In this post:
- io.net, a GPU cloud network, has joined hands with Phala Network to revolutionize decentralized AI and secure computation.
- The strategic alliance will equip Phala Network with GPU hardware through IO Cloud, io.net’s cloud network.
- Phala Network has built a 40,000 TEE (Trusted Execution Environment) CPU nodes infrastructure since its 2021 mainnet.
In a statement released on September 26, the io.net GPU cloud network announced its strategic collaboration with Phala Network to advance decentralized AI and secure computation. Phala would be equipped with GPU hardware through io.net’s cloud network, IO Cloud, to extend its decentralized AI capabilities.
According to the statement, Phala’s enhanced ability to support TE CPU nodes would provide an environment and the necessary infrastructure to flourish computationally complex dApps. Currently, Phala Network relies on Nvidia H100 and H200 GPUs to run its TEE tech, but the partnership with io.net would ensure Phala’s access to the ‘essential GPUs that formed the cornerstone of its decentralized AI ecosystem’.
io.net’s IO Cloud supports Phala’s on-demand GPU access
The io.net statement clarified that IO Cloud would enable Phala to deploy and manage on-demand decentralized GPU clusters. Phala Network could access powerful GPU resources without investing in costly hardware or the need to manage complex infrastructure.
IO Cloud democratized GPU access for ML developers and engineers by delivering a cloud-like experience to make advanced computing more accessible and up to 90% cheaper compared to traditional cloud services, according to the statement. Phala and io.net would combine efforts in benchmarking, research, and testing cutting-edge technology, starting with Nvidia H100s and H200s.
According to the statement, the partners were keen on exploring the deployment of Phala Network’s autonomous AI agent contracts and AI agents and integrating Phala’s TEE hardware and workers into the IO’s GPU network.
Phala sets new standards for TEE-enabled GPUs
As stated in the press release, Phala Network benchmarked its first TEE-enabled GPUs in August to set new decentralized AI standards. The partnership with io.net ensured that crucial AI workloads were processed securely with advanced cryptographic protection by extending TEE to include AI accelerators like GPUs, the statement claimed.
The statement explained that Phala’s security layer would further be enhanced using secure boot and encrypted memory-equipped Nvidia H100 Tensor Core GPUs. Additionally, Phala would be able to offer computational power to run and provide training for large AI models like Microsoft Phi and LLaMa 3 by accessing Nvidia H100 and H200 GPUs IO Cloud.
As per the statement, Phala Network built a robust 40,000 TEE CPU nodes infrastructure since its 2021 mainnet launch to enable Web3 developers to transfer complex computations smoothly from smart contracts to Phala’s secure off-chain network. The statement confirmed that the nodes were key in maintaining security and data privacy. The statement said they further delivered verifiable oracles and proofs to support various Web3 applications ranging from AI-driven agents to social apps.
