Any successful blockchain must create a flywheel that looks like this.
The problem that Ethereum has faced is simple – we have broken every part of this flywheel.
Note: This article approaches the high level technical roadmap of Ethereum and does not focus on the social layer roadmap. These two must exist together for a full picture.
New apps, developers, and users are on L2s! Innovation takes place on L2s! And economic progress is shifting towards L2s as well.
This would not be a problem for Ethereum if those L2s fed back into the flywheel themselves, but too often this is not the case.
Where is the root of this break in the flywheel?
Ethereum (circa 2020 onwards) thought scaling with rollups was the only way to scale, and it drastically overestimated how much L2s would contribute back to the broader flywheel of Ethereum.
Rollups were presented as a scaling solution. When compared to sharding, rollups seemed simpler, avoided diluting the security of Ethereum L1, and could even bring better composability.
But rollups are more than just a scaling architecture, they are an incentives architecture too. A simplified train of logic might look like this:
The first major misstep here, imo, was point #2. Empirically it is blindingly obvious that we were wrong (at least to a degree). Solana and Monad, for example, have both shown sound scaling roadmaps that do not involve sharding of any form. Meanwhile, many in Ethereum (especially core devs) have shown we can push L1 performance much further than we have today.
While I am not convinced one chain will ever be enough for everything, I think we pushed to that endgame before pursuing the opportunities for L1 scaling we had.
Point #4 in this line of reasoning also falls short. We failed to correctly evaluate the potential downsides/cons of a rollup-centric roadmap on the network effect flywheel of the L1.
I believe we can reframe the network effect flywheel like this:
Instead of drawing network effects away from the flywheel, L2s should accelerate the flow of network effects between each one.
Concretely, that means:
This sort of interaction has the desired effect for Ethereum but also for L2s – a rising tide lifts all boats.
A slide from my talk at Sequencing Day in November 2024
To effectively reboot the flywheel, we need a strong L1. An L1 worth composing with. An ETH worth holding in your treasury. A coordination point for innovation.
How? The answer couldn’t be simpler. Aggressively scale the L1.
We start with Innovation at the L1 level.
There are three reasons:
Most people reading this probably understand what scaling the L1 means in practice, but the core is increasing TPS and gas per second while reducing slot times. We must build Ethereum L1 into the strongest network for settlement, yes, but also for execution.
Combined, this is the firm foundation that L2s need.
As the L1 scales and builds its own network effect, there is no time to waste in optimizing L2s to contribute to the ideal flywheel.
It is important to balance a couple of things here:
Any sort of pivot back to L1 scaling must be careful not to completely alienate major L2s (although some L2s have no good reason to exist anyways, and should absolutely die).
I propose a simple rollup design:
Rollups designed like this have been called “ultrasound rollups” and “based+native rollups.” I have written about them in more detail already!
Ultrasound rollups are not possible on today’s Ethereum. In order to enable the native part, Ethereum would need to hard fork to add a new opcode called the execution engine opcode. Based sequencing also has some practicality problems for now. The things required are all closely related to scaling the L1.
Now, assume for a second we can do this. What do we get?
Ultrasound rollups contribute to the network effect flywheel of Ethereum by maintaining composability and enabling customization. Their combined scale is theoretically very high, and any one ultrasound rollup can crank execution just like MegaETH or RISE. Ultrasound rollups aren’t a step backward, they need to be a step forward.
Ultrasound rollups are so synergistic with Ethereum that I think of them as network effect extensions of Ethereum. Solana had the right idea with network extensions, but ultrasound rollups are more than just adding to the capabilities of Ethereum. They are Ethereum’s network.
It is possible to transition existing rollups into ultrasound ones. In fact, some teams have already committed to exploring the option further. New rollups and appchains should prioritize this.
A unified Ethereum with Universal Synchronous Composability, bringing insane scale but also infinite expressivity, is absolutely possible with this approach.
In this world, user and developer activity takes place on L1 or a specialized rollup. Valuable and highly contentious state likely remains on L1. Developers build cross-chain applications without needing to pay any attention to the underlying gap between chains. Users experience chain abstraction within the broadened economic zone of Ethereum.
This is the integrated era of Ethereum.
Ethereum is building premium DA, based rollups are exporting sequencing that we are improving too, and native rollups will provide amazing execution.
The Ethereum L1 bundles core rollup services into the integrated ultrasound rollup. While chains can still be modular (the market is permissionless after all), Ethereum itself provides such important and polished services that any competitor is irrelevant.
Value accrual in the form of fees is simple under this model: Provide the most valuable services, access to the largest synchronous economic zone, the strongest economic security, the most censorship resistance sequencing, the most reliable settlement layer, and the most secure DA.
Narrative falls into place too. “Ethereum is the best” ➡️ Ethereum is the best.
Scale the L1.
Bring the rollups home.
Integrate everything.
And ship it as fast as possible.
Any successful blockchain must create a flywheel that looks like this.
The problem that Ethereum has faced is simple – we have broken every part of this flywheel.
Note: This article approaches the high level technical roadmap of Ethereum and does not focus on the social layer roadmap. These two must exist together for a full picture.
New apps, developers, and users are on L2s! Innovation takes place on L2s! And economic progress is shifting towards L2s as well.
This would not be a problem for Ethereum if those L2s fed back into the flywheel themselves, but too often this is not the case.
Where is the root of this break in the flywheel?
Ethereum (circa 2020 onwards) thought scaling with rollups was the only way to scale, and it drastically overestimated how much L2s would contribute back to the broader flywheel of Ethereum.
Rollups were presented as a scaling solution. When compared to sharding, rollups seemed simpler, avoided diluting the security of Ethereum L1, and could even bring better composability.
But rollups are more than just a scaling architecture, they are an incentives architecture too. A simplified train of logic might look like this:
The first major misstep here, imo, was point #2. Empirically it is blindingly obvious that we were wrong (at least to a degree). Solana and Monad, for example, have both shown sound scaling roadmaps that do not involve sharding of any form. Meanwhile, many in Ethereum (especially core devs) have shown we can push L1 performance much further than we have today.
While I am not convinced one chain will ever be enough for everything, I think we pushed to that endgame before pursuing the opportunities for L1 scaling we had.
Point #4 in this line of reasoning also falls short. We failed to correctly evaluate the potential downsides/cons of a rollup-centric roadmap on the network effect flywheel of the L1.
I believe we can reframe the network effect flywheel like this:
Instead of drawing network effects away from the flywheel, L2s should accelerate the flow of network effects between each one.
Concretely, that means:
This sort of interaction has the desired effect for Ethereum but also for L2s – a rising tide lifts all boats.
A slide from my talk at Sequencing Day in November 2024
To effectively reboot the flywheel, we need a strong L1. An L1 worth composing with. An ETH worth holding in your treasury. A coordination point for innovation.
How? The answer couldn’t be simpler. Aggressively scale the L1.
We start with Innovation at the L1 level.
There are three reasons:
Most people reading this probably understand what scaling the L1 means in practice, but the core is increasing TPS and gas per second while reducing slot times. We must build Ethereum L1 into the strongest network for settlement, yes, but also for execution.
Combined, this is the firm foundation that L2s need.
As the L1 scales and builds its own network effect, there is no time to waste in optimizing L2s to contribute to the ideal flywheel.
It is important to balance a couple of things here:
Any sort of pivot back to L1 scaling must be careful not to completely alienate major L2s (although some L2s have no good reason to exist anyways, and should absolutely die).
I propose a simple rollup design:
Rollups designed like this have been called “ultrasound rollups” and “based+native rollups.” I have written about them in more detail already!
Ultrasound rollups are not possible on today’s Ethereum. In order to enable the native part, Ethereum would need to hard fork to add a new opcode called the execution engine opcode. Based sequencing also has some practicality problems for now. The things required are all closely related to scaling the L1.
Now, assume for a second we can do this. What do we get?
Ultrasound rollups contribute to the network effect flywheel of Ethereum by maintaining composability and enabling customization. Their combined scale is theoretically very high, and any one ultrasound rollup can crank execution just like MegaETH or RISE. Ultrasound rollups aren’t a step backward, they need to be a step forward.
Ultrasound rollups are so synergistic with Ethereum that I think of them as network effect extensions of Ethereum. Solana had the right idea with network extensions, but ultrasound rollups are more than just adding to the capabilities of Ethereum. They are Ethereum’s network.
It is possible to transition existing rollups into ultrasound ones. In fact, some teams have already committed to exploring the option further. New rollups and appchains should prioritize this.
A unified Ethereum with Universal Synchronous Composability, bringing insane scale but also infinite expressivity, is absolutely possible with this approach.
In this world, user and developer activity takes place on L1 or a specialized rollup. Valuable and highly contentious state likely remains on L1. Developers build cross-chain applications without needing to pay any attention to the underlying gap between chains. Users experience chain abstraction within the broadened economic zone of Ethereum.
This is the integrated era of Ethereum.
Ethereum is building premium DA, based rollups are exporting sequencing that we are improving too, and native rollups will provide amazing execution.
The Ethereum L1 bundles core rollup services into the integrated ultrasound rollup. While chains can still be modular (the market is permissionless after all), Ethereum itself provides such important and polished services that any competitor is irrelevant.
Value accrual in the form of fees is simple under this model: Provide the most valuable services, access to the largest synchronous economic zone, the strongest economic security, the most censorship resistance sequencing, the most reliable settlement layer, and the most secure DA.
Narrative falls into place too. “Ethereum is the best” ➡️ Ethereum is the best.
Scale the L1.
Bring the rollups home.
Integrate everything.
And ship it as fast as possible.