Microsoft has announced a new category of Windows laptops designed to run its Copilot chatbot directly on the device, eliminating the need for a constant internet connection. The company made the announcement on Monday, introducing what it calls Copilot+ PCs. Every unit sold this year will include a specialized chip known as a neural processing unit, or NPU, which is engineered to handle the computational demands of a large language model without sending data to remote servers.
The NPU represents a fundamental shift in how the company’s AI assistant operates. Since Copilot launched in 2023 as a replacement for the defunct Cortana, it has relied on OpenAI’s GPT-4 and GPT-5 models and required a live internet connection for every interaction.
The new hardware changes that dynamic. With the NPU, key features will work offline, latency is reduced, and the assistant remains available even when the network is not. Microsoft first signaled its intentions at the Build 2023 conference, where it announced plans to integrate Copilot into Windows 11 through the taskbar.
That move was software-based. The current announcement is hardware-based.
The NPU is a dedicated processor designed to perform the mathematical operations required by large language models without draining the battery or slowing the main CPU. For developers, this enables local inference. For users, it means the assistant does not pause while waiting for the network to respond.
The timing of the announcement follows a broader consolidation of Microsoft’s AI products. The company introduced its chatbot as Bing Chat in February 2023 and later unified its various AI offerings under the Copilot brand by the end of that year.
The “copilot” analogy—an AI that works alongside the user—has been central to that strategy. Monday’s announcement completes that vision by placing the AI brain inside the computer itself. The implications for the user experience are significant.
A cloud-dependent assistant is a shared resource that can be slow and subject to outages. It cannot access local files or understand offline workflows.
A local assistant, by contrast, is personal, fast, and always on. This changes what developers can build and what users can expect from their machines. For the PC industry, the stakes are high. Laptop sales have been flat, and traditional arguments for upgrades—such as a new processor or a thinner chassis—have lost their appeal.
An NPU that runs AI locally provides a genuinely new reason to upgrade. It also creates a lock-in effect: once users depend on local Copilot features, switching to a machine without an NPU means losing that capability.
Microsoft did not disclose specific partners, pricing, or which models will ship first. The company stated only that these are the first Windows laptops built from the ground up for local AI. The proof of concept will depend on developers writing software that uses the NPU and users noticing the difference.
For now, Microsoft has drawn a clear line. Before Monday, Copilot was a cloud service.
After Monday, it is a hardware feature. The distinction between a connected PC and a standalone AI device has become thinner.
