Anaconda brings native Python to Windows on ARM and Surface Laptop Ultra

Native Python replaces emulation

Until now, Python developers on Windows on ARM have depended on emulation, Anaconda says. Anaconda and Miniconda ran through Prism, the translation layer that lets x64 apps run on ARM chips. The company says this worked, but every instruction went through an extra step, and packages with hardware-specific code did not always cooperate.

Native support removes that step. The Python runtime, the conda package manager and the packages in Anaconda's trusted repository are now built for ARM64 and run directly on the processor. Anaconda argues this matters most for AI work, because local AI depends on a long chain of software, from Python down to the libraries that talk to the GPU. When every link is native and trusted, the company says, the hardware can do what it was designed to do.

Kilo Desktop and local models

Developers can manage the tools in Kilo Desktop, which Anaconda describes as an agentic engineering tool for developers and data scientists building AI applications and analyzing data. It lets users import and build environments, manage local models, and serve both to their device.

Environments can be imported from a single environment.yml file, the format conda uses to describe a workspace with its own Python version and packages. A teammate with the same file gets the same setup, so a project built on one Surface device can be rebuilt on another machine with less guesswork, according to Anaconda.

Once packages and models are downloaded, everything runs on the device and no data is sent to the cloud. Users load an open-weights model and start a local model server with a few clicks. The server uses the widely adopted OpenAI-compatible API, so existing code and tools can point at the local model by changing a single address.

From new device to working app

Anaconda outlines a workflow for a new Surface device:

• Install the native Windows on ARM build of Kilo Desktop and access Anaconda packages.
• Create the project environment from a single environment.yml file.
• Serve models locally from Kilo Desktop.
• Build a Python app with Anaconda packages, such as a Streamlit dashboard over a local dataset that answers questions using the model on the device.

The company says every part of the stack stays on the laptop: the Python runtime, the packages, the model, the agent, the data and the finished app. It also points to Microsoft's new agent sandboxing in Windows on ARM, which it says gives developers a complete path on one device to explore, design and develop a production AI agent or application running locally.

What it means for developers and teams

For individual developers, Anaconda says the result is a laptop that can run real AI workloads locally, useful for prototyping, for working with sensitive private data, or for building on the go.

For teams, the company says it adds reproducibility and control. Environment files keep setups consistent across devices, and sandboxing lets teams give agents access to real projects while limiting what they can reach. For the wider Windows on ARM ecosystem, Anaconda says native packages unlock richer capabilities and building possibilities.

Anaconda's native Windows on ARM support is available now. The company links to Kilo with Anaconda packages for Windows on ARM, its vetted packages for the platform, and Microsoft Execution Containers on GitHub.