Hello!
Thank you for taking the time to speak with us today. We’re excited to learn more about Gevulot and your upcoming innovations, including ZkCloud and Firestarter. Let’s dive into it!
1. Can you provide a brief overview of Gevulot, its core mission in the ZK ecosystem, and how ZkCloud differentiates itself from other ZK solutions?
Gevulot is building ZkCloud, which is the first credibly neutral, decentralized
cloud for zero-knowledge proof generation and verification. ZkCloud allows users to generate ZK proofs for any proof system at a fraction of the cost.
2. Gevulot is soon launching Firestarter, your production-ready permissioned network. Can you explain its significance and how it contributes to the development of ZkCloud?
Firestarter is an end-to-end implementation of Gevulot’s ZkCloud, just running in a permissioned fashion. It is a high-performance compute network optimized for ZK, designed to scale to thousands of prover nodes handling both CPU and GPU workloads. The full system is made up of a blazingly fast CometBFT-based chain, which handles Credits, workload allocation and monitoring, and a network of prover nodes that provide compute for generating zk proofs.
3. You’re working with notable collaborators like Taiko and Risc0. How are these partnerships shaping the future of Gevulot?
From now on, users can generate proofs using one of the pre-deployed provers, such as Risc0, SP1, Nexus, Aztec, ZKsync & Polygon or by deploying their own. Gevulot Firestarter is designed to be able to support thousands of prover nodes with minimal overhead making it ideal for large-scale parallelized workloads such as parallel RISC0 segment proving or Aztec proof tree construction.
4. ZkBoost connects various proof demand sources. Can you tell us more about this consortium and its importance in the ZK landscape?
Gevulot is a founding member of The ZkBoost Consortium, a neutral and open-source software for generic ZK technology. Along with leading ZK-focused companies like Starkware, ZKsync, and Linea, a total of 35 companies are onboarded, developing the ZkBoost API that simplifies and unifies the ZK proof generation process across different networks. It’s designed to help streamline proof outsourcing, making it easier and more cost-effective for developers and platforms to tap into decentralized proving services without getting bogged down by the complexities of multiple APIs or fragmented solutions.
That means developers can focus more on building innovative applications while ZkBoost handles the complexities of proof generation, ultimately accelerating the adoption of ZK technology across industries
5. Could you elaborate on the hardware cluster you are building and how node operators can benefit from it?
Gevulot is currently onboarding incentivized prover nodes to Firestarter. Node operators can apply to run hardware for Firestarter and earn rewards with the following preconditions:
6. How does Gevulot’s network ensure high performance, low costs, and security for its users?
Gevulot’s network is designed to carry the best cost-to-performance ratio. Users can define the exact resource requirements for their workload and only pay for what they use. It is up to 90% cheaper than hyper scalers on comparable hardware.
Hourly rates:
ZkCloud’s decentralized architecture, combined with protocol-level fallback mechanisms, ensures high availability and liveness, censorship resistance, and trustless execution.
7. With the upcoming launch of ZkCloud in 2025, what are some key milestones we should look forward to?
Gevulot Deluge, a permissionless and incentivized testnet, will be launched during Q4/2024. Deluge will cover a full consensus implementation and run parallel with Firestarter. The network will have an E2E test of decentralized workload allocation, proof of capacity and incentive structures.
8. Lastly, how do you see Gevulot contributing to the broader adoption of ZK technology in the blockchain industry?
Traditionally, there have been two main approaches to ensuring that a computation was done correctly: trust and re-execution. These methods are opposites in terms of efficiency—trust minimizes redundancy and maximizes performance, while re-execution maximizes redundancy at the expense of performance. Zero-Knowledge (ZK) technology is transforming this dynamic by making it cost-effective and efficient to verify that a third party has performed a computation correctly.
With ZK, we can combine the strengths of both approaches—trust and re-execution—without their trade-offs. The ultimate goal of ZK technology is to reduce the computational and cost overhead of verifiability to near zero. At Gevulot, we believe that the future lies in making all the world’s computations verifiable—or as close to that as the laws of physics will allow. By contributing to the development and adoption of ZK technology, Gevulot is helping to push the boundaries of what’s possible in decentralized computation, driving efficiency, security, and scalability in the blockchain industry.
We look forward to hearing your insights!