Just saw something interesting happening in the ZK space. Jordi Baylina and the core zkEVM team from Polygon have spun out to launch their own project called Zisk, and it looks like they're taking a fundamentally different approach to zero-knowledge virtual machines.

So here's what went down. The Polygon Foundation decided to shelve their zkEVM initiative, which was apparently burning over a million dollars annually. Then Sandeep Nailwal took over the foundation and pivoted the roadmap toward Polygon PoS and their AggLayer protocol instead. For Baylina and his team, the signal was clear - either adapt or move on. They chose the latter.

What's interesting is that Baylina retained all the intellectual property and transferred it to a Swiss entity he controls. The team had been incubating this within Polygon since May 2024, so they formalized the separation in June. Basically, they're taking the open-source codebase and building something independent from scratch.

The key difference with Zisk is the focus. While most zkEVMs obsess over EVM compatibility, Jordi Baylina's team is prioritizing low-latency proofs instead. That matters for real-world applications like DEXs and gaming where speed is everything. Early benchmarks suggest they could cut verification times by 40 to 60 percent compared to existing solutions, though obviously we'll need independent audits to confirm.

It's a smart move, honestly. Polygon invested nine figures into zkEVM but never quite delivered on the vision. Baylina is essentially saying he can do better with fewer constraints and a clearer technical direction. Whether Zisk actually delivers remains to be seen, but the fact that he's committing to keeping the codebase open-source and permissionless is a decent signal. Worth keeping an eye on if you're interested in how zero-knowledge tech evolves.