Key Takeaways
- Xage Security integrates with NVIDIA’s DOCA and Vera BlueField-4 STX to enhance AI security.
- The combined solution provides real-time visibility, policy enforcement, and control for AI agent interactions.
- Xage offers a comprehensive Zero Trust framework to secure critical data and workflows in AI environments.
New Era of AI Security
Xage Security has announced its support for the advanced NVIDIA DOCA security capabilities and the NVIDIA Vera BlueField-4 STX platform. This new collaboration aims to establish a robust security framework specifically designed for AI environments, facilitating secure data storage and protecting real-time interactions of AI agents. NVIDIA Vera BlueField-4 STX utilizes DOCA security to deliver secure-by-design AI storage, embedding real-time protective measures directly into silicon. This approach empowers organizations with trusted agentic AI while ensuring operational efficiency and performance are maintained without modifications to existing systems.
Duncan Greatwood, CEO of Xage Security, emphasized the evolving role of AI factories as critical infrastructure. He highlighted the need for organizations to maintain “unbypassable visibility” over their AI agents, which are increasingly accessing sensitive information. By integrating with NVIDIA’s security architecture, Xage aims to instill a foundational layer of Zero Trust security within AI factory operations, crucial for safeguarding sensitive workflows and data.
Securing AI Interactions
Xage Zero Trust for Agentic AI enables businesses to transition AI agents from pilot projects to full production with enhanced security measures. The solution provides complete visibility and controls throughout the AI interaction lifecycle, encompassing various platforms and resources. In contrast to traditional security measures that focus solely on outputs or prompts, Xage’s approach regulates access and actions for each AI agent, ensuring comprehensive governance.
The partnership with NVIDIA’s BlueField-4 STX enables Xage to implement detailed security policies in real-time, detect potential threats, and secure workflows at scale. Key components such as Xage Resource Gateway, Xage Agent Sentry, and Xage Extended Protection will function within the NVIDIA DOCA architecture, ensuring AI resources and agents are secured effectively.
Comprehensive Integration for Maximum Security
Xage’s collaboration with NVIDIA DOCA involves multiple integration points that enhance overall security. The Xage Resource Gateway can connect with NVIDIA DOCA Vault, providing visibility and control for file access. Additionally, the Xage Fabric’s policy engine uses DOCA-generated information for real-time decision-making regarding agent interactions.
Furthermore, Xage will work seamlessly with DOCA Argus and DOCA Flow, sharing critical event data for improved threat detection capabilities. For hardware-based control, integration with DOCA Open vSwitch (OvS) allows Xage to manage permitted interactions while blocking unauthorized ones. Xage can also respond to detected threats in real-time, containing potential risks by isolating suspect agents or users.
The concerted efforts of Xage and NVIDIA DOCA create a closed-loop security framework that offers visibility into runtime behaviors, evaluates identity and policy compliance, and enforces necessary controls to mitigate business risks.
Empowering Organizations with Enhanced AI Security
With the support of NVIDIA Vera BlueField-4 STX and DOCA, Xage equips organizations to effectively monitor and manage AI agents at both the infrastructure and workload levels. Key benefits include enforcing least-privilege access, applying granular Zero Trust policies, and enhancing visibility throughout the AI lifecycle.
Xage Zero Trust for Agentic AI represents a significant advance in securing interactions among AI agents, applications, and critical resources. This deep integration supports enterprises, government agencies, and critical infrastructure operators in adopting AI technologies with increased security, resilience, and control.
The content above is a summary. For more details, see the source article.