Microsoft’s recent event reveals a strategic pivot towards Nvidia-powered PCs, signalling a move away from its original NPU-centric AI vision for on-device computing, with implications for developers and enterprise users.
Microsoft has quietly shifted the centre of gravity in its AI-PC story. At its event on 7 October, the company spent only a brief moment on Copilot+ PCs and far more time on Nvidia hardware, a contrast that underlines how far the message has moved since last year’s launch.
When Microsoft introduced Copilot+ PCs on 20 May 2024, it defined the category around an NPU capable of more than 40 trillion operations per second, along with at least 16GB of memory and 256GB of storage. The promise was clear: on-device AI, low power draw and a new Windows experience that would not depend on the cloud. Microsoft said at the time that the NPU was the key component, not the graphics processor, and promoted features such as real-time translation and image generation as examples of what that silicon could do.
That framing was much less visible in San Francisco. According to Axios, the company used the event to unveil a new Nvidia-powered Surface Laptop Ultra for business users, priced from $2,599, alongside a broader push towards what Microsoft now describes as a hybrid AI model. The machine is built around Nvidia’s RTX Spark chip and Blackwell RTX graphics, and Microsoft is positioning it for developers and enterprises that need substantial local AI performance without relying entirely on cloud computing.
The presentation also made clear that Microsoft’s current priority is not the NPU-first formula it promoted in 2024. Pavan Davuluri, who leads Windows and Surface, described the company’s direction as a PC for everyday AI, but most of the live demonstrations ran on GPU and memory rather than on the kind of NPU found in standard Copilot+ machines. The new workloads shown on stage, including large local models and code tools, require far more memory than the baseline Copilot+ specification, which makes them unsuitable for many of the laptops that still carry the label.
That creates an awkward gap between the marketing of 2024 and the product reality of late 2026. Microsoft has not said that Copilot+ PCs are obsolete, and the company still argues that NPUs remain useful for persistent background tasks such as webcam blur and live captions because they consume very little power. But the new emphasis on large local models, heavier memory configurations and Nvidia GPUs suggests that the NPU is now a supporting component rather than the centrepiece. For buyers who chose a Copilot+ laptop on the assumption that it would be the main platform for Microsoft’s on-device AI ambitions, the company’s latest messaging amounts to a quiet but significant retreat.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
