NVIDIA Confidential Bare Metal: Rent The GPUs, Keep The Model Private
NVIDIA published an announcement on October 21, 2026, during its GTC Berlin event, expanding NVIDIA Confidential Computing with a new deployment option called Confidential Bare Metal. Everything below comes from that first-party post on NVIDIA's own blog.
Transcript
NVIDIA said today that model builders can rent someone else's hardware. Their models stay private.
Confidential Bare Metal is a new option for sensitive workloads on third party infrastructure while the builder keeps control of models and data.
New security controls protect data in transit and enforce hardware level access restrictions, using cryptographic attestation.
Confidential Computing now offers three deployment options across Grace Blackwell and Rubin systems, while Cohere and Google Cloud are named adopters.
This is NVIDIA's own announcement of capability, not availability. It gives no availability date, no pricing and no benchmarks, and both named integrations are future tense.
Inboxsmith builds AI receptionists for small business. Please like and subscribe for more news.
Sources
Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:
- NVIDIA announced Confidential Bare Metal, a new deployment option with NVIDIA Confidential Computing that enables frontier model builders to run sensitive workloads on dedicated third-party infrastructure while retaining control of their models and data.(NVIDIA (blogs.nvidia.com), first party announcement titled "NVIDIA Confidential Computing Expands, Enabling Frontier Model Builders to Securely Scale Across Global Infrastructure", October 21, 2026)
- NVIDIA is introducing new platform security controls that protect data in transit, measure platform integrity and enforce hardware-level access restrictions, and NVIDIA Confidential Computing uses hardware-rooted isolation and cryptographic attestation to help protect models and data during execution.(NVIDIA (blogs.nvidia.com), first party announcement titled "NVIDIA Confidential Computing Expands, Enabling Frontier Model Builders to Securely Scale Across Global Infrastructure", October 21, 2026)
- NVIDIA Confidential Computing now offers a converged security architecture with three deployment options: Confidential Bare Metal for bare-metal infrastructure, Confidential Containers for Kubernetes environments, and Confidential Virtual Machines for virtualized environments.(NVIDIA (blogs.nvidia.com), first party announcement titled "NVIDIA Confidential Computing Expands, Enabling Frontier Model Builders to Securely Scale Across Global Infrastructure", October 21, 2026)
- Support now extends to NVIDIA Grace Blackwell NVL72 rack-scale systems, as well as x86-based NVIDIA HGX Rubin NVL8 and NVIDIA DGX Rubin NVL8 systems, and NVIDIA DSX OS will support all NVIDIA Confidential Computing modes as a reference architecture for cloud providers building AI factories.(NVIDIA (blogs.nvidia.com), first party announcement titled "NVIDIA Confidential Computing Expands, Enabling Frontier Model Builders to Securely Scale Across Global Infrastructure", October 21, 2026)
- Cohere plans to integrate Confidential Bare Metal into Cohere Model Vault, and Google Cloud is extending support for Confidential Bare Metal to A4X Max instances powered by NVIDIA GB300 NVL72 systems; providers listed as supporting NVIDIA Confidential Computing deployments include CoreWeave, Firmus, GMI Cloud, Google Cloud, IREN, Lambda, Microsoft Azure Infrastructure, Nebius, Nscale and Vultr.(NVIDIA (blogs.nvidia.com), first party announcement titled "NVIDIA Confidential Computing Expands, Enabling Frontier Model Builders to Securely Scale Across Global Infrastructure", October 21, 2026)
- The announcement names no general-availability date, no pricing and no performance or security benchmark figures, and both named customer integrations are future tense: Cohere "plans to integrate" Confidential Bare Metal and the Google Cloud instances "will combine" non-virtualized accelerated computing with cryptographic protection; every claim is NVIDIA's own, published on NVIDIA's blog, with no third-party verification.(NVIDIA (blogs.nvidia.com), first party announcement titled "NVIDIA Confidential Computing Expands, Enabling Frontier Model Builders to Securely Scale Across Global Infrastructure", October 21, 2026)