Ocean Network, a decentralized computing platform for AI workloads, has announced the beta launch of its peer-to-peer GPU orchestration layer. The system is designed to simplify access to high-quality hardware and reduce dependence on traditional cloud providers.
During the beta phase, the company is purchasing GPUs through its 2025 partnership with Aethir, a distributed cloud computing provider. This gives users access to hardware like NVIDIA H200 at competitive rates.
Ocean Network is offering $100 in compute credits to early users. The platform is used globally, with capacity scaling based on demand. In the future, individuals and businesses will be able to connect idle GPUs and generate revenue by offering computing resources.
An important component is the Ocean Orchestrator, an extension that is integrated into development environments such as VS Code and Cursor. It allows users to select hardware, deploy tasks, and monitor execution in real time from a single interface.
The pricing model is based on actual usage. Payments are handled through an escrow system on Base, an Ethereum Layer 2 network for low-cost transactions. Money will only be released after tasks are completed.
For sensitive data, the platform uses a compute-to-data model. This means that algorithms are executed where the data is stored and only results are returned, without transferring raw data.
The beta phase focuses on users running workloads, with plans to expand into a broader decentralized computing marketplace.
Image: Freepik