AI factories are fast becoming the backbone of modern enterprises. These environments power millions of data exchanges every second between models, agents, and applications, but that speed and autonomy also make them harder to secure. With its new integration,
Xage Security is moving Zero Trust enforcement directly onto the
NVIDIA BlueField Data Processing Unit (DPU), ensuring every transaction, identity, and data flow is governed in real time.
Duncan Greatwood, CEO of Xage Security, told ChannelE2E that this step represents a major leap in both scale and performance.
“Inside large data center environments, customers have been deploying our XEP XL for identity-based enforcement. XEP XL manages up to 10Gbps and 100,000 simultaneous flows, with multiple appliances in parallel where needed,” he explained. “With AI factories and our NVIDIA BlueField integration, the hardware-accelerated solution supports up to 800Gbps and millions of managed flows - about 80× faster than existing approaches, with very low latency.”
That acceleration is crucial for organizations running AI workloads that depend on instant reasoning and continuous data exchange. As Greatwood put it, this performance ensures “AI factories can operate securely at the speed and scale needed for real-time reasoning for millions of users and applications.”
Continuous Security Without Friction
A common challenge in AI-driven operations is balancing innovation with security. Continuous enforcement often slows things down or adds layers of complexity. The Xage-NVIDIA integration was designed to remove that friction by running Zero Trust controls at the hardware level, where traffic and decisions happen.
“By running our controls directly on the BlueField DPU, security enforcement occurs at line speed,” said Greatwood. “Policies are enforced and visible in real time across all interactions. Even as AI systems self-evolve, our closed-loop framework keeps policies intact and adapts automatically.”
In practice, this means security no longer trails innovation; it keeps pace with it. Organizations can expand or modify their AI environments without losing visibility or control. Real-time enforcement ensures data and model interactions remain trustworthy, even as autonomous agents make independent decisions or generate new behaviors.
Bringing Zero Trust to Critical Infrastructure
Beyond AI, the integration extends to operational technology and critical infrastructure—sectors where uptime, reliability, and security are tightly intertwined. By embedding identity-based segmentation into BlueField’s DPU, Xage delivers security that scales across everything from data centers to industrial systems.
For sectors like energy, utilities, and manufacturing, this approach helps prevent lateral movement between workloads and isolates potential breaches before they spread. The result is consistent, hardware-enforced protection that supports both IT and OT environments, without adding latency or operational overhead.
Governance, Compliance, and Real Accountability
Security and compliance now go hand in hand, especially as AI regulations tighten globally. Greatwood said the integration was built with governance in mind, ensuring that organizations can meet standards without compromising performance.
“We provide auditable, identity-driven controls aligned to NIST 800-207, EU NIS2, and U.S. federal Zero Trust requirements,” he said. “Every policy change, access decision, and flow-control event is recorded immutably and integrated with SIEM and observability tools via Syslog or REST. Running these controls on BlueField means compliance doesn’t cost performance.”
This approach gives organizations the ability to prove compliance through verifiable audit trails while maintaining the agility needed to keep AI systems responsive and productive. In short, compliance becomes part of the workflow, not an afterthought.
New Opportunities for Partners and MSPs
The integration also creates new opportunities for partners and MSPs. With Xage’s Zero Trust Fabric now hardware-accelerated on NVIDIA BlueField, partners can deliver scalable, AI-ready security architectures across diverse customer environments.
“Partners can deliver line-speed, identity-based segmentation and enforcement across customer environments - from AI models to operational systems - while simplifying rollout through our distributed Fabric,” Greatwood said. “It helps differentiate services, accelerate secure AI adoption, and support clients’ regulatory goals - turning Zero Trust from a checkbox into a business enabler.”
For MSPs, this means more consistent service delivery and stronger security outcomes, backed by the ability to monitor, govern, and enforce Zero Trust policies in real time at the hardware layer
At its core, this partnership signals a shift in how enterprises secure AI. Instead of layering software controls on top of dynamic systems, Xage and NVIDIA are embedding Zero Trust into the fabric of computation itself, making security intrinsic to every data flow, transaction, and decision AI makes.
