Secure AI: How Google and NVIDIA Protect Data

Technology giants Google and NVIDIA have announced an expanded partnership with a focus on the development of AI in response to increased security demands. The main goal is to provide organizations with tools for implementing agentic AI in critical infrastructures, such as healthcare and finance, with an increased level of privacy and protection.

Security Is a Priority

The main focus of the project is to integrate AI without losing the confidentiality of information by encrypting data. For example, creating the new AI agents is possible with the help of NVIDIA Blackwell HGX and DGX systems, which work with Gemini through their own firewalls; accordingly, data security is not affected in any way.

NVIDIA's Confidential Computing Stack also plays a major role here. This stack encrypts all requests that Gemini processes.

Next-Generation AI

The next generation of AI systems will be able to make autonomous decisions and perform tasks with many steps. This also opens up opportunities for creating new generation applications in various industries.

Google is betting on the development of such agent systems and is already investing in tools for their monitoring and scaling. The company has announced plans to implement NVIDIA Dynamo, an open source library for scaling ML models, as well as the imminent release of RayTurbo, an improved version of the Ray framework developed in collaboration with Anyscale. It will be available on Google Kubernetes Engine (GKE) this year and will speed up the execution of AI pipelines.

What It Means for Companies

By integrating Google's Gemini multimodal models and NVIDIA's high-performance AI chips, companies will be able to access advanced technologies without compromising on compliance and data security.