Today, Nesa is announcing its biggest breakthrough since inception - a privacy preserving technology for AI that delivers the first fully encrypted inference, end-to-end, as an API with zero latency.
Introducing Equivariant Encryption, called EE for short. EE is the first adaptation of Homomorphic Encryption that works on a vast body of models to provide complete end-to-end privacy on your data insertion, query and response, but with zero latency.
What this means is your request is returned in nearly the exact same amount of time as it would if unencrypted, and now no one, including your cloud provider, has the ability to see your AI inference requests. No TEE, specific hardware or local set-up required.
You can learn more about EE here:
github.com/nesaorg/nesa
We have spent countless months perfecting this technology. Today it will go live with several of the largest enterprises in the world in pilots and will be rolled out to all Nesa users in the coming weeks.