Satya Nadella and Jensen Huang beside a display of Windows laptops
Satya Nadella and Jensen Huang beside a display of Windows laptops
+ AI News

How Microsoft is changing Windows for Nvidia's RTX Spark laptops

Microsoft is tuning Windows for RTX Spark's local agents, shared memory and Arm apps. Copilot Plus features and some game support arrive later.

Microsoft is changing Windows to make local agents easier to install and large AI models easier to run on Nvidia’s RTX Spark laptops. Its work also includes tuning x64 app emulation for Nvidia’s Arm CPU and managing the memory shared by the CPU and GPU, according to Tom Warren’s report in The Verge.

The software arrives in stages. Agent installation is part of the launch experience, while Copilot Plus PC capabilities such as Recall are due later. That makes the particular Windows feature a buyer needs as important as the laptop’s hardware specifications.

Easier agent setup, with an early snag

Microsoft’s new Get Started experience lets RTX Spark owners choose and install an agent in two clicks, principal program manager Peter Dawoud told The Verge. The Windows Store handles agent installation and updates. Hermes, Perplexity and OpenClaw are the first officially supported agents, with Microsoft working to add others.

Warren found Hermes largely matched the two-click setup claim. OpenClaw required an uninstall and reinstall after it complained about a missing local gateway. His early use also found large local models slower to answer prompts than cloud models. He suggested background or overnight tasks as a possible use, while saying more testing was needed to judge the agents’ value beyond coding.

Memory capacity and app compatibility

RTX Spark combines an Arm CPU, a Blackwell GPU and unified memory. Microsoft has tuned memory management so large models can use much of that capacity while Windows and other apps remain responsive. Surface Laptop Ultra configurations can reach 128GB, according to the report.

Nvidia’s Windows app also has a slider for adjusting how much memory goes to the GPU. The report points to its use on 24GB and 32GB machines, where users may want to change the allocation for particular apps or games.

Apps without native Arm versions use Microsoft’s Prism emulator. Dawoud says Microsoft tuned translation of x64 software for RTX Spark’s CPU design. He also says the Windows team ships compatibility improvements every month.

Gaming support shows why that work extends beyond emulation. The report lists Fortnite and Counter-Strike 2 among titles supported at launch. Call of Duty’s Ricochet support is planned for 2027, while EA targets early 2027 for Battlefield’s Javelin anti-cheat. Microsoft is also bringing advanced shader delivery to RTX Spark so the Xbox app can preload shaders during a game download.

More Windows features are still coming

Microsoft expects Copilot Plus PC experiences on RTX Spark to first reach Windows Insiders later this year, Windows marketing executive Adrienne Brewbaker told The Verge. Warren reports that full availability appears likely next year. Recall is among the capabilities absent at launch.

Microsoft is also developing a faster Windows Search interface with actions such as switching Bluetooth on or off from the search bar. The report gives no firm release date for that interface.

Sources

The AI Feed Desk

The AI Feed Desk

Editorial desk

The AI Feed Desk tracks AI provider updates, model releases, agent tooling, and enterprise adoption, turning fast-moving announcements into source-linked context for builders and operators.

Noticed a typo, incorrect information, or translation error?

Tell us so we can fix it.

Help Improve This Article

Related Articles

Satya Nadella and Jensen Huang standing beside a display of Windows laptops

Microsoft and Nvidia launch RTX Spark PCs, with laptops priced up to $7,000

RTX Spark laptops start at $2,599. The $7,000 model is an Asus, while the buying case against a MacBook Pro depends on memory and the software you run.

The AI Feed Desk

By The AI Feed Desk

NVIDIA's official COMPUTEX article card for RTX Spark and local AI agents

NVIDIA announces RTX Spark PCs for local AI agents

RTX Spark puts 1 petaflop of AI performance and up to 128GB of unified memory into Windows PCs designed for local agents.

The AI Feed Desk

By The AI Feed Desk

A pull request review panel connects to Jira cards and a guarded plugin marketplace

GitHub turns Copilot governance into a code-review and Jira story

GitHub's June 25 Copilot updates pair cheaper code-review analysis with Jira agent visibility and stricter controls over plugin marketplaces.

The AI Feed Desk

By The AI Feed Desk

A Copilot routing panel sends developer tasks either to automatic model selection or to private model providers

GitHub makes Copilot routing more automatic while opening BYOK in the app

GitHub's latest Copilot changes push Free and Student users toward automatic model routing while giving Copilot app users bring-your-own-key model providers.

The AI Feed Desk

By The AI Feed Desk

Bar charts comparing LFM2.5-24B-A2B prompt processing, token generation and memory use against three other models on Apple and AMD hardware

Liquid AI releases LFM2.5-24B-A2B for fast on-device agents

Liquid AI's 24B mixture-of-experts model uses 2B active parameters for local reasoning and tool calls. Its agent workflows differ from d1's single-pass decisions.

The AI Feed Desk

By The AI Feed Desk