Palo Alto puts AI Runtime Security on NVIDIA BlueField DPUs

Palo Alto Networks frames the shift from AI that answers questions to agents that take action as a change in what security must protect: an agent's identity, the authority it carries and the actions it takes. Its answer is to move enforcement away from the host. AI Runtime Security on NVIDIA BlueField DPUs sits in the path between agents and the models, data and services they reach, so policies are enforced in silicon and a compromised host cannot be trusted to report accurately. Offloading security from the host node is said to bring two benefits: increased token throughput and stronger runtime protection.

The identity side is being addressed through an integration between agentic identity security and NVIDIA Sentry on BlueField, with the stated goal of a verifiable identity for each agent, access appropriate to its task and attributable actions. For visibility, Palo Alto announced an integration of Cortex XSIAM with NVIDIA DOCA Argus, which monitors host activity from BlueField's independent trust domain and streams alerts for correlation.

Palo Alto also points to partners for deployment. Deloitte's Silicon-to-Service offering and Secure AI Foundations, and WWT's AI Proving Ground, are presented as ways for enterprises to move from individual security components to a repeatable approach. The body does not give customer names, pricing or a date for the identity integration, so those remain open.