@fermah_xyz has launched its incentivized testnet, intended to stress-test its Universal Proof Market under real-world conditions: https://fermah.xyz/blog-posts/testnet-is-live…
The testnet will roll out in three phases, increasing the number of prover nodes and testing key infrastructure components like matchmaking, payments, and peer-to-peer systems. Participants will earn 'points' based on contributions.
The aim is to validate Fermah’s promise of delivering fast, cheap, and reliable ZKP generation at scale. Each phase honors major cryptographers and introduces progressive testing milestones. External prover onboarding, smart contract audits, and performance monitoring are key parts of this final step toward mainnet readiness.
@thezkcloud has integrated with IBC Eureka, enabling seamless, low-cost ZKP transfers between @cosmos and @ethereum. This allows $PROOF to enter Ethereum DeFi, with support for @solana, @base, and @arbitrum coming soon. IBC Eureka offers EVM compatibility and simplified bridging, expanding ZkCloud’s reach across multiple ecosystems in theZK compute space: https://x.com/thezkcloud/status/1910284036049617225…
Additionally, the Deluge Testnet has expanded with 15 new validators, including @Xangle_official, @equilibrium_co, @StakingFac, @a41_allforone, @keplr_infra, @RockawayX_Infra, @Allnodes, @lavender_five, @NodesGuru, @piconbello, @staketab, @LunaNovaTech, @Figment_io, @Alchemy, and @silk_nodes.
These additions bring extensive infrastructure expertise, institutional-grade systems, and years of experience supporting blockchain networks - further strengthening ZkCloud’s decentralized infrastructure for efficient and cost-effective ZKP generation and verification: https://x.com/thezkcloud/status/1908057746911469693
@SindriLabs has launched 'Explorer' - a platform that simplifies ZK app development by offering ready-to-use circuits and tools: https://sindri.app/blog/2025/03/24/explorer/…
Developers can now deploy in minutes what once took months, using Sindri’s infrastructure and API. With over 10 million proofs already generated, Explorer supports uploads, sharing, and soon, rewards for creators based on API usage.
@zerobasezk is launching the world’s largest Zero-Knowledge Proof Trusted Setup Ceremony, inviting global participants to help build a universal cryptographic foundation using Groth16 and Gnark. Each person will contribute an irreversible proof embedded in future systems: https://x.com/zerobasezk/status/1906637233576390666…
In other news, @Gink5814 from ZeroBase took part in the Boundless House of ZK event, co-hosted by @boundless_xyz and @HouseofZK as part of @festival_web3 in Hong Kong.
He joined @Ismael_H_R, Founder of @lagrangedev, and @anuragarjun, Founder of @AvailProject to share candid, spontaneous views on current challenges and the future direction of the ZK industry: https://x.com/HouseofZK/status/1908798109938856209
@zan_team published an article in which @pignard_web3 explores how the universal AI agent Manus has sparked significant attention for its autonomous capabilities, inspiring broader AI agent development, especially in Web3.
The piece examines Model Context Protocol’s role in connecting AI with blockchain, outlining new deployment strategies and existing challenges, while emphasizing the importance of patience as integration between AI and Web3 continues: https://medium.com/@zan.top/from-manus-and-mcp-the-cross-boundary-exploration-of-ai-agents-in-web3-39c61032219c…
In addition, the project has released a monthly report (https://x.com/zan_team/status/1909566855381238025…), with key highlights including:
• Deployment of a new high-concurrency architecture for improved node stability and performance during traffic spikes.
• @solana data consistency enhancements across RPC and WSS.
• Launch of @base Sepolia Testnet node.
• GPU acceleration for PowerZebra’s Plonky3 components with significant speed gains.
• CEO Cobe Zhang’s participation in major Web3 and blockchain events.
• Co-hosting of the @Web3LabsClub ECO 2025 ecosystem event.
ZAN has also partnered with @GravityChain by @Galxe to support the Gravity Alpha mainnet. As a high-speed, stable RPC provider in the Asia-Pacific, ZAN Node now allows developers to build Web3 projects within the Gravity ecosystem.
This integration aims to streamline development on Gravity’s fast, parallel EVM L1 blockchain infrastructure: https://x.com/zan_team/status/1907357337863835678
@MarlinProtocol shared an article explaining how @Verida_io uses Marlin’s Trusted Execution Environments to enable AI agents to securely compute on private user data without exposing it: https://blog.marlin.org/veridas-use-of-marlin-to-compute-on-personal-data-securely-in-tees… This collaboration allows users to retain full control over their sensitive information while benefiting from AI-driven services. It supports secure APIs, private credential handling, and confidential data processing - enabling personalized automation in finance, health, and identity without compromising privacy.
In other news, Marlin is integrating their TEE infrastructure into @ragafinance’s off-chain oracle, enhancing the security and reliability of cross-chain operations. This enables Raga to optimize yields on assets like ETH and BTC on @ethereum while accessing returns across other networks: https://x.com/MarlinProtocol/status/1905590934986797141
@leofanxiong, Co-founder of @cysic_xyz, explained the fundamentals of ZK hardware in an educational piece for @CoinMarketCap Academy, outlining how specialised ASICs and optimised GPUs accelerate ZK proofs, enabling scalable, private, and energy-efficient blockchain operations. These advances support real-time use cases in payments, identity verification, AI compliance, and regulatory auditing: https://x.com/cysic_xyz/status/1911782108730917336…
Leo also joined the @boundless_xyz House of ZK panel at @festival_web3 in Hong Kong, where @yq_acc, Founder of @alt_layer, @no89thkey, Founder of @brevis_zk, and @sinka2022, Founder of @DelphinusLab, discussed the competition for attention in the ZK industry. Moderator Tina Liu, Head of Global Marketing at @RiscZero, challenged the participants with sharp questions about ZK narratives:
https://x.com/HouseofZK/status/1908802899196445054
@Ingo_zk has launched its second $100k research grant round for cryptography projects using its ICICLE library. Researchers saw up to 80x and 60x speed improvements in threshold encryption and collaborative zk-SNARKs by shifting from Arkworks to ICICLE: https://medium.com/@ingonyama/updated-ingonyama-research-grants-2025-455984201f3a…
ICICLE v3.7 adds native support for the Goldilocks field and curve pairings, enhancing performance for cryptographic protocols in Go using Apple Metal GPU acceleration.
The update also includes new APIs, Rust wrappers, documentation improvements, and backend-specific optimizations. Incentives are offered for early adopters using Goldilocks arithmetic or GPU protocols with the updated library: https://x.com/Ingo_zk/status/1911794172446617732…
The project also published a study analyzing the Sumcheck protocol in HyperPlonk, confirming that performance is primarily limited by memory access rather than computation: https://hackmd.io/@Ingonyama/Hardware-Friendliness-HyperPlonk-Part2…Profiling their CUDA implementation in ICICLE showed memory throughput consistently near peak while compute remained underutilized. As rounds progressed and data fit into cache, compute performance improved, reinforcing the memory-bound conclusion. The findings suggest future optimizations should target memory efficiency over computational improvements.
Lastly, @yuval_domb of Ingonyama proposed a simplification of the BaseFold polynomial commitment scheme by replacing the Sum-Check protocol with a Point-Check protocol: https://medium.com/@ingonyama/basefold-a-simple-improvement-to-basefolds-multilinear-polynomial-commitment-scheme-using-e150e0f13d85…His approach reduces the complexity of the original method by lowering the degree of the interactive proof from two to one, while preserving correctness guarantees. The new method still uses FRI for proximity testing but avoids an additional multiplication step by directly verifying the claimed value.
Domb also discusses how this structure links randomness and target evaluations through interleaved protocol execution.
