“This month may well be remembered as an inflection point in the acceleration toward real-time proving of transaction validity for blockchains.”, — write: www.coindesk.com
The journey to transformative breakthroughs often follows this familiar pattern and this month may well be remembered as an inflection point in the acceleration toward real-time proving of transaction validity for blockchains.
First, Polygon Labs went public last week with its purchase of Fabric Cryptography’s specialized chips – verifiable processing units (VPUs) – claiming a significant leap forward for zero knowledge (ZK) proving capabilities.
Then, in collaboration with Optimism, a leading layer-2 blockchain, Succinct Labs separately unveiled a framework to upgrade optimistic rollup chains in the OP Stack into a ZK rollup – in just an hour. Users will be able to withdraw funds sooner after a rollup, or batch of transactions, is recorded on-chain, because the dispute resolution period required with optimistic rollups is replaced with the certainty afforded by ZK proofs.
And this week, we had news from RiscZero on its plan to become the distributed verifiable compute layer for all blockchains.
Together this series of announcements reveals the acceleration of the ZK space toward true scalability, with better proving speeds and costs.
Fabric’s VPUs promise a remarkable boost, thanks to their custom architecture and cryptographic compute power. As a researcher, builder, and strategic investor in Fabric, I’m excited and eager to test the chips in the wild. But as an industry, we need to keep our eye on a bigger prize. Hardware acceleration is only one component of overcoming a broader, more complex challenge: achieving real-time proving for blockchains, a.k.a. proof singularity.
Proof singularity represents a critical milestone in the evolution of blockchain technology because it promises to eliminate the bottlenecks that hinder scalability, privacy, and interoperability across blockchain networks. By achieving real-time proving, applications can execute complex computations securely without exposing sensitive data, opening doors to use cases such as private transactions, confidential smart contracts, and highly efficient rollups. This transformation would not only enhance the user experience by significantly reducing latency but also drive broader adoption by making zero-knowledge technology practical for real-world applications, from decentralized finance to enterprise blockchain solutions.
The broader impact of proof singularity extends beyond individual blockchain networks, as it paves the way for a more interconnected and scalable Web3 ecosystem. As ZK proofs become faster and more efficient, cross-chain communication and interoperability can be greatly improved, enabling seamless, secure interactions between various blockchain protocols. This could lead to a paradigm shift where data privacy and security are inherently built into the infrastructure, fostering trust and compliance in industries that require rigorous data protection standards, such as healthcare, finance, and supply chain management.
Ultimately, proof singularity has the potential to redefine the foundational principles of blockchain technology, merging performance, security, and privacy in unprecedented ways, thereby driving the next wave of blockchain innovation.
Hardware acceleration: necessary, not sufficient
Against that ambitious backdrop, Fabric’s VPU offers up to 900% increased big-integer operations performance compared to conventional GPUs. Polygon’s investment underscores a commitment to optimizing proof generation, betting on VPUs to enhance applications like Polygon zkEVM (a zero-knowledge version of the Ethereum Virtual Machine). However, while these numbers are great, many technical barriers remain to end-to-end, real-time proving, because big-integer operations are only a part of end-to-end proving time.
We know from experience that true real-time ZK proving cannot be realized through hardware alone. The industry’s approach must go beyond raw computational power, and should focus on a full-stack integration of hardware and software to address the range of challenges that stand in the way of seamless, real-time ZK applications.
A fully vertically integrated approach
- Better zkVM architecture: Let’s rethink the zkVM architecture from the ground up: the current generation of zkVM is limited by many sequential components. We need to re-architect zkVM so that zkVM execution and proving can be done in parallel from start to finish.
- Real-time proof aggregation: By incorporating this, the industry can address a key bottleneck often overlooked—verification cost and latency. Real-time proof aggregation allows cheap verification of proofs, on-chain, with minimal latency.
- Hardware/software co-design: Real-time proving needs to leverage different kinds of hardware, for example, CPU, GPU, VPU, and field programmable gate arrays (FPGAs), because each kind has different trade-offs between performance and energy efficiency. To fully realize the power of hardware, we need to co-design the hardware together with the software so that there is no leakage of performance or between different layers of the stack.
Bottom line
Polygon’s VPU investment signals an exciting step forward, but the journey to prove singularity will require more than hardware innovation. The true potential of ZK technology will be unlocked through a balanced combination of advanced circuits, optimized cryptography, and system-level breakthroughs. Let’s keep pushing the boundaries of ZK technology, striving to realize a future where real-time ZK proving is not just a possibility but a reality.
The race is on, and it’s about much more than faster chips—it’s about reimagining the entire ZK stack.
Shumo Chu is the CEO of NEBRA, a research and development organization, building technologies, infrastructure, and products to facilitate the mass adoption of zero-knowledge proofs.
Note: The views expressed in this column are those of the author and do not necessarily reflect those of CoinDesk, Inc. or its owners and affiliates.
Edited by Marc Hochstein.