Docker's new VMM: faster containers, smoother workflows

Docker just shipped a major under-the-hood upgrade most developers won’t notice, until they do. The new Docker VMM replaces third-party virtualization with a purpose-built engine, and the difference shows up in real-world use: faster container starts, smoother file sharing, and memory that actually gives back when you’re done with it.

Why rebuild the wheel

For years Docker Desktop leaned on someone else’s virtual machine monitor. That worked, but it also meant Docker couldn’t fix the things that annoyed users most: sluggish first launches, memory leaks, and the occasional crash that took the whole app down. By owning the stack end-to-end, Docker can now optimize for container workloads instead of generic VMs.

What got faster

• Container startup times dropped across the board, including the first run after install.

• Switching between projects no longer feels like waiting for a VM to boot.

• File sharing between host and container is snappier, especially in edit-compile-test loops.

• Memory management improved; idle containers release RAM back to the host.

Windows users get a better deal

Docker VMM on Windows aims to deliver strong isolation without the performance tax. Early tests show it closing the gap with WSL2, which has been the go-to for Windows container devs. The opt-in setting in the latest Docker Desktop release lets teams test it without breaking existing workflows.

Beyond the desktop

This isn’t just a point release. Docker VMM provides the foundation for other Docker runtimes, including those used in cloud environments. The long-term plan is a unified runtime that works the same on laptops, cloud instances, and on-prem servers, reducing surprises when you scale up.

When and where you can use it

The new VMM is live in the latest Docker Desktop release for macOS (enabled by default) and Windows (opt-in). Linux support is under development, with details to be announced in future updates. If you’re on Windows, flip the switch in settings and see if your containers feel lighter. On macOS, you’re already running it.

The bottom line

Docker rebuilt its virtualization layer so you don’t have to think about it. The payoff is containers that start faster, use memory smarter, and behave more predictably across platforms. That’s the kind of upgrade that saves time without asking for any.