Today, Linea and Ethereum Mainnet upgrade simultaneously for the first time. In 2025 alone, four Ethereum hard forks from Paris through Prague were delivered to get Linea up to speed. Now we’re showing that zkEVMs can keep pace with Ethereum's upgrade cadence, something many thought impossible.

The timeline shows the story of our evolution:

• zkEVM to London: 4 years

• London to Prague: 4 months

• Prague to Osaka: real-time 

Most zkVM teams take a RISC-V approach: they run an Ethereum client inside a RISC-V virtual machine, which means they don't need to rewrite ZK proofs for each fork. We believe the trade-off is this sacrifices efficiency and performance.

Linea chose a different path. We develop ZK proofs directly for the EVM, which gives us superior efficiency and performance. This requires manual work for every single fork. Most teams gave up on this approach, convinced it was too difficult to maintain.

With Osaka, we take another step in demonstrating it can actually be viable at scale.

Over the past four years, we have created most of Linea’s EVM arithmetization modules manually. This hard work laid the foundation we can now build upon to deliver upcoming upgrades much faster, as demonstrated with Fusaka. In parallel, we also implemented  zkASM, a high-level "source language" that generalizes EVM constraint logic and compiles directly into circuits. zkASM will allow us to implement any ZK circuits much faster.