November 4, 20262 min read0 views

Cysic Shifts GPU Network Toward AI Inference as ZK Proof Demand Eases

Cysic is shifting its GPU network from zero-knowledge proof generation for layer 2 networks toward artificial intelligence inference, as demand for proof work eases and layer 2 fees rise above fees on Ethereum.

Zero-knowledge proofs let one party demonstrate that a computation was completed correctly without requiring another party to repeat the work. Layer 2 networks use proofs…

Reading Level🟢BeginnerNo prior crypto knowledge needed

Listen to this article

Zero-knowledge proofs let one party demonstrate that a computation was completed correctly without requiring another party to repeat the work. Layer 2 networks use proofs as they bundle transactions and settle them on Ethereum. AI inference is the process of using a trained model to respond to a request. Both workloads use GPUs, giving Cysic a way to redirect existing computing capacity.

Cysic’s mainnet launched Sept. 24, 2025. Since then, it has handled more than 13 million transactions and generated more than 10 million zero-knowledge proofs. The network lists more than 118,000 provers and 200,000 verifiers, and has worked with Scroll and Aleo.

The network runs as a Proof-of-Compute layer 1 compatible with the Ethereum Virtual Machine. It coordinates tasks for both proof generation and AI workloads. Cysic’s Venus zkVM and GPU optimizations have improved proof-time efficiency by more than 9%, while its custom C1 chip benchmarks at 1.31 million Keccak functions per second. Keccak is the hashing function used by Ethereum.

Cysic’s AI marketplace, CyAI, checks AI tasks through random statistical audits rather than generating a full zero-knowledge proof for each task. AI data centers’ demand for GPUs has constrained supply, adding to Cysic’s interest in specialized hardware.

Fan says Cysic continues to pursue zero-knowledge proving alongside the shift. The company’s goal is to prove Ethereum blocks in under eight seconds.