Nutanix and Nvidia launch AI Factory for governments

"Nutanix, together with Nvidia and others, is introducing an integrated AI Factory solution for governments and highly regulated sectors. The solution features FIPS encryption, standard Kubernetes management, and continuous security monitoring. Nutanix and Nvidia hope that the solution will deliver AI capabilities where complex legislation can be a roadblock. The companies note that various countries, for example, hopes to build an AI factory of sorts, about which few details are generally available."

"Nutanix bases the solution on NVIDIA's AI Factory for Government architecture. NVIDIA containers run on Nutanix Kubernetes Platform (NKP) and integrate with Nutanix Enterprise AI (NAI). The combination provides hardened configurations, FIPS-compliant encryption, and vulnerability monitoring. NKP can run on both virtual machines and physical servers. A similar level of security would be possible from conventional infrastructures based on these solutions. The difference here is that an integrated solution guarantees that level."

"Nutanix talks about encryption at every layer, robust governance, and a fast time-to-first-token, i.e., the time span before AI models are actually deployable. This is largely due to the way Nvidia's AI Factory for Government works. Everything from networking to DPUs, GPUs, and cooling has been validated and is derived from Nvidia's Enterprise Reference Architecture. Customers are free to choose their model, but above all, they will not be dependent on external APIs and clouds."

Nutanix and Nvidia present an integrated AI Factory solution targeted at governments and regulated sectors, emphasizing FIPS-compliant encryption, standard Kubernetes management, and continuous security monitoring. The solution builds on NVIDIA's AI Factory for Government architecture with NVIDIA containers running on Nutanix Kubernetes Platform (NKP) integrated with Nutanix Enterprise AI (NAI). Validated configurations cover networking, DPUs, GPUs, and cooling following Nvidia's Enterprise Reference Architecture. NKP supports both virtual machines and physical servers and includes hardened configurations and vulnerability monitoring. The package enables local model deployment to avoid reliance on external APIs and clouds while supporting open-weight models and rapid model deployment timelines.

Read at Techzine Global