Edge Microvisor Toolkit 3.1 — What’s New for Edge AI Developers

[magerstam]

Edge Microvisor Toolkit 3.1 — What’s New for Edge AI Developers?

The Edge Microvisor Toolkit (EMT) 3.1, release brings significant improvements to the platform—targeted at developers building secure, AI-powered applications at the edge.

With this release, we’re making it easier than ever to deploy edge and AI workloads. EMT is a lightweight, container-host operating system designed to unlock the full potential of Intel® platforms. Built on the Intel® 6.12 kernel, EMT provides an optimized foundation for running demanding workloads at the edge.

Let’s explore the technical enhancements and what they mean for your deployments.

Overview of Key Enhancements

1. Updated Base: Azure Linux Merge

This release syncs with Azure Linux 3.0.20250423, capturing upstream improvements in stability, performance, and security.(GitHub)

2. Enhanced ISO Installer with Docker Variants

The ISO installer has been revamped for usability and versatility. It now supports four installation modes:

• Standard Kernel + Core Packages
• Real-Time Kernel (RT) + Core Packages
• Standard Kernel + Core Packages + Docker
• Real-Time Kernel + Core Packages + Docker(GitHub)

Download the ISO

Why this matters:

Enabling Docker directly in the image simplifies containerized workload deployment—ideal for AI workloads that run in Dockerized environments at the edge.

3. Validated Hardware & Networking Standards

This release includes validation and native support for:

• Discrete GPUs: Intel B850 dGPU, NVIDIA P100 and 3090
• Time-Sensitive Networking (TSN): Essential IEEE 802.1 standards (AS-2011, Qav, Qbv, Qbu/3br, TBS)
• Intel RDT (Cache Allocation Technology): Enables fine-grained resource management(GitHub)

Impact for AI workloads:

Edge AI applications can now leverage GPUs and TSN, improving inference performance and network reliability in deterministic environments such as autonomous systems and industrial automation.

4. New Agents: Discovery & Reporting

One new agent deliver critical enhancements:

Device Discovery Agent: Automatically identifies hardware resources present on the device. These tools help with runtime hardware awareness and operational visibility—especially useful when using the toolkit within a fleet managed by the Edge Manageability Framework (EMF).

5. Component Version Updates & Bootkit Support

Key platform components have been updated for improved compatibility and functionality:

• AMT, cJSON, cluster-agent, hardware-discovery-agent
• Helm, kernel, linux-firmware, node-agent, os-update (renamed from os-ab-update)
• platform-observability-agent, platform-telemetry-agent, platform-update-agent, QEMU(GitHub)

Additionally, Edge Microvisor Bootkit support has been added, unlocking possibilities for custom secure boot flows and bootloader extensibility.

6. Documentation Enhancements

This release also includes enriched documentation—covering:

• New demos and usage instructions
• An improved architecture diagram
• Clarified release policy for developers(GitHub)

What This Means for Edge AI Infrastructure Developers

• Faster DevOps: Docker-enabled ISO simplifies testing and rapid iteration for edge AI workloads.
• Hardware-Accelerated AI: GPU and TSN support let you run low-latency, high-throughput inference.
• Secure, Immutable Deployments: Leveraging Bootkit and updated OS agents supports scalable, secure OTA fleet updates.
• Greater Insights: Discovery and Reporting Agents enhance observability and resource management.

A Demonstration Vehicle for What’s Possible

EMT is not intended as a commercial OS with service-level agreements. Instead, it is a demonstration vehicle — a reference toolkit that shows what’s possible when Intel hardware features are fully exposed and optimized for edge and AI workloads.

Our long-term strategy includes:

🚀 Major feature releases every 3–4 months to keep pace with emerging workloads.

🔧 Active maintenance for 18–24 months per release.

📅 Yearly upgrades to the latest Linux LTS kernel, ensuring EMT stays aligned with the upstream ecosystem while continuing to unlock Intel platform innovations.

This cadence ensures EMT remains a living showcase for developers and partners who want to experiment with cutting-edge Intel technologies at the edge.

Try It Today

Experience how simple it is to deploy your Edge & AI workloads with EMT. Download the new EMT-D 3.1 ISO and start building today: 👉 Download EMT 3.1 ISO

References

Release Notes

Documentation

Demos

We’d love to hear your thoughts, plans, and technical experiences with this release—drop a comment below!

Read the original article: https://www.linkedin.com/pulse/introducing-edge-microvisor-toolkit-emt-31-mats-agerstam-auq8c/?trackingId=N26r0B%2FHRsWWVtmVoTQxDQ%3D%3D