The first Windows PCs built on Nvidia processors will debut next week at Computex in Taipei and Microsoft Build in San Francisco, according to Axios. Nvidia-powered PCs are expected from Microsoft’s Surface brand and from Dell, marking Nvidia’s biggest push into the mainstream PC market.
The hardware announcement came with coordinated social media teasers. The official X accounts of Windows, Nvidia, and chip design firm Arm posted “A new era of PC” alongside what appeared to be geographic coordinates for Taipei, where Computex opens this week.
Scout and Autopilot: Local Agent Software
Alongside the hardware, Microsoft is preparing to unveil new agent software for Windows. A purportedly leaked screenshot shows an app called Scout, described as an OpenClaw-like agent, with a new agentic layer called Autopilot built in, according to Indian Express citing an Axios report.
The screenshot shows a user asking the agent to send daily briefings pulled from their inbox, calendar, and Teams, then triage email and draft replies for review. Scout is expected to consolidate Microsoft’s scattered Copilot tools for chat, coding, and the Claude Cowork task assistant into a single destination. It is unclear which AI model will power the Autopilot harness, or how Autopilot and Claude Cowork will coexist within the same interface.
The Chip Landscape Shifts
Until now, Intel and AMD have dominated the Windows CPU market, with Qualcomm supplying Arm-based processors for select laptops. Nvidia’s entry adds a fourth contender to the Windows silicon race. Reuters first reported Nvidia’s plans to design Windows-compatible Arm CPUs in 2023, according to Indian Express.
The timing follows Apple’s March 2026 MacBook refresh with M5-series chips. Microsoft’s partnership with Nvidia appears aimed at narrowing Apple’s lead in power-efficient, locally capable hardware.
Agents Move to the Edge
The announcement extends Microsoft’s agent strategy from cloud infrastructure (Azure, Copilot Studio) to local hardware. NCT previously covered Microsoft’s Build 2026 software roadmap, which includes taskbar-based autonomous agents, MCP integration at OS level, and a WinUI 3 agent framework. The Computex announcements represent the hardware side of that vision: Nvidia silicon optimized for running AI agents locally rather than routing every request through cloud APIs.
For agent builders, the split between Google’s cloud-native approach (Gemini Spark, which runs even when the device is off) and Microsoft’s local-execution bet (agent tasks on Nvidia hardware) defines two competing architectures. Which model wins may depend less on raw capability and more on whether enterprises trust their agent workloads to run in someone else’s cloud or insist on keeping them on-premises.