Author: mteam.eth
Compiled by: ShenChao TechFlow
Any successful blockchain must create a flywheel effect as shown below:
Economic progress (such as TVL, price, revenue, transaction volume, etc.) must bring attention and visibility to the chain, thereby:
Supporting new applications with funding, allowing new developers to learn relevant technologies, enabling new users to utilize everything we build to improve their lives will inevitably lead to:
Innovation, along with improvements in infrastructure and applications, to enhance efficiency and explore new use cases and architectures. In the innovation phase, collaboration is particularly critical, but this phase is often also prone to team dispersion due to natural incentive mechanisms. Innovation drives economic progress, and the cycle begins again.
The problem Ethereum faces is simple — we have broken every part of this flywheel.
Note: This article explores Ethereum's high-level technical roadmap and does not focus on the social-level roadmap. Both must be combined to present a complete picture.
First, accept the problem
New applications, developers, and users are all on L2! Innovation happens on L2! Economic progress is also shifting to L2.
If these L2s can provide feedback to the flywheel, it is not a problem for Ethereum, but this situation is often not common.
Where is the root of the flywheel's breakage?
Ethereum (around 2020) considered scaling through Rollups as the only way to expand, severely overestimating the contribution of L2 to Ethereum's overall flywheel.
Rollup is seen as a scaling solution. Compared to sharding, Rollup appears simpler, avoids diluting the security of Ethereum L1, and can even bring better composability.
But Rollup is not just a scaling architecture; it is also an incentive architecture. A simplified logical chain might look like this:
We need to scale Ethereum.
Some form of sharding is a necessary condition for scaling blockchains with the features we need.
In protocol execution, sharding is too complex and has other issues.
Therefore, Rollup is the only way to scale Ethereum.
In my view, the second point here is the first major mistake. It is empirically evident that we were wrong (at least to some extent). For example, Solana and Monad both demonstrate reasonable scaling roadmaps that do not involve any form of sharding. At the same time, many core Ethereum developers have proven that we can push the performance of L1 further than it is now.
While I don't believe a single chain can meet all needs, I think we pushed toward this endgame before pursuing L1 scaling opportunities.
The fourth point in this reasoning is also insufficient. We failed to accurately assess the potential downsides of a Rollup-centric roadmap on the L1 network effect flywheel.
Ideal flywheel
I believe we can rebuild the network effect flywheel like this:
L2 should not draw network effects away from the flywheel but should accelerate the flow between each network effect.
Specifically, this means:
Providing nearly unlimited flexible scaling as overflow
Driving customization, specialization, and bold experimentation
Attracting users and developers
Increasing the total revenue of the Ethereum ecosystem and the revenue of Ethereum L1 itself
Maintaining high composability with Ethereum
This interaction has had the anticipated effect for both Ethereum and L2 — a rising tide lifts all boats.
My presentation slides from "Sequencing Day" in November 2024
Solid foundation
To effectively restart the flywheel, we need a strong L1. A composable L1. An ETH that is worth holding in your vault. An innovative coordination point.
How to achieve it? The answer couldn't be simpler. Actively expand L1.
We start with innovation at the L1 layer.
There are three reasons:
Scaling L1 increases network effects under the ideal flywheel
Scaling L1 raises the competitive threshold for any L2
Scaling L1 is beneficial to L2! (Especially the kind I will discuss in the next section)
Most readers of this article may understand what scaling L1 means in practice, but the core is to increase TPS and Gas per second while reducing slot time. We must build Ethereum L1 into the most powerful settlement network, yes, but also an execution network.
Combined, this is the solid foundation that L2 needs.
Let Rollup return
As L1 scales and builds its own network effects, optimizing L2 to contribute to the ideal flywheel cannot be wasted.
A few things need to be balanced here:
Ethereum gives the impression to Rollups that they will be prioritized within Ethereum.
Rollup has successfully grown its own network effects.
Any shift back to L1 scaling must be cautious and not completely alienate the main L2s (although some L2s have no reason to exist and should absolutely cease to exist).
I propose a simple Rollup design:
Rollup uses Ethereum for Data Availability (DA).
Rollup uses Ethereum for execution. This means it is a native Rollup.
Rollup uses Ethereum for ordering. This means it is a Rollup-based.
Rollup uses ETH as its native Gas token.
This design of Rollup is called "Ultrasound Rollup" and "Based + Native Rollup." I have written about them in detail!
"Ultrasound Rollups" are not yet achievable on current Ethereum. To enable their native parts, Ethereum needs to add a new opcode through a hard fork, namely the execution engine opcode. The ordering-based design currently also has some practical issues, all closely related to scaling L1.
Assuming we can achieve this, what would we get?
Ultrasound Rollups contribute to Ethereum's network effect flywheel by maintaining composability and achieving customization. Their combined scaling capabilities are theoretically very powerful; any Ultrasound Rollup can push execution like MegaETH or RISE. Ultrasound Rollups are not a regression but a necessary step forward.
The synergy between Ultrasound Rollup and Ethereum is so strong that I see it as an extension of Ethereum's network effects. Solana's ideas on network scaling are correct, but Ultrasound Rollups are not just about increasing Ethereum's capacity; they are themselves the network of Ethereum.
Existing Rollups have the potential to transition to Ultrasound Rollups. In fact, some teams have already committed to further exploring this option. New Rollups and application chains (Appchains) should prioritize this direction.
Through this approach, a unified Ethereum ecosystem can achieve "Universal Synchronous Composability," bringing insane scaling capabilities while also possessing infinite expressiveness.
In this ecosystem, user and developer activity occurs on L1 or specialized Rollups. Important and controversial states may still remain on L1. Developers can build cross-chain applications without focusing on the underlying gaps between chains. Users experience chain abstraction within the economic zone of Ethereum's scaling.
This is the integration era of Ethereum.
From "multiple choices" to "the obvious choice"
Ethereum is building top-tier Data Availability (DA), with ordering-based Rollup outputting our continuously improving ordering technology, while native Rollup will provide excellent execution capabilities.
Ethereum L1 integrates core Rollup services into a unified Ultrasound Rollup. While the market remains permissionless and chains can remain modular, Ethereum itself provides such important and comprehensive services that any competitor becomes irrelevant.
In this model, the accumulation of value (in the form of fees) becomes straightforward: providing the most valuable services, accessing the largest synchronized economic zones, the strongest economic security, the most censorship-resistant ordering, the most reliable settlement layer, and the safest data availability.
The narrative also naturally forms: "Ethereum is the best" -> Ethereum truly is the best.
Scale L1.
Let Rollup return.
Integrate everything.
And launch as soon as possible.
