Summary: Fermah is best cataloged as proof-generation infrastructure rather than as a generic ZK startup. Its current official materials describe a programmable execution layer whose live product is a universal proof market: a two-sided network matching proof demand with GPU/CPU supply, initially bootstrapped by EigenLayer operators. The docs, status page, explorer, and ZKsync case study together make Fermah look like an operational control plane for proof workloads — handling resource matching, prover-node onboarding, workflow orchestration, and proof delivery across proof systems, chains, and VMs.
What it does:
Runs a “ZK Proof Market” that accepts proof requests from seekers and routes them through Fermah Core to prover nodes, then stores and returns completed proofs
Frames the network as a universal marketplace for proof generation, with supply-side GPU machines and demand-side proof buyers
Publishes mainnet onboarding docs for prover-node operators, including hardware requirements, node registration, telemetry, whitelisting, and reliance on registered EigenLayer operator accounts
Exposes a proof explorer and public status page, which is a useful signal that the product is being run as live infrastructure rather than remaining a purely research-stage concept
Shows at least one concrete production-style integration narrative around proving blocks for ZKsync, positioning Fermah as outsourced proof-generation infrastructure for heavy ZK workloads
Key claims:
The docs describe Fermah as a universal proof market that supports “any proof system, chain, or VM” and is optimized for cheap, fast, and reliable ZK proof generation
The architecture docs say the network consists of seekers, Fermah Core, and prover nodes, with completed proofs stored and submitted back through callback flows
The mainnet docs indicate the network is operational enough to require whitelisted Ethereum addresses, node registration, telemetry, and specific GPU-oriented hardware/software setups
The blog argues that the core economic value is higher machine utilization, better orchestration, competitive pricing, and abstracting complex proving workflows into simpler request flows
The ZKsync case study claims Fermah is already being used to prove blocks for one of the largest proof producers in crypto, strengthening the case that this is real infrastructure rather than only a thesis
Whitepaper: No canonical Fermah whitepaper or litepaper was found in this pass. The strongest primary sources were the official docs, blog posts explaining the proof-market thesis and ZKsync integration, the status/explorer surfaces, and the Fermah docs repo; see ../whitepapers/fermah-primary-sources-2026-04-27.md.