NEAR AI, the artificial intelligence arm of the NEAR Protocol ecosystem, is developing infrastructure that lets AI models run without exposing their data or code to the underlying cloud provider. The system relies on confidential computing, a hardware-based approach that encrypts data while it is being processed.
What confidential computing brings to AI
Confidential computing uses trusted execution environments (TEEs) — isolated enclaves inside a processor — to keep data and algorithms private even from the operating system or the cloud host. For AI workloads, that means a model can be trained or run on sensitive information without the risk of leakage. The technology is already used in finance and healthcare, but applying it to decentralized AI is relatively new.
NEAR AI’s infrastructure is designed to work with the broader NEAR Protocol, a blockchain platform focused on usability and scalability. By integrating confidential computing, the project aims to offer a way for developers to deploy AI applications that are both verifiable and private. The company has not disclosed which hardware vendors or specific TEE implementations it is using.
Why private AI execution matters
Privacy concerns around AI have grown as more companies move models to the cloud. Users often have no guarantee that their data isn’t being seen by the cloud provider or used to improve other models. Confidential computing addresses that by ensuring the data is encrypted throughout the entire computation cycle — at rest, in transit, and in use.
For decentralized networks, the challenge is bigger. Nodes that run AI models might be operated by untrusted parties. NEAR AI’s approach would let those nodes execute models without ever seeing the underlying data or model weights. That could open the door to AI applications in regulated industries like finance, law, and medicine.
The project is still in development. NEAR AI has not announced a timeline for a public release or a testnet. The company says it is working with researchers and infrastructure partners, but has not named them.
What’s next
NEAR AI plans to share more technical details in the coming months. The team is also exploring how to make the system compatible with existing AI frameworks like PyTorch and TensorFlow. No specific launch date has been set.
