Back to timeline

Thu, October 1522:30ToolsAI codingAgentsInfra & costAI coding guide

Microsoft optimizes Windows for Nvidia RTX Spark laptops, first to support Hermes and two other Agents

Decision Brief

What changedMicrosoft optimized Windows for Nvidia RTX Spark laptops with Arm architecture and Blackwell GPUs, first supporting three local Agents—Hermes, Perplexity, and OpenClaw—with game support via the Prism emulator, unified memory management, and anti-cheat adaptation.
Why it mattersAgents install and update via the Windows Store with a two-click launch flow, and unified memory tops out at 128GB with a slider to adjust GPU share, letting teams running local open-source models keep large models resident in laptop memory.
Who should careAI coding tool users, Inference / infra teams
Affected stackCopilotNVIDIA
Builder actionUpdate Copilot to the latest version
Source confidenceMedium · Reliable media or first-hand reporting

Microsoft made several adaptations for the RTX Spark platform: a new Get Started experience appears on first entry into Windows, letting users find, install, and open the Agent they need in two clicks. The first officially supported Agents are Hermes, Perplexity, and OpenClaw, with more to come. Agents install and update uniformly through the Windows Store, which handles dependencies. Under the hood, x64 apps run via the Prism emulator with ongoing tuning for the chip's microarchitecture, and Microsoft says it ships compatibility updates monthly; unified memory reaches up to 128GB, with a slider in Nvidia's Windows app to manually adjust how much is allocated to the GPU. For gaming, titles with anti-cheat engines like Fortnite, Counter-Strike 2, and Dota 2 work out of the box, while Call of Duty, Battlefield, PUBG, and League of Legends will be supported around 2027. For developers and teams doing local inference, 128GB of unified memory plus the slider means large models can fit entirely in memory, but in hands-on testing large local models still respond noticeably slower than cloud models, making them better suited to background or overnight tasks. For those coding on Arm laptops, OpenClaw's hands-on install wasn't smooth, repeatedly reporting that the local gateway wasn't installed correctly and requiring uninstall-reinstall to work, showing stability gaps in these Agents' first-run experience. Also, Copilot Plus features like Recall aren't available at RTX Spark launch, expected to reach Windows Insiders later this year.

Summary basis: full article readCompiled from the source scope noted above; the original remains authoritative.

Sources

  • The Verge:AI

    Consumer AI products, platform shifts, and policy/culture impact.

  • The Verge:AI

Related intel

留言

登入后即可留言,和其他 builder 交换实测心得。

还没有留言,抢头香。