I have been saving this for last. My goal was to demonstrate that monolithic blockchains are a technological dead end. Over 30 posts and hundreds of comments (particularly on Reddit) over the last year or so, I think I have written pretty much everything I wanted to say on the matter, and if you’re still not convinced, nothing else I say ever will. So, the last question is — what can monolithic blockchains do to remain relevant in the brave new era of specialization? Specialize, of course. It’s like asking what would farmers crafting their own homebrew sickles and using horseshit do after the industrial revolution? Use tractors and fertilizers built by others who specialize in those instead, of course. Lastly, I’m taking a long-term view. Here are their options:
Remain monolithic, accept technological obsolescence, but focus on marketing, memes & build network effects and niches before modular chains dominate
Let’s get the bored ape in the room out of the way. We have countless examples from history where the inferior tech won due to marketing, memes & network effects. I’m not sure if they’ll be able to keep up with 100x-10,000x inferiority, though. Nevertheless, there are certainly niche use cases which don’t require modular architectures. Bitcoin is a decent example — it’s happy catering to a sizeable niche — a store-of-value linking metaverse with meatspace, which doesn’t necessarily require scalability or cutting-edge tech. Another potential case would be Cardano — they have built a strong cult through by far the best marketing & memes in the industry. There’ll be people who’ll swear by it for years to come— just like there are people who continue using CRTs. Side note: CRTs, while obsolete, do have some very niche benefits. Same can be true of monolithic chains — though I’m not sure what these niche cases are just yet.
[A word of advice to projects that opt for this: please don’t be obnoxious lying and FUDing through your teeth. Accept that there are other, newer technologies out there, but you’ll make it work regardless.]
Expand into a validium
The reason I say that is because a monolithic chain can simply retain everything and become a validium. This is the path of least resistance. You lose nothing, but now share security with whatever the most secure layer is. All that needs to be done here is generate ZKPs and verify on the top security layer. Of course, that’s a huge challenge right now, but as StarkNet and zkSync 2.0 overcome it — and Polygon Hermez, Scroll & the EF have native zkEVMs, the knowledge is going to permeate and it’s going to get progressively easier.
The cost per transaction will be negligible — particularly once we have GPU/ASIC provers. For a busy validium with many transactions amortized over one ZKP, the cost could be fractions of a cent long term (currently ~$0.01). It’s just a huge increase in security for very little cost — absolute no-brainer.
Once this transition is made, the new validium can actually start cutting back on their consensus mechanism — due to the new security inherited — and push scalability higher, be more innovative with execution layer features etc. It’s not just about security, of course, you also benefit from the network effects and ecosystem support. A great case is Immutable X — despite off-chain DA, that it’s partially secured by Ethereum is evidently a huge plus point, and why it’s the runaway winner in the NFT space.
Become a volition or rollup
This is arguably the most attractive option. In addition to expanding into a validium, you also give users the choice to settle on the top security & DA chain to inherit maximum possible security & scalability. This makes you a volition. The other option is to abandon your data availability layer and just focus on being a rollup with maximum security. I used to think this is the most pragmatic approach, but I now think there’s too much capital and hubris invested in monolithic projects for them to take this rollup-only approach any time soon. The one that does will be a pioneer and gain immense network effects, though. As mentioned above — it’s not just security, but also inheriting networks effects and ecosystem support. We have seen how every major application on Ethereum has committed to deploying on Arbitrum One — it’s the most adopted smart contract platform by developers after Ethereum itself.
Become a security & data availability layer
There are two ways to do this — rearchitect your monolithic structure to be modular friendly. Or, build a data availability layer with a minimal security layer like Polygon Avail or Celestia are doing.
Of course, Ethereum is taking the former approach as a security & data availability layer. For other sharded networks like Polkadot and NEAR, this is actually a fairly straightforward pivot to make. Replace execution shards (parachains) with data shards; leverage rollups/volitions as execution layers instead of execution shards (parachains). Potentially, you can continue having execution on shards, just reorient to focus on data & rollups. It’s harder for single-ledger chains or non-shared-security multi-chain networks — they’ll need to build new data availability layers to remain competitive.
Needless to say, Bitcoin & Ethereum have a gargantuan advantage in “security” — which covers credibly neutrality, monetary premium, social consensus etc. But these less secure chains can be strong competitors in the data availability space, and build their own niches as a security + DA chain.
Become a security-only layer
Speaking of Bitcoin, it’s the only realistic competitor to Ethereum on “security”. The easiest way forward is for Bitcoin to add functionality to verify ZKPs. This makes it a security-only layer where validiums can settle. I doubt this’ll apply to anything other than Bitcoin — but perhaps we’ll see new innovations around revolutionary consensus mechanisms that make proof-of-stake obsolete. Lastly, yes, Bitcoin can build a DA layer, but realistically I doubt that’ll ever happen.
Build a data availability layer
Focus on building the best data availability layer for validiums and volitions. In the “security & data availability layer” section — we saw that certain data availability layers like Polygon Avail and Celestia are actually using consensus mechnanisms from the monolithic era, and are acting as both a security and DA layer. However, focusing on data availability exclusively, you can innovate on new security models beyond monolithic consensus mechanisms which could potentially unlock new efficiencies.
It’s abundantly clear that technologically and pragmatically modular architectures are orders of magnitude better and obsolete monolithic blockchains. However, technological obsolescence does not mean irrelevance. Monolithic chain projects still have plenty of options to be relevant in the modular world. Let’s hope they are pragmatic and make the right choices to not only survive, but also thrive. I fear there’s too much ego and hubris in this industry and many will become irrelevant though.