We are entering the age of a thousand little blockchains
To quote Fleetwood Mac: (App) Chain Keep us Together
The world is a giant cyclic joke. Fashion designers bring back baggy jeans every decade; atrocities rejected by a whole generation always gain new audiences later (nipplegate); my TikTok is so full of witchcraft that I fear the puritans might sneak in at night and burn me alive.
And it must be early 2022 again because Twitter is once again seeing Tom Brady’s retirement and Avalanche bulls. With its new high-profile customers flocking to its subnet, Avalanche brough application-specific chains (aka “app chain”) back to the public’s attention.
NOTE: I’m using “app chain” as a general term for blockchains/rollups designated for the purpose of 1 application or 1 use case. The concept is simple: as opposed to building on a large blockchain, start a blockchain just for your own purpose! Your blockchain can communicate with a larger blockchain (for security or for data sharing purposes), but your app and your ecosystem alone runs on it.
Here’s the thing: app chains have huge potential and we might see huge adoption in the future. But before we get started, let me tell you that this is an article for a more expert crowd. For everyone else, I’ll be back with dumb NFT takes. In the meantime, please enjoy this handsome (and royalty-free) raccoon.
App Chains are more than just scaling; they enable new use cases
Yes, Cosmos and Avalanche originally positioned the narrative as “app chains are the future of scaling.” That may be a fair point in the age of Ethereum killers. But to me, the strongest argument for app chains lies in the fact that different applications require differently optimized infrastructure – like how optimus prime cannot be a school bus. Where would the kids go when he turns into a robot? Definitely wouldn't work and they’re definitely dead.
A few strong value propositions for app chains beyond buzzwords like “sovereignty” and “flexibility" include:
You define your validator set
You own all your blockspace
You write your own rules at protocol level
Of course, pick and choose what extent you want to engage with these benefits:
Oracle networks like Bandchain select trusted validators for price reliability
dYdX decided that owning block space is key to building a central limit order book on-chain
Osmosis writes superfluid staking into its protocol to maximize liquidity for AMM
Axie infinity and Crabada would like to own all its blockspace so that its users aren’t priced out from on-chain volatility
KKR is the taking a shot at tokenizing assets in the US through highly permissioned validators, KYC-embedded wallets, and likely some custom asset features such as claw backs
Of course, app chain preachers would tell you even more reasons, such as ability to edit VM directly or the flexibility of tokenomics design. I find those to be secondary reasons that require a bit more fine tuning. In any case, I hope the message is clear: app chains can make things happen in ways application-agnostic blockchains cannot.
Even private, permissioned app chains deserve to exist
Permissioned chains are ideologically mid at best, not unlike midnight by Taylor (fight me). But hear me out, permissioned chains (usually built around a single application) have a genuine place in the reality we are in - even if they would be 100x cooler once they are permissionless.
Regulation demands highly controlled environments. Smh. Under the current regime, perhaps the only way to prove “safety and soundness” to regulators is by spinning up a permissioned environment. I’ve even heard about people developing on private blockchains as “corporate testnet” until they feel comfortable deploying on permissionless infrastructure. This doesn’t have to be the reality, but it is - like how Timothée Chalamet gave everyone chlamydia.
Permissioned chains can still be improvements upon existing tech stack. At that point, you are essentially building, once again, a better distributed ledger with better security, automated processes, and faster settlement. Better yet, you can just fork Ethereum as opposed to building something custom. Thief.
Permissioned blockchains are not without success. Signature bank has pioneered cash payment and settlement with its permissioned chain, Signet™, for a while. There are also many solutions out there using permissioned blockchains to streamline asset issuance, automating processes and driving down cost for asset issuers.
So, if the use case is to automate internal processes using a more secure infrastructure, then even the 12 seconds of settlement on an uncool blockchain is cooler than what we have right now: a sentient fax machine that runs on DOS.
App Chain as a Service is a promising business model
I’ve always wondered: when Polygon’s business development team goes to Starbucks, how the hell do they explain what the hell an app chain is. It’s not like explaining blockchain is hard enough on its own.
My hypothesis is that they are probably like: launching an app chain means accessing blockchain without dealing with all the other crap out there. Better yet: we got you! We will set everything up for you! You can sit back and chill! Drink a kombucha!
This model of “helping people set up their own blockchain” can be called App-Chain-As-A-Service, or rather just 2023 PaaS (Platform-as-a-Service). The Avalanche x AWS partnership (and Tencent Cloud too) is perhaps the best example. It’s quite smart: a foundation for a permissionless technology is but an as-a-Service provider who takes care of setting up the necessary platform on optimized infrastructure. You pay a service fee to the foundation, and a security budget to the permissionless infrastructure. Decentralized players win; centralized players win; only the Philadelphia Eagles loses. Yay?
Open-source protocols have been doing this for a long time to varying degrees of success, but, at a bare minimum, app-chain-as-a-service might actually help Alt-L1s and rollups find paths to revenue. That’s cool because revenue is yet another retro concept that VCs are bringing back. History is indeed cyclical.
Interoperability is a key unlock
The success of app chains hinges on blockchains being interoperable, meaning assets and data on one chain can travel to other chains with ease.
Without interoperability, composability dissolves, liquidity gets strained, and user experience goes straight to shit. The way out will be more interoperable tools. Whether it is Avalanche or Cosmos or Eigenlayer or LayerZero or Osmosis or Ren Protocol, a significant portion of the community believes in breaking down the boundaries between blockchains.
But interoperability needs to be done safer and faster. Security is definitely the top concern. Trust-minimal bridges are not quite there, as is manifested through the hacks and attacks we’ve witnessed. In terms of speed, current bridges simply aren’t fast enough. This isn’t just an issue for traders, but also for regular folks looking for a seamless user experience.
For permissioned players, interoperability provides a balance between crypto and regulation. There is no reason why stablecoins cannot be used to fund a compliant security on a private chain, or vice versa. In the utopia of zero knowledge proofs, we can expect settling private blockchain information onto public blockchains. Of course, there is a long way to go.
Those who reject crypto wholly will need to really justify that blockchain is the best decentralized infrastructure for their use case - like how I insist on using Apple Music. What if I like fewer features?? Fight me.
Which App Chain will prevail? We might have to wait for more use cases.
The field of App Chain is ever evolving:
L1 blockchains such as Avalanche Subnets, Polygon Supernets, Polkadot Parachains, Cosmos at large, and more
Ethereum roll-ups such as zkSync’s Layer 3 vision
Some private blockchain solutions (e.g., hyperledger fabric) could also be considered since they are deployed to enable a very small set of applications
The debates about app chain designs is too big-brained for me to process, but in practice, app chain designs haven’t yet been pushed to their limits.
Apart from a few high-profile projects, the use of app chains is still relatively limited compared to L1s and L2s. As economic incentive mounts, we will start to see huge influx of users, high blockspace demand, and eventually, sophisticated attacks
I’m sure some of them will cave under pressure, some part of my conjecture will be proven false, and some part of my ego will get a lil hurt. But until then, baggy jeans be baggy, witches be witchy, and I be bullish on app chains and Apple Music. Fuck Spotify.
Ty Evan Ian and Julian.