THE APEX TIMES
NVIDIA expands confidential computing, adding “Confidential Bare Metal” for developers running sensitive AI workloads on third-party infrastructure
The company says its updated security controls and new deployment option are designed to let frontier model builders keep control of models and data while scaling training and inference across global clouds and bare-metal environments.
NVIDIA on Oct. 21 expanded its confidential computing offering with a new deployment option aimed at frontier AI model builders who want to run sensitive workloads on third-party infrastructure without giving up control of their models and data. The update, described in a company blog post, adds “Confidential Bare Metal” to NVIDIA Confidential Computing and introduces new platform security controls intended to protect data and verify the integrity of the computing environment during execution.
NVIDIA positions confidential computing as an infrastructure layer that safeguards the models and data that AI agents depend on while they are running. In the announcement, it ties the update to its OpenShell Reference Architecture, which NVIDIA describes as a broader blueprint intended to help organizations deploy agents at scale. The thrust of the new release is to extend protection beyond virtualized environments, giving developers an additional way to match security needs to their trust boundaries.
A core part of the announcement is the “Confidential Bare Metal” option. NVIDIA says it enables frontier model builders to run sensitive workloads on dedicated third-party bare-metal infrastructure while retaining control of their models and data. The company also describes this as a way for organizations to own and operate their “bespoke software stack,” including host and platform infrastructure software, when using NVIDIA accelerated computing on third-party cloud, “neocloud,” and enterprise infrastructure.
NVIDIA also says it is introducing new platform security controls for the environment in which workloads execute. According to the blog, the controls are intended to protect data in transit, measure platform integrity, and enforce hardware-level access restrictions. NVIDIA describes its approach as relying on hardware-rooted isolation and cryptographic attestation, mechanisms intended to help verify that the platform is in an expected state and to protect secrets during execution.
The company frames the update as part of a broader evolution of NVIDIA Confidential Computing, which it says now offers a converged security architecture with three deployment options. NVIDIA says organizations can select the right model based on their trust boundaries and operating requirements, implying that different scenarios may require different levels or forms of isolation.
On the hardware side, NVIDIA says its confidential computing support is expanding across its Grace Blackwell and Rubin systems. The update includes support for Grace Blackwell NVL72 rack-scale systems and x86-based HGX Rubin NVL8 and DGX Rubin NVL8 systems. NVIDIA also says Confidential Computing is integrated with DSX OS as a reference architecture for cloud providers building “AI factories,” with DSX OS described as supporting all NVIDIA Confidential Computing modes.
The announcement includes several named ecosystem participants. Cohere plans to integrate Confidential Bare Metal into Cohere Model Vault, Cohere’s model protection system intended to provide a secure and confidential platform for inference. In the blog, Phil Blunsom, Cohere’s chief technology officer, says the company is extending hardware-rooted security directly to the GPU using NVIDIA Confidential Computing trusted execution environments, with the goal of protecting customer data in memory and across interconnects during execution.
Google Cloud is also extending Confidential Bare Metal support, NVIDIA says, to A4X Max instances powered by NVIDIA GB300 NVL72 systems. NVIDIA describes the configuration as combining non-virtualized accelerated compute with cryptographic protection for training data and model intellectual property. Google Cloud’s statement in the blog emphasizes hardware-isolated, non-virtualized compute environments and end-to-end cryptographic protection, while positioning the instances for privacy-preserving, high-performance serving.
NVIDIA’s ecosystem language is broad in the hardware and infrastructure it expects to support the update. In the blog, it lists multiple cloud and infrastructure providers supporting NVIDIA Confidential Computing deployments, including CoreWeave, Firmus, GMI Cloud, Google Cloud, IREN, Lambda, Microsoft Azure Infrastructure, Nebius, Nscale, and Vultr. It also cites global systems partners such as Cisco, Dell Technologies, HPE, Lenovo, Supermicro, and Bull as supporting the capabilities through NVIDIA-Certified Systems, and says DSX OS will support all confidential computing modes for cloud providers building AI factories.
Still, the blog does not provide implementation specifics such as performance impact, availability timelines by region, or detailed configuration steps for each deployment option. It also does not disclose any independent benchmarking or certification results in the announcement text, focusing instead on security capabilities at the architectural level. As with many confidential computing features, organizations will likely need to validate fit for their threat models, data flows, and compliance requirements in practical pilots rather than relying solely on the high-level design descriptions.
Why It Matters
- Confidential computing is becoming a differentiator for AI platforms as more model builders seek ways to run sensitive training and inference workloads across third-party infrastructure without exposing models or data.
- A bare-metal option targets cases where customers want non-virtualized isolation or tighter control over the host and platform stack, potentially expanding adoption beyond traditional virtualized confidential environments.
- By tying the update to DSX OS reference architecture and “AI factories,” NVIDIA is indicating a push to make confidentiality features easier to deploy at scale across cloud providers.
- Support across multiple NVIDIA systems and rack-scale configurations suggests the company is aiming for broader hardware compatibility rather than a narrow capability limited to one platform.
Sources
Key Facts
- NVIDIA says Confidential Computing is an infrastructure layer intended to protect models and data while AI workloads run.
- The company is adding “Confidential Bare Metal” as a new deployment option for running sensitive workloads on dedicated third-party bare-metal infrastructure.
- NVIDIA says it is introducing platform security controls to protect data in transit, measure platform integrity, and enforce hardware-level access restrictions.
- NVIDIA describes its security approach as using hardware-rooted isolation and cryptographic attestation.
- Confidential Computing support is expanding across Grace Blackwell NVL72 rack-scale systems and x86-based HGX Rubin NVL8 and DGX Rubin NVL8 systems.
- The update includes ecosystem support from providers such as Google Cloud and Cohere, and lists multiple other cloud and systems partners supporting Confidential Computing deployments.
Technology Related
Netflix posts a profit gain, but investors want proof on hits and ads ahead of January
A stronger bottom line is not yet convincing Wall Street. The market is pressing Netflix to show it can deliver a true global breakout and to fix what it views as shortcomings in its ad-supported rollout, while also demonstrating that revenue momentum is stabilizing.
Armada says it will build “superintelligence” factories using NVIDIA DSX POD, positioning deployments as a months-long rollout
At NVIDIA GTC, Armada announced support for the NVIDIA DSX POD reference design, aiming to help partners assemble standardized, factory-integrated compute systems faster than traditional large-scale deployments.
Tanium and NVIDIA team up to bring enterprise endpoint management to local AI systems
The companies say the integration will extend Tanium’s enterprise-grade device management capabilities to NVIDIA’s local AI infrastructure, aiming to make deployments easier to control at scale.
TrendAI adds NVIDIA Nemotron open models to its enterprise security platform for sovereign and private cloud deployments
The Trend Micro unit says the move is aimed at letting regulated organizations run AI-driven security controls with model and data staying within customer-managed environments.
Nvidia debuts DSX POD, a shipping-container-sized data center concept at GTC Berlin
The company introduced DSX POD at its GTC Berlin event, indicating continued push to make AI compute deployments more standardized and faster to install, even as details of configurations and pricing remain unclear.
Nvidia unveils AI agent safety push and new humanoid robot design elements at GTC Berlin
The company used its Berlin developer conference appearance to highlight software and robotics themes aimed at limiting unsafe behavior in AI agents and advancing humanoid robot concepts.
Volkswagen Group and partners build a shared “physical AI” data flywheel for autonomous-vehicle development, with Microsoft among integrators
Porsche is set to lead the initial implementation of a joint vehicle-data platform, combining computer-vision and AI tooling with cloud and platform integrations aimed at accelerating AV training and validation.
Yahoo Finance column pins hopes on Apple’s next leg after Oct. 28
A market-focused prediction from Yahoo Finance says Apple’s stock could benefit after Oct. 28, citing investor enthusiasm around the company’s innovation push. Apple did not provide details in the cited post beyond general business framing.
Crusoe backs NVIDIA DSX POD, aiming to plug its modular AI factory approach into NVIDIA’s data center stack
The cloud infrastructure provider says its Crusoe Spark platform, built from years of running modular AI factories, will be integrated with NVIDIA’s DSX POD offering, expanding a modular ecosystem for enterprises that want faster deployment of AI capacity.
Nvidia CEO Jensen Huang pitches “AI factories” that can be deployed faster, with new DGX Pod aimed at financing hurdles
In remarks at Nvidia’s GTC Berlin, Huang said the company wants AI factory deployments to take weeks, not months, and introduced DGX Pod as a way to package compute and simplify the path to implementation.