The upgrade of $ETH is beneficial for L2, theoretically reducing costs by 40%~60%, but this does not mean that tokens like $ARB , $OP , and strk will make any progress.

With a poor token model, endless unlocking and dumping, and a fragile ecosystem combined with declining TVL, the second-layer network seems to be entering a negative spiral.

The only thing that can save L2s seems to be the support of AI applications. However, based on underlying network factors, L2 is still not the first choice for AI business.