"L2s will make Ethereum deflationary again"
No they won't - and we already have examples of why that's true.
A report on L2s, blobs and why the Ultrasound Money meme is dead without mainnet users
TLDR
🔹L2s will only make ETH deflationary if they saturate both the blob and regular fee markets
🔹L2s, largely, aren't beholden to a specific batch posting cadence to L1
🔹The combination of the above means L2s will constantly maneuver around each other to avoid creating a high fee environment for themselves
--------------------------------------------------
BACKGROUND
First, a brief reminder of how Ethereum looks post-Dencun (the update from March that made L2s ~10x cheaper for user).
Dencun introduced the concept of blobspace, which is an expansion of blockspace on Ethereum meant for Layer 2s to post their batch data to.
A visual:
This new area has a few properties worth noting:
🟦"Blobspace" is currently limited to 6 max individual blobs within it per block
🟦It's a separate fee market from regular blockspace, but uses a similar mechanism
🔹Blockspace: If the current block is >50% utilized, increase the base fee in the next block
🔹Blobspace: If 4 (i.e. >50%) or more blobs are in the current block, increase base fee for the next block by ~12%
NOTE: L2s can use either blockspace or blobspace for their purpose, at their discretion.
LOOKING FORWARD
Given the above, we can predict a few things about expected behavioral flow of these L2s that will lead to burn:
🔹L2s will first saturate blobspace to the point where it stops being ~free (3 blob/block)
🔹Once it hits this level, they will begin doing the calculus of posting costs/reputational savings and opt to either:
a.) Use calldata
b.) Slow their posting frequency / spread out / coordinate to cool the fee market
That is all to say that L2s modus operandi will be to constantly adjust their behaviors to avoid creating a high fee environment for themselves, despite the top L2s already making millions per month in profit (w/ near-100% margins).
The ETH vampire death loop:
🔹Prolong low-fee environment
🔹Wait for Ethereum to scale, resetting market
🔹Prolong low-fee environment
🔹Wait for Ethereum to scale, resetting market...
"When does Eth plan to scale?" you ask
As early as next year with Pectra update, which may expand blob capacity ~2-3x.
--------------------------------------------------
WE'RE SEEING IT ALREADY
The market hasn't presented this opportunity a ton, but we have two examples of L2s intentionally modifying behavior to save costs and avoid fees (as you would for any business or rational actor).
Ex 1: Blobscriptions Craze
🔹In late March blobs saw "price discovery" from a blobscription phase and despite that: L2s and builders on average limited their blobs to the non-price-increasing level of 3 blobs/block:
h/t @hildobby_ for the Dune
They had no obligation to go beyond the fee-increasing level and so they didn't.
Ex. 2: LayerZero Airdrop
🔹On 6/20 we saw the $ZRO airdod, which created a substantial increase in transactions on Arbitrum and led them to flood the blob market.
TLDR
-> More blobs from ARB
-> higher blob costs for all L2s
-> ~$800k in ETH burned by L2s that didn't have the infra in place to switch to regular blockspace
I covered it when it happened if you want to read more:
The big takeaway from the event was how the teams reacted to the high-fee environment:
h/t @blocknative for the article
🔹Scroll, a zk rollup with no obligation to post, stopped posting altogether (lol).
🔹Taiko, a based rollup with a strict cadence, slowed their batch pace.
Both market participants acted in a rational way to reduce their overhead.
Going forward, you bet your ass they're going to have automated processes in place to avoid burning hundreds of thousands of dollars before they switch.
All L2s will, which brings me to...
--------------------------------------------------
CONCLUSION, WHAT CAN WE DO
When I set out to write this report it was to simply calculate something like "It would take another 4 Base-level L2s to make ETH Ultrasound Money™️ again" but the more I dug the more realized the goalpost would never be hit.
L2s will constantly adjust their behavior to avoid heavy costs (read: ETH burn). They're business - of course they will.
So, what do we do given the above? We make mainnet cool again, attracting both users and builders.
No more of this:
More of this:
You have to balance scaling through L2s with keeping your powerusers on mainnet - not pushing them off indiscriminately to one of a dozen ecosystems of various constructions that give minimal economic nod back to Ethereum.
Star by adjusting the messaging
> Fill to capacity
> Scale to your comfort level
> Tell users and apps to gtfo (skip this step)
> Scale through L2s
You also implement things like EIP-7623 () which defend L1 users blockspace from being used as L2 cost-saving failover space.
Relying solely on L2s as actors and mainnet as purely settlement isn't it. Not with the current incentives.
Thanks for coming to my Bread Talks.
No they won't - and we already have examples of why that's true.
A report on L2s, blobs and why the Ultrasound Money meme is dead without mainnet users
TLDR
🔹L2s will only make ETH deflationary if they saturate both the blob and regular fee markets
🔹L2s, largely, aren't beholden to a specific batch posting cadence to L1
🔹The combination of the above means L2s will constantly maneuver around each other to avoid creating a high fee environment for themselves
--------------------------------------------------
BACKGROUND
First, a brief reminder of how Ethereum looks post-Dencun (the update from March that made L2s ~10x cheaper for user).
Dencun introduced the concept of blobspace, which is an expansion of blockspace on Ethereum meant for Layer 2s to post their batch data to.
A visual:
This new area has a few properties worth noting:
🟦"Blobspace" is currently limited to 6 max individual blobs within it per block
🟦It's a separate fee market from regular blockspace, but uses a similar mechanism
🔹Blockspace: If the current block is >50% utilized, increase the base fee in the next block
🔹Blobspace: If 4 (i.e. >50%) or more blobs are in the current block, increase base fee for the next block by ~12%
NOTE: L2s can use either blockspace or blobspace for their purpose, at their discretion.
LOOKING FORWARD
Given the above, we can predict a few things about expected behavioral flow of these L2s that will lead to burn:
🔹L2s will first saturate blobspace to the point where it stops being ~free (3 blob/block)
🔹Once it hits this level, they will begin doing the calculus of posting costs/reputational savings and opt to either:
a.) Use calldata
b.) Slow their posting frequency / spread out / coordinate to cool the fee market
That is all to say that L2s modus operandi will be to constantly adjust their behaviors to avoid creating a high fee environment for themselves, despite the top L2s already making millions per month in profit (w/ near-100% margins).
The ETH vampire death loop:
🔹Prolong low-fee environment
🔹Wait for Ethereum to scale, resetting market
🔹Prolong low-fee environment
🔹Wait for Ethereum to scale, resetting market...
"When does Eth plan to scale?" you ask
As early as next year with Pectra update, which may expand blob capacity ~2-3x.
--------------------------------------------------
WE'RE SEEING IT ALREADY
The market hasn't presented this opportunity a ton, but we have two examples of L2s intentionally modifying behavior to save costs and avoid fees (as you would for any business or rational actor).
Ex 1: Blobscriptions Craze
🔹In late March blobs saw "price discovery" from a blobscription phase and despite that: L2s and builders on average limited their blobs to the non-price-increasing level of 3 blobs/block:
h/t @hildobby_ for the Dune
They had no obligation to go beyond the fee-increasing level and so they didn't.
Ex. 2: LayerZero Airdrop
🔹On 6/20 we saw the $ZRO airdod, which created a substantial increase in transactions on Arbitrum and led them to flood the blob market.
TLDR
-> More blobs from ARB
-> higher blob costs for all L2s
-> ~$800k in ETH burned by L2s that didn't have the infra in place to switch to regular blockspace
I covered it when it happened if you want to read more:
The big takeaway from the event was how the teams reacted to the high-fee environment:
h/t @blocknative for the article
🔹Scroll, a zk rollup with no obligation to post, stopped posting altogether (lol).
🔹Taiko, a based rollup with a strict cadence, slowed their batch pace.
Both market participants acted in a rational way to reduce their overhead.
Going forward, you bet your ass they're going to have automated processes in place to avoid burning hundreds of thousands of dollars before they switch.
All L2s will, which brings me to...
--------------------------------------------------
CONCLUSION, WHAT CAN WE DO
When I set out to write this report it was to simply calculate something like "It would take another 4 Base-level L2s to make ETH Ultrasound Money™️ again" but the more I dug the more realized the goalpost would never be hit.
L2s will constantly adjust their behavior to avoid heavy costs (read: ETH burn). They're business - of course they will.
So, what do we do given the above? We make mainnet cool again, attracting both users and builders.
No more of this:
More of this:
You have to balance scaling through L2s with keeping your powerusers on mainnet - not pushing them off indiscriminately to one of a dozen ecosystems of various constructions that give minimal economic nod back to Ethereum.
Star by adjusting the messaging
> Fill to capacity
> Scale to your comfort level
> Tell users and apps to gtfo (skip this step)
> Scale through L2s
You also implement things like EIP-7623 () which defend L1 users blockspace from being used as L2 cost-saving failover space.
Relying solely on L2s as actors and mainnet as purely settlement isn't it. Not with the current incentives.
Thanks for coming to my Bread Talks.
https://x.com/0xBreadguy/status/1807444529731211594
https://x.com/0xBreadguy/status/1821602952139960694
https://x.com/0xMert_/status/1822732571836314005
https://x.com/0xBreadguy/status/1823073712314139060
FYI -
The best math we can do is "how many L2s would it take to get to non-zero cost blobs and L2s changing behavior?" and for that, I have numbers.
Assuming behavior stays constant, we could add to current load and still be ~free blobs:
- 1 Base, or
- 2 Arbitrums, or
- 3 Blasts, or
- 4 OP Mainnets, or
- 9 Scrolls, or
- 12 Starknets, or
- 21 ZkSyncs, or
- Couple thousand alt-DA chains (Mantle, Eclipse, etc)
Reminder that once the above is added, they'll just jostle around to return to a low-fee environment and then coast until Eth scales again - resetting the market.
The best math we can do is "how many L2s would it take to get to non-zero cost blobs and L2s changing behavior?" and for that, I have numbers.
Assuming behavior stays constant, we could add to current load and still be ~free blobs:
- 1 Base, or
- 2 Arbitrums, or
- 3 Blasts, or
- 4 OP Mainnets, or
- 9 Scrolls, or
- 12 Starknets, or
- 21 ZkSyncs, or
- Couple thousand alt-DA chains (Mantle, Eclipse, etc)
Reminder that once the above is added, they'll just jostle around to return to a low-fee environment and then coast until Eth scales again - resetting the market.
They're already upgrading to automate exactly what I said they'd do (blob to calldata switch).
From an engineer for one of the OP Stack chains:
From an engineer for one of the OP Stack chains:
https://x.com/StackDigest/status/1823970385957937529?t=fzi0gbFb9OKD1mCBDZ997w&s=19
• • •
Missing some Tweet in this thread? You can try to
force a refresh
