Here we report on the progress of the leading builders in the Proof Verification ecosystem, documenting recent significant releases, technical breakthroughs and general updates
@ZKVProtocol has released Runtime Upgrade 0.14.0, featuring integration of the new @RiscZero Verifier 2.0. The update enhances performance, restores previously paused functionality, and improves reliability.
The upgrade also includes a critical OpenSSL security patch, clearer build documentation, and various fixes across tests, CI workflows, and scripts, ensuring a faster, more stable, and developer-friendly zero-knowledge verification platform: https://blog.zkverify.io/posts/zkverify-runtime-upgrade-0-14-0-faster-integration-enhanced-performance…
The project has also upgraded its protocol, expanding Groth16 proof support from 16 to 64 public inputs. This gives developers greater flexibility in circuit and app design.
The release also adds support for a new EVM parachain and introduces automatic delivery fee collection for select cross-chain destinations, enhancing performance and future cost control readiness: https://x.com/ZKVProtocol/status/1928465987214729436…
@0xMilica shared a video explaining verifiers - core components in ZK systems that validate proofs using inputs and keys, returning true or false.
zkVerify’s verifier is modular, gas-efficient, and cross-chain compatible. It supports multiple proof systems like Groth16 and Plonky2, letting teams avoid building custom contracts. The video covers verifier logic, inputs, and integration: https://x.com/0xMilica/status/1924525005271961953…
Finally, zkVerify underwent a security audit by @trailofbits, a well-known firm assessing Web3 projects. The audit confirmed the security of its modular blockchain for ZK proof verification: https://x.com/ZKVProtocol/status/1922712321119617491
@wyatt_benno, Core Contributor at @NovaNet_zkp, announced the open-source release of a high-performance MSM implementation for the BN254 curve using WebGPU. Compatible with Windows, macOS, AMD chips, and mobile devices, it runs directly in the browser. Built for speed and parallelism, it’s based on the cuZK paper and works with Arkworks for in-browser ZKP proving:
https://x.com/wyatt_benno/status/1922992638677958882…
Additionally, Wyatt released an article arguing that AI systems will only scale meaningfully if they can verify their own knowledge using zero-knowledge proofs. Without internal, succinct verification, models remain fragile and reliant on human oversight. By embedding cryptographic guarantees into machine learning outputs, zkML and advanced zkVMs could enable more trustworthy, autonomous AI across complex domains requiring reliability and auditability: https://blog.icme.io/succinct-verification-the-key-to-ai/…
NovaNet also published an article explaining how zkML, once considered impractical, is becoming more feasible. Using techniques like GKR, CCT, and batched sum-check, developers can now verify complex model operations with far less memory and time. One-hot encoding, efficient lookups, and elliptic curve methods further optimize performance, making zkML a promising solution for private AI in decentralized applications: https://novanet.xyz/blog/verifiable-ai-with-zero-knowledge-proofs…
Finally, Wyatt, spoke at @HouseofZK’s College Tour in New York about how ZKPs are shifting blockchain design. He explained how instead of re-executing computation for consensus, blockchains are now becoming shared layers for verifiable compute, data storage, and orchestration, offering better performance and privacy: https://x.com/HouseofZK/status/1922280291973857488
@RosuGrigore, Founder & CEO @Pi_Squared_Pi2, posted the latest update on progress toward launching public-ready infrastructure. The team advanced on several fronts last week, focusing on system readiness and demo improvements: https://x.com/RosuGrigore/status/1927440983299371337…
- Core verification engine finalized
- New @ethereum validator using KEVM now handles mainnet blocks
- AI verifier integrated with blockchain mirrors
- Claim Explorer and wallet support in development
- UI updates and devnet launch preparations underway
In other news, Pi Squared proposes replacing Total Value Locked with Total Value Unlocked, shifting Web3’s focus from siloed capital to accessible, verifiable, cross-chain value. Their Verifiable Settlement Layer enables universal validation of contracts, data, and AI outputs across blockchains. TVU supports composability and usability, aiming to unify fragmented systems and prioritize practical interoperability over isolated growth metrics in the evolving Web3 landscape: https://blog.pi2.network/from-tvl-to-tvu-why-web3-needs-a-new-narrative-2/…
Additionally, they shared an article explaining the role of “claims” in Pi Squared’s VSL. In this system, a claim is a verifiable statement supported by proof, enabling secure, trustless interoperability across Web3. VSL doesn’t execute logic but validates and stores claims, allowing off-chain verifiers to confirm results. This approach boosts efficiency, privacy, and scalability across decentralized applications and networks: https://blog.pi2.network/unpacking-the-vsl-what-are-claims/…
Finally, the project is preparing to launch its Devnet, promising fast, verifiable, and cross-chain functionality. Devs can now sign up for early access through an online form as the release nears: https://x.com/Pi_Squared_Pi2/status/1920161779357003867
@LumozOrg updated its Rollup-as-a-Service platform following the activation of @ethereum's Pectra upgrade on May 7, 2025. With support for new features like blob data integration, ZK proof acceleration, and modular data availability, Lumoz offers developers faster and more cost-efficient L2 deployment. An account abstraction SDK will be released soon, making it easier to integrate advanced Web3 interactions: https://mirror.xyz/lumozorg.eth/im3PBZ0PebhMMYZI3cfQVC4Xdw0NU6BS6AiYArfMKAQ…
The project also launched its first MCP Server, enabling AI Agents to interact with Web3 applications through natural language via Claude Desktop. Based on the standardized Model Context Protocol, this innovation lets AI perform complex on-chain tasks, lowering barriers for non-technical users. By merging AI’s intent recognition with Web3’s smart contracts, Lumoz is advancing decentralized app functionality and driving scalable, user-friendly ecosystem growth: https://mirror.xyz/lumozorg.eth/LjEJB-jhW6HQZQtSrCinUzeRbf-t-kvY2v1Qg18SHKk…
Finally, Lumoz RaaS now supports RISC-V-based zkEVM, aiming to enhance Layer2 scalability and efficiency.
This move aligns with @VitalikButerin’s proposal to adopt RISC-V for Ethereum’s execution layer. RISC-V offers better performance, lower costs, and improved hardware compatibility. Lumoz provides prebuilt modules and a developer-friendly toolchain, making zkEVM deployment easier while boosting Rollup transaction throughput and reducing operational overhead: https://mirror.xyz/lumozorg.eth/GSXe3fVN-UF3Cb8gpvHDuMYZbIcZWJu7UkQDAUgW77k
@alignedlayer’s Proof Aggregation Service launched on the @ethereum Holesky testnet. It reduces the cost of verifying ZKPs by combining them into a single recursive proof. The service supports SP1 by @SuccinctLabs and R0VM by @RiscZero zkVMs and is suited for applications like rollups and DA layers.
Upcoming improvements include broader system support, better SDKs, and enhanced payment infrastructure: https://blog.alignedlayer.com/aligned-aggregation-service-is-live-on-the-ethereum-holesky-testnet/…
Also, the project shared the latest updates, including SDK refactors for improved structure and reduced indirection. New features allow users to verify aggregated proofs on-chain using updated contract functions. A new chunk-based aggregation method is in review.
Other enhancements include robust failover handling for Operators, full aggregation mode examples, expanded search for aggregated proofs, and multi-network support for proof verification. Version 0.15.3 is in draft: https://blog.alignedlayer.com/aligned-engineering-updates-may-19th-23rd-2025/…
In addition, Aligned and @class_lambda have completed a demo of their integration between Aligned’s Rollup as a Service and Ethrex.
The video shows a local L1 setup with transactions on an Ethrex L2 node. L2 execution is verified through Aligned’s AVS and aggregated proof on Ethereum, confirming deposits and withdrawals via tested batch verification: https://x.com/alignedlayer/status/1928515927789703314
@maqstik, COO of @alt_layer, took part in a panel at the ZK/AI Summit, co-hosted by @lagrangedev and @HouseofZK, discussing how to verify AI model outputs using cryptographic tools.
The conversation explored ZKPs, determinism, explainability, circuit design, trusted execution, and crypto-economic trust. He was joined by @1HowardWu, Co-founder of @AleoHQ; @0xnavkumar, CEO of @UngateAI; @samcassatt, Co-founder of @LayerOnEth; and @omarryehia1, Chief Strategy Officer at @lagrangedev.
Together, they examined the trade-offs between transparency and security, and the role of cryptography in shaping the future of verifiable AI: https://x.com/HouseofZK/status/1919390766444482664
