Microsoft and NVIDIA Team Up for Secure AI Agents on Windows

Key Takeaways

Microsoft and NVIDIA Team Up for Secure AI Agents on Windows

Microsoft and NVIDIA have introduced new tools to support the development of secure, on-device AI agents on Windows. These tools include Microsoft eXecution Containers (MXC) for enhanced security and NVIDIA OpenShell for runtime integration. NVIDIA RTX Spark desktops and laptops deliver powerful AI performance with 1 petaflop capability and up to 128 GB memory. Microsoft offers a developer edition preloaded with tools for AI development.

Windows Platform Security for AI Agents

Microsoft is focusing on security for AI agents on the Windows platform. The company has introduced Microsoft Execution Containers (MXC) and a composable sandbox to ensure containment, identity, and manageability for AI agents. This allows developers to build and deploy AI agents with confidence, while maintaining control and trust.

Run Local AI Agents with Faster Models and Multi-Node Clustering

NVIDIA has introduced updates to its DGX Spark system software, which enables developers to run local AI agents with faster models and multi-node clustering. The updates include a streamlined installation process for NemoClaw and support for multi-node clusters with full ConnectX-7 performance.

Cisco Cloud Control for Secure Agentic AI

Cisco has introduced Cloud Control, a unified operations platform that brings together networking, security, AI infrastructure, observability, and collaboration. The platform provides a secure harness for agentic AI, allowing agents to act on real systems without breaking them or exposing them to risk.

Cisco Live: Cisco Unveils Cloud Control, AI Canvas, and Quantum Security

Cisco has unveiled several new technologies at Cisco Live, including Cloud Control, AI Canvas, and a quantum security push. Cloud Control is a unified platform that spans Cisco's networking, security, and observability portfolios, while AI Canvas is a multiplayer, generative environment for human operators and AI agents to investigate and resolve complex issues.

Ainekko's Edge AI Silicon Platform Goes Open Source

Ainekko's CORE-ET Silicon Platform has been made an open source project by the OpenHW Foundation. The platform provides a collection of open hardware and software building blocks for fast, energy-efficient AI inference at the edge.

NVIDIA Partners with Robotics Companies for Humanoid Robots

NVIDIA has announced plans to accelerate development of humanoid robots with an open humanoid robot reference design. The design combines a Unitree H2 Plus humanoid robot, Sharpa five-fingered hands, and NVIDIA Jetson Thor onboard compute for advanced reasoning and control.

DOD Focuses on Cyber Integration and AI Security

The US Department of Defense is prioritizing the integration of cyber operations into all its activities and ensuring security in AI usage. The DOD's top cyber policy official emphasized the importance of considering cyber effects in operational planning and building security into AI tools from the beginning.

Microsoft Aims to Make Users Addicted to AI Assistant

Microsoft's internal documents reveal that the company plans to make users addicted to its new AI assistant, Scout. The goal is to make people dependent on the tool before adding new features.

MediaTek to Hire More Staff for AI and Chip Development

MediaTek, a Taiwanese chipmaker, plans to hire more staff to boost its push into new AI spheres and chip development. The company has good order visibility for its new data center business over the next few years.

Has AI Transformed Religion?

The Pope has released an encyclical titled 'The Soul of AI,' which examines the intersection of technology and spirituality. The encyclical argues that AI has the potential to both enhance and diminish human existence, depending on how it is used.

Report Warns of Potential for Exploitation by AI Agents

A report from the Open Markets Institute warns that AI agents may work on behalf of corporations rather than users, highlighting the potential for exploitation. Policymakers must ensure that AI agents work for users or that users understand when AI agents are working for corporations.

Supermicro and Arm Partner for Energy-Efficient AI Solutions

Supermicro has collaborated with Arm to deliver energy-efficient rack-scale infrastructure for enterprise agentic AI. The new solutions feature Arm AGI CPUs and are designed to maximize compute performance within power envelopes and physical footprints of enterprise data centers.

Sources

• Build Personal AI Agents on Windows PCs with New Tools from Microsoft and NVIDIA
• Windows platform security for AI agents
• Run Local AI Agents with Faster Models and Multi-Node Clustering on NVIDIA DGX Spark
• Cisco Cloud Control: The Secure Harness for the Agentic Era
• Cisco Unveils Cloud Control, AI Canvas, Quantum Security Push At Cisco Live
• Ainekko’s Edge AI Silicon Platform is Now an OpenHW Foundation Open Source Project
• Nvidia has a robot partner. It isn’t Tesla.
• DOD wants to integrate cyber in all operations, and integrate security into AI
• Microsoft Wants to 'Make People Addicted' to its New AI Assistant, Internal Documents Reveal
• MediaTek Pledges Hiring Spree to Boost Push Into New AI Spheres
• Chatbots And The Soul: Has AI Transformed Religion?
• Who Do AI Agents Work For? New Report Warns of Potential for Exploitation
• Supermicro Collaborates with Arm to Deliver a New Class of Energy-Efficient Rack-Scale Infrastructure for Enterprise Agentic AI