Critiques of Ethereum's Scaling Design

Crypto activity continues to be in a summer doldrum pattern as the newsflow has largely subsided after a frenzy in May/June around ETFs, the Ripple rulling, etc. Although there are new crypto-related bills being floated in Congress, it is unlikely that these will be enacted in the near term, which means the regulatory uncertainty will likely endure. This uncertainty has undoubtedly fueled the crypto market’s recent underperformance vs the stock market, which continues to melt up.

However, the Ethereum ecosystem’s technological development remains vibrant. A landmark Ethereum conference, EthCC, just concluded in Paris, where a multitude of infrastructure projects (new L2s, ZK tech upgrades, novel scaling mechanisms) and applications (Aave’s new stablecoin, Gnosis’s new crypto debit card, Uniswap’s new UniswapX product) were announced. While the 2020-21 bull market, in retrospect, was driven by speculative demand without the underlying fundamental infrastructure or application use cases required for sustainable adoption, it does seem that the groundwork is being laid during the bear market for a healthier growth cycle for the Ethereum ecosystem.

Last week, we discussed Ethereum’s fractal scaling design, which arguably provides the most flexible architecture for continued scaling to support onboarding the next wave of crypto users without hitting capacity constraints. The alternative to the fractal scaling model (which consists of L2s and L3s in a web of rollup chains settling on Ethereum) is the monolithic blockchain model – where all users and apps live on the base layer. Although we have covered the main reasons why monolithic chains cannot work – namely because they sacrifice decentralization by having high hardware requirements to run nodes – the monolithic blockchain camp frequently points out the flaws of fractal scaling. Let’s examine some of these critiques and address them:

Critique 1: L2s are centralized. This critique is largely valid today. Most L2s operate a single sequencer (which processes and orders user transactions), and the sequencer is typically controlled by the L2 team in early stages. However, the key property of an L2 is the ability to force-withdraw assets to ETH without being censored. Although L2s are relatively new and are launching with training wheels, users ultimately retain ownership of their L2 assets. The rest of the infrastructure – sequencers, provers, validators – will decentralize over time.

Critique 2: 7-day withdrawal time for Optimistic Rollups. First, ZK Rollups, which are likely the endgame (although far away from competing on cost), do not have a long withdrawal time. However, even so, we believe most users will live entirely on L2s, eliminating the need to bridge back and forth to L1 ETH.

Critique 3: Lack of composability. This is the most common critique from monolithic blockchain supporters. If different apps live on different L2s, then how can users interact with the full ecosystem? Isn’t it better to just have all apps on one monolithic L1? First, all monolithic L1s will reach capacity limits with enough users unless they sacrifice on decentralization (and become Google or Apple). Second, part of the magic of ZK technology is ZK bridging, which will ultimately allow for instant cross-L2 communication and therefore allow L2s and L3s to interact seamlessly under the hood.

Critique 4: L2 will just become their own L1s. This is being disproven; indeed, many previous L1s are transitioning to becoming L2s on Ethereum to inherit ETH’s security and settlement strength.

ETH Scaling Critiques - Why Modular vs Monolithic?

• Despite modular blockchain scaling being more complex than a monolithic base layer, modularity has been a superior solution across many systems. The Internet is modular, with TCP/IP as a base layer and a variety of websites, servers, and combinations deployed on top. Hardware is modular, with parts that can be swapped out to optimize use cases.
• Having a modular stack for blockchains allows for max experimentation, which will facilitate more customized adoptions. If banks want KYC-gated payment systems, they can deploy their own centralized L2 requiring KYC/AML and having transaction reversing mechanisms – and they can still enjoy the security and universal settlement provided by Ethereum. If governments want their own CBDCs with permissioned control, these can be deployed as their own L2 with composable settlement on ETH.
• Ultimately, the separation of power via modular architecture may win.

ETH Market Critiques - ETH Economic Snapshot

• Lastly, we revisit weekly Ethereum statistics, this time to highlight why having Ethereum as a universal settlement layer is so powerful. By having multiple L1 settlement networks, each blockchain system must provide its own security and censorship-resistance guarantees. Instead, Ethereum has ossified enough after the Merge and EIP-1559 to become a self-sustaining settlement ecosystem that various L2s can plug into, thereby outsourcing security to focus on user acquisition and applications.
• The amount of staked ETH continues to climb weekly, with 18.4% of the supply now staked (just over 22mm), and another 80,489 validators (2.5mm ETH) in the queue to stake. The wait time given the large amount of validators waiting to stake is 33 days, while the exit queue is empty, with zero validators waiting to exit.
• Fees continue to be at bear market lows, keeping staking yield sub 5%.