Microsoft has been vocal about the transformative potential of neural processing units (NPUs) in enhancing the intelligence of Windows, even though these advanced chips are not yet part of the official hardware requirements. NPUs, which are typically integrated into the CPU package, serve as a bridge between hardware capabilities and software applications, waiting for the right use cases to become essential.
These units are designed to accelerate local inference while consuming less power, making them a common feature in devices ranging from smartphones to Copilot+ PCs. Microsoft emphasizes that NPUs enable manufacturers to deliver sophisticated AI experiences at more accessible price points. The company states, “What once required compute that cost thousands of dollars can now be done on a device with options that cost hundreds, making the technology more accessible to a wider cohort of people.”
A few OS features require local processing, though none greatly boost productivity.
However, the current utility of local AI processing for everyday users remains limited, a fact Microsoft seems to acknowledge. The company notes that the architecture of NPUs positions Copilot+ PCs favorably for future advancements, but this raises questions about the immediate value for consumers. Aside from enhancing hardware vendors’ profit margins, what tangible benefits do customers gain from having AI hardware in their laptops?
While a few operating system features necessitate local processing, their impact on productivity is minimal. Notable examples include Microsoft’s privacy-focused Recall feature, which allows users to maintain a searchable log of their activities, and the Semantic Windows Search, which enables natural language queries. Yet, many users might prefer that Microsoft first address existing shortcomings in Windows Search functionality.
Microsoft has also begun integrating AI capabilities into various applications, such as Notepad and Photos, which are set to leverage local NPU functionalities. However, these changes have not been driven by user demand, and aside from features like simulated eye contact in Windows Studio Effects, there remains a lack of compelling applications that would justify the presence of NPUs.
AI Agents and Future Prospects
Despite the lukewarm reception, Microsoft continues to advocate for NPUs. The company traces the development of NPU technology back to the Surface Hub 2 Smart Camera, asserting that the latest innovations involve AI agents within Windows. The agent in Settings can already interpret natural language prompts, such as “My mouse is too small,” and provide relevant suggestions. Microsoft envisions that on-device agents will soon tackle more complex tasks.
One of the primary advantages of NPUs is their ability to run multiple AI-driven applications concurrently. Unlike generative AI features that rely on large language models and cloud connectivity, NPUs can support smaller language models directly on the device. Yet, aside from vague references to upcoming features like Click To Do—which will soon allow users to create bulleted lists from selected text—there is little clarity regarding the specific applications that will utilize this capability.
Concerns About Future Hardware Requirements
As Microsoft expands Windows functionalities that depend on NPUs, concerns arise regarding the user experience. The company’s enthusiastic promotion of this technology may be a precursor to an eventual decision to include NPUs in its hardware requirements list. This concern echoes past experiences when Windows 10 machines became obsolete due to stringent CPU and TPM requirements.
Esben Dochy from Lansweeper expressed uncertainty about future AI hardware requirements, noting that companies will evaluate the technology based on workforce efficiency and cost savings. Meanwhile, Marie-Christine Pygott, a senior analyst at Context, observed that while there has been growth in the adoption of Copilot+ PCs, this increase stems more from marketing strategies and price reductions rather than substantial productivity enhancements.
Some buyers are motivated by the desire to future-proof their purchases, recognizing that while a must-have AI application may still be years away, having an NPU now could prevent the need for an upgrade sooner than anticipated. Microsoft has already prompted one hardware refresh with the requirements for Windows 11, and enterprises may justifiably worry that the company could revisit hardware specifications in its pursuit of AI integration.
According to Pygott’s data, AI-enabled notebooks, featuring NPUs regardless of TOPS, constituted 40.5 percent of the European distribution channel as of early September. This figure is expected to rise, making the inclusion of NPU requirements in future Windows versions a plausible scenario. As the landscape shifts towards AI-equipped devices, it is clear that the conversation around hardware compatibility will only intensify, leaving many Windows 10 users with outdated systems to ponder their options.
