to change the current MEV supply chain, impacting ownership of orderflow and who gets to extract its value. This part is a bit wonky, so if you’re not familiar with MEV, feel free to skip this section.

ERC-4337 allows users to submit user operations (aka “meta-transactions”) to alternative mempools for inclusion in a transaction. These user operations are bundled into a transaction by a Bundler who initiates the transaction and is reimbursed for paying gas for the transaction by users. Paymasters (a third party that can reimburse Bundlers for transaction fees on behalf of users) are an additional optional step for transaction sponsorship.

These two new actors — Bundlers and Paymasters — could be run by a number of market actors such as searchers, builders, and Order Flow Auctions (“OFA”) providers, who are able to extract MEV from user operations (they’ll get a sneak-peek into orderflow before anyone else).

Personal wallets and applications using embedded wallets may want to build their own OFAs to capture the MEV from user ops, similar to how Robinhood sells its orderflow to market makers and large trading firms. In the same vein, personal wallets and applications using embedded wallets could use the MEV they capture to subsidize gas fees for their users.

ERC-4337 will also open the door for experimentation with mempool design as it allows for the creation of alternative user op mempools with different rulesets. For example, wallets could run their own user ops mempool that only shares partial data on incoming user ops to their OFA to reduce harmful MEV.

Below is an example of what the MEV supply chain could look like with a custom user op mempool that only shares partial information on incoming orders.

The visualization above is just an example that shows how ERC-4337 could be used to change the MEV landscape and make it fairer to users.

This higher level, alternative mempool layer could potentially be used by Bundlers to match intents (user ops) before they hit the chain to make settlement cheaper and more efficient. This would be a bit similar to how the Clearing House Interbank Payments System's clearance system continuously matches, nets, and then settles payments via Fedwire. 

Web2 Meets Web3

Web2 apps have struggled to connect to the web3 ecosystem and reap its benefits because they have to build their own infrastructure to solve a number of issues (e.g., wallet integrations, chain support, and on/off ramps).

One of the most exciting new use-cases of ERC-4337 is the ability for existing web2 apps to embed wallets in-app. This will let the mainstream applications we use today easily connect to the blockchain ecosystem and integrate crypto-native features, thereby converting their users into web3 without them having to manually spin up wallets.

For example, payments companies like Revolut could use embedded wallets to allow their users to convert fiat into stablecoins just as they let their users exchange between fiat currencies today.

Wallets as a Replacement for Third Party Cookies

ERC-4337-enabled SC wallets may be used as a replacement for third party cookies to help applications convert and retain users. 

By using embedded wallet middleware to create wallets on users' behalf and add metadata to them, for example, apps could use click-through ads to seamlessly onboard users and track where they came from. This would also make identifying single users on their platforms much simpler as they could map different wallets to the same email address. 

ERC-4337 may also make it easier for apps to engage with existing users and resurrect activity. For example, DEXs may be able to more accurately track how much a user trades and issue fee rebates or dynamic fees based on how much volume they contribute, and, more generally, may be able to create loyalty programs to retain and incentivize users.

While this may seem far out there initially, companies such as Garden Intelligence are already emerging to help applications better track their users. Garden acts as a wallet oracle that application developers can use to better understand their users and inform their user acquisition and retention strategy. 

Conclusion

ERC-4337 is an extremely exciting development for the crypto industry as a whole. It will allow wallets to take many different forms, kickstarting a new wave of experimentation that will unbundle and rebundle wallets in many different ways to bridge the gap between web2 and web3 UX. 

While no one really knows how ERC-4337 will affect market dynamics long-term, it's clear that application developers should think carefully about how ERC-4337 will impact them and how they can leverage the new use-cases that come with it to build better products. If you don’t, you risk losing to applications that see where the puck is headed before you do.

If you're building in this space and want to brainstorm how ERC-4337 may impact you, the industry, and/or to tell us why we’re thinking about it in the wrong way, please reach out to us. We'd love to chat.

Many thanks to Myles O'Neil, Ankit Chiplunkar, Rohan Garg, Daniel Marzec, Evan Weiss, John Rising, Trevor Aaron, Adi Sidapara, Sam Hart, Barry Plunkett, and Larry Sukernik for their feedback on this post.