Sei today released a Giga upgrade roadmap, aiming for 200k TPS and 400ms finality.

In a time when the L1 performance race has already fatigued market appetite, these numbers remain eye-catching—especially since it plans to abandon Cosmos SDK and fully switch to an EVM-only architecture.
Giga includes four components: Ares execution client, Eidos state management (a flat key-value store replacing Merkle trees), Autobahn consensus, and Sedna private memory pool.
Among them, Autobahn has already reached 5 Gigagas processing capacity in the distributed testnet.
If fully implemented, Sei will completely detach from the Cosmos ecosystem and become an independent high-performance EVM L1.
However, in the history of L1 performance narratives, there is often a gap between promises and reality.
Solana’s multiple outages and Sui’s two halts within five months serve as reminders: high TPS often comes with complexity and stability risks.
Sei has also experienced network interruptions before; can Giga’s radical overhaul avoid repeating the same mistakes?
From a funding perspective, the TVL and user activity in the Sei ecosystem have not shown significant changes following the roadmap release.
Market pricing for the L1 performance narrative has become increasingly rational—TPS without application layer demand is just a number.
The true test of Giga is not whether it can reach 200k TPS, but whether it can sustain an application ecosystem.
$atom #sei #sol #sui #Blockchain