Xage Security

Xage Security announced a strategic partnership with NVIDIA integrating NVIDIA DOCA security and NVIDIA Vera BlueField-4 STX for zero trust solutions in agentic AI

Leadership read: Xage has moved zero-trust enforcement from the network perimeter to the silicon layer. Embedding policy logic directly into NVIDIA's BlueField-4 STX data-processing unit means access decisions and visibility run at line speed inside the AI infrastructure fabric itself, not as a software overlay sitting above it. That is a materially different security architecture than what Xage shipped before this partnership: the company has now committed to a hardware-coupled delivery model, which reshapes its integration surface, its customer deployment lifecycle, and the technical depth required to sell and support at AI factory scale. This is one of twelve partnership signals we have tracked in the last 90 days, though the comparable set is diffuse, spanning export programs, consumer goods, and defence, and none maps cleanly to critical-infrastructure cybersecurity. The more useful frame is the narrower one: hardware-anchored security integrations at the AI infrastructure layer are accelerating as enterprises build out GPU clusters and agentic workloads that existing perimeter tools were never designed to protect. Cadence's AI-infrastructure collaborations from the same period point to the same pull: silicon vendors and software specialists are converging on co-engineered solutions because neither side can credibly address the problem alone. Across companies operating at this convergence of OT/critical-infrastructure security and AI-infrastructure scale, the pattern concentrates demand in a specific functional cluster: deep-stack systems engineering capable of working across hardware and software stacks, commercial leadership with experience selling into regulated and critical-infrastructure buyers, and solutions architecture that can bridge the gap between AI platform teams and security operations functions.

curated · 2026-06-01 · context →