Brevis ProverNet | The Next Paradigm Turning Point of ZK Track: @Vitalik Buterin once again forwarded Brevis @ brevis_zk on how to verify the three fastest proofs in Brevis Pico ZKVM and @ ziskvm, and complete the final block determination without re executing the block transaction. When the market is not very awesome, I still need to move forward. Just last night, I read the ProverNet white paper released by Brevis from beginning to end. In my opinion, it is more like a paradigm shift instruction manual: ZK's paradigm shift from "proof system competition" to "computing power market competition"! one ️⃣ First impression: The real problem with ZK is not technology, but 'computing supply' The ZK track has been very similar to the early Rollup in the past two years: Overall, it's very overwhelming! But what is the most realistic thing in the on chain world? Supply. Just like EVM TPS increasing by 10 times, it's meaningless if no one comes to execute the transaction. Brevis has not accumulated 'performance metrics' in the past two years, but rather the computing needs of real users. You can tell from its partners: Uniswap v4, PancakeSwap Euler, MetaMask, Linea, Aave ecosystem; They ran over 250 million production grade ZK proofs before coming to a very counterintuitive conclusion: the bottleneck of ZK is not prover performance, but the lack of computing power to adapt to all workloads. I was stunned when I read this sentence. Because this means that ZK's future is not 'a great prover', but 'a market for diversified computing power'. two ️⃣ Second feeling: ProverNet seems to have been "forced out" rather than imagined: The white paper mentions many real workloads: 30 day VIP trading volume inquiry (urgent, within 2-3 seconds); 100000 address reward distribution (not urgent, but in large quantities); ZkVM chunk → recurrence → merge → wrap → final proof (multi-stage); Historical data scanning (with strong I/O throughput); When these tasks are put together, you will find that they do not belong to the same hardware at all, and different types of provers are only good at some of them. You used to think that building a GPU Farm with a wealthy team would suffice, but the reality is that one farm cannot cover all ZK workloads; It's like finding a food delivery platform that only makes fried chicken, only to have it deliver seafood, grilled meat, desserts, and late night snacks at the same time. It simply can't handle it. So you see, the emergence of ProverNet is not a design, but a natural evolution; It is a requirement of the process. three ️⃣ Third feeling: TODA is the most underestimated part of the white paper: I know many people skip 'Truly Online Double Auction (TODA)' when they see it. But the reality is: This thing is the soul of the entire ProverNet. Why? Because what you need to do is the 'heterogeneous computing market', Then you need a mechanism: Can automatically match different providers; Can ensure the authenticity of supply and demand; Can make malicious quotations unprofitable, It can also approximate the optimal solution in NP hard matching problems (the white paper uses the "phantom requester" here to forcefully guarantee integer solutions, which is very clever); This reminds me of the early auction and liquidity mining incentive curves of MEV - algorithms are actually more important than performance breakthroughs. Because technological breakthroughs can be caught up with, but market mechanisms belong to the "basic rules" and can be reused for a long time and accumulated across cycles. four ️⃣ Fourth feeling: Pico Prism makes everything realistic, not idealistic! I have watched this part several times. The white paper is written very simply, but the data is extremely heavy: 99.6% of Ethereum L1 blocks (45M gas) can be proven within 12 seconds 96.8% completed within 10 seconds, with an average of 6.9 seconds; It uses 64 x RTX 5090 (consumer grade hardware). This means that 'real-time proof' is no longer a concept, but a reality. Why is it important? Because once ZK proves that it can achieve "near real-time", you can "outsource" more on chain logic to ZK, such as personalized rates, VIP transactions DEX loyalty、trustless incentive、 Intelligent router execution, multi chain state synchronization; This is exactly the demand side that ProverNet wants to undertake. Performance breakthroughs are the foundation, while market mechanisms are the superstructure. five ️⃣ Fifth feeling: Brevis is transforming ZK from a "hardcore technology" to a "market resource": I really like one sentence in the white paper: ProverNet transforms proof from an 'internal tool' to a 'publicly tradable commodity'. I have always felt that a technology moving from niche to mainstream has an important sign: it becomes a market resource rather than an internal capability of the manufacturer. GPU is like this, bandwidth is like this, and computing power is like this now (Nvidia's story). The problem with ZK is not whether it is difficult or not, but whether the supply is always insufficient and the cost is always too high. So when I saw Brevis say that a prover can: stay, take orders, report costs, be auctioned off, be fined, and be rated Coordinated execution of multi-stage tasks by multiple parties; I realized: Is it possible that ZK is no longer just cryptography, but a part of the 'computational economy system'. You can even imagine that in the future, someone will specialize in: zkVM prover、STARK aggregator、wrap specialist、GPU-based latency specialist、CPU-based batch specialist； An ecosystem will naturally grow. six ️⃣ Finally: This will be one of the main storylines for the ZK track in the next two years! Because it addresses a fundamental issue in the industry: 'Who will provide computing power for Web3?'? ” Rollup solves execution, DA solves storage, and ProverNet solves computation! This is an infrastructure layer that is' stacked on all chains'. Whoever can make it, is the next L2, the next EigenLayer. Even in a sense: ZK Prover Market may be the "decentralized AWS" of Web3. The white paper describes the mechanism, but what I see is that ZK has finally moved from an "engineering breakthrough" to a "market era".