ZDNET key takeaways

WSL 3 enhances Linux capabilities on Windows.
Linux gains improved direct connections to GPUs and NPUs.
A native Linux desktop remains preferable for AI developers.

Microsoft advances its Linux desktop integration through the beta launch of the open-source Windows Subsystem for Linux version 3.

During Microsoft Build 2026 in San Francisco, the company unveiled WSL 3 as the latest evolution in its Linux-on-Windows initiative. The feature appears first as a preview and will expand gradually across Windows 11 users. It arrives as a free, independently updatable component via standard WSL channels rather than a distinct product.

WSL 3 represents an architectural shift rather than a complete overhaul. Earlier iterations used syscall translation or a lightweight virtual machine with a genuine Linux kernel. The new release preserves the familiar shell interaction while supporting Linux containers directly.

The core advancement involves streamlined communication between Linux processes and hardware, reducing intermediate layers for Windows devices.

Performance improvements

Microsoft highlights GPU and NPU access free from prior performance penalties. Earlier GPU features in WSL 2 still routed through the virtual machine, introducing overhead that hindered AI tasks even on hardware with dedicated NPUs.

WSL 3 adopts a paravirtualized VM design that combines virtualization with native hardware access. Linux containers now operate directly on Windows without extra setup steps.

This redesign lowers overhead by exposing accelerators straight to the Linux environment. AI tools such as PyTorch and TensorFlow inside WSL approach performance levels seen on bare-metal Linux systems. Microsoft positions WSL 3 as the quickest method for Linux-based AI work on Windows PCs short of a full Linux installation.

Benefits appear first on Copilot+ devices and systems using Qualcomm Snapdragon X Elite, Intel Meteor Lake, or Lunar Lake processors, with AMD support planned later.

The experience stays consistent for users through standard wsl commands, while internal changes improve how accelerators reach Linux workloads.

This aligns with Microsoft's AI PC focus, enabling local AI execution on GPU- and NPU-equipped hardware without requiring a switch away from Windows.

For optimal AI development, a standalone Linux distribution still delivers the best results. Yet for those restricted to Windows by organizational policies, WSL 3 offers a capable substitute.

Users retain flexibility in choosing Linux distributions such as Ubuntu, Debian, OpenSUSE, Kali, or Alpine. WSL became open source in mid-2025, though certain kernel and filesystem elements stay proprietary.

Participation options

Microsoft plans to integrate WSL 3 into the default WSL experience, making it standard on new Windows installations. This transition is anticipated with Windows 11 26H2 later this year.

Access currently requires joining the Windows Insider Program and selecting a preview channel. Once enrolled on a qualifying build, WSL updates arrive automatically through the OS and Store packages.

Steps to participate include joining the Insider Program on a Dev or Beta channel, confirming a recent build, installing WSL via elevated PowerShell, and allowing updates to complete.

Commands such as wsl --version confirm the active release. For those avoiding full Insider enrollment, a standalone pre-release package provides another route through the wsl --update --pre-release command.

Overall, WSL 3 strengthens Microsoft's developer offerings by focusing on accelerator performance and seamless integration. It targets users running Linux-based AI, container, or development workloads who prefer or must remain on Windows.