Content
@
0 reply
0 recast
0 reaction
Brent Fitzgerald
@bf
What’s the thinking or dialog around L2s as value extraction middlemen? Win by attracting users, being sticky, then eventually exploiting them. Meanwhile mainnet validators they derive security from see less and less revenue. It’s getting siphoned off at the layer above. Just a hunch as I’m seeing more projects launch their own chains.
1 reply
0 recast
6 reactions
shazow
@shazow.eth
What's the distinction of L1 getting more efficient for L2 vs L1 undercharging for L2? The entire point of increasing efficiency is less revenue per unit, to support dramatically more units. Can we make a straight-face prediction that we'll scale ETH L2s to 1000+ TPS and ETH will be making less revenue than before L2s? (Also don't forget EigenDA and other restaking brings in revenue through restaking rather than displacing calldata fees into blob fees.) https://x.com/materkel/status/1845011100816773380
2 replies
0 recast
1 reaction
Brent Fitzgerald
@bf
Understood at an efficient market theory level. But L2s are quite centralized right now, and I genuinely wonder how many of them will eventually become decentralized, esp if they are successfully capturing value as sequencer fees. So will it be "worth it" for L1 validators to just chug along with minimal revenue while centralized L2 operators rake it in? Is there an opportunity cost there? Or at a human level, a resentment that builds? Does resentment translate to rational market behavior? Could L1 validators figure out a way to collude/unionize to reclaim value? Or will there always be defectors?
1 reply
0 recast
0 reaction
shazow
@shazow.eth
Honestly don't think it matters if L2s are centralized or not, from an L1 perspective. Every OP Stack rollup will shortly be Stage 1, Arbitrum is also stage 1. We have testnet prototypes for Stage 2. I have no reason to believe that we won't have Stage 2 rollups. But I'm not sure what difference it makes in terms of value extraction? Keep in mind, the privilege of choosing transaction ordering (sequencing) is MEV, not gas fees. From an L1 validator perspective, the math is: What's my expected staking revenue vs cost? If revenue goes down (e.g. too many people stake, so the stake rewards drop) and costs go up (e.g. higher bandwidth requirements) then we should expect fewer validators. But this is by design, right? We don't want infinite validators. Don't forget that restaking is going to also be a primary source of revenue for stakers.
0 reply
0 recast
0 reaction
