Think the focus in the L2 world has been on "plain vanilla" EVMs for too long, which especially in the Ethereum ecosystem has led to a lot of replication but brought little innovation. While Base (CB distribution power), Arbitrum (first mover) and increasingly Mantle (app layer innovation) are quite successful, most other EVM L2s have never really gained traction, or have been bleeding significantly in recent months/years (e.g. OP mainnet which lost 2/3 of TVL since ATH). Imo that's not too surprising bc the Ethereum ecosystem and the broader market don't need 50 EVM L2s. At least not unless they bring something actually new to the table. I always liked the "network extension" idea from the Solana ecosystem in this context, which while being a rather broad term wrt what it can entail from an architectural perspective (e.g. Bullet, MagicBlock, Neon, etc.), is quite clear wrt the intention (extending the network's functionality). Makes sense given @solana's much stronger focus on scaling L1 execution kind of removes the need for L2s that replicate the L1s functionality on L2. Yet, there are use cases and apps that might still be hard to realize or that at least suffer under the constraints of the L1 execution environment. Bullet with it's perp DEX is a good example here on Solana, while on Ethereum similar appchains (e.g. @Lighter_xyz or @Paradex) that specialize instead of building another general-purpose EVM, and EVM L2s that innovate beyond what we've seen so far (e.g. @megaeth_labs or @rise_chain enabling realtime execution or @fluentxyz pioneering blended execution), now actually bring innovation to the Ethereum ecosystem that significantly broadens the design space and the network's functionality. While from the perspective of the base layers (Ethereum and Solana) the controversy around value accrual potential that L2s actually provide for the L1 (especially if they are fully modular, reducing L1 role to settlement only) vs enabling those apps on L1 remains, I think the rise of more opinionated and specialized L2s is a healthy and promising development. Yet, demand for the block space and all the theoretical TPS of these chains (not to forget all the old and new L1s that are out there too) has to grow as well to support the increasing number of chains. What's certain tho, is that the battle for attention, users, liquidity and ultimately relevance, in what measured by current standards seems like a rather saturated market, will be fierce. Strong technical moat and real innovation are a good start, but don't guarantee long-term success.