NVIDIA Confidential Computing Expands, Enabling Frontier Model Builders to Securely Scale Across Global Infrastructure
NVIDIA Blog John Fanelli ● Covered by 3 sources
NVIDIA added a bare-metal option for confidential computing. It lets AI builders keep models and data private on third-party infrastructure.
Based on reporting by NVIDIA Blog, John Fanelli — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
NVIDIA is broadening its confidential computing stack with a new option called Confidential Bare Metal. The pitch is simple enough: let frontier model builders run sensitive workloads on dedicated third-party infrastructure without giving up control of their models and data.
The company is pairing that with new platform security controls that protect data in transit, measure platform integrity, and enforce hardware-level access restrictions. In practice, that means customers can own and operate their own software stack on NVIDIA accelerated computing infrastructure, even when the underlying hardware sits in third-party cloud, neocloud, or enterprise environments.
NVIDIA now describes its confidential computing setup as a converged security architecture with three deployment choices: Confidential Bare Metal for bare-metal infrastructure, Confidential Containers for Kubernetes, and Confidential Virtual Machines for virtualized environments. The idea is to let organizations line up the security model with their own trust boundaries and operating needs.
The reach is widening across NVIDIA hardware too. Support now extends to Grace Blackwell NVL72 systems, as well as x86-based HGX Rubin NVL8 and DGX Rubin NVL8 systems. NVIDIA says cloud providers such as CoreWeave, Firmus, GMI Cloud, Google Cloud, IREN, Lambda, Microsoft Azure Infrastructure, Nebius, Nscale and Vultr support confidential computing deployments, while systems partners including Cisco, Dell Technologies, HPE, Lenovo, Supermicro and Bull will back the features through NVIDIA-Certified Systems. NVIDIA DSX OS is also set to support all of these confidential computing modes.
My take — AI-written commentary, not fact-checked reporting
This is NVIDIA doing the grown-up version of AI infrastructure: less hype, more control over who touches the machine and when. The interesting part is not the buzzword pile, it’s that privacy is becoming a feature of the stack, not a promise from the sales deck. That’s where the real competition is headed.
Read more about this at: NVIDIA Blog