AI PCs with NPUs are here and Windows 11 is the modern OS layer that unlocks edge AI on new silicon (including ARM), enabling faster, private on-device intelligence. Competitors using edge AI can gain material productivity advantages. Download our ebook to learn more.
Thanks for downloading – your ebook is on its way to your inbox. In the meantime, you can explore more insights on Datacom’s Microsoft partnership here.
Computing is set for a seismic shift as high-performance computer processors, AI-driven edge computing and the Windows 11 system converge to deliver a new standard for speed, efficiency and real-time intelligence.
Greg Furlong, Datacom’s Associate Director, Strategy and Innovation, and a veteran of hardware makers, including Acer, IBM, Lenovo and Toshiba, outlines the changes in store for computing and what it means for the devices and applications you use every day.
Greg Furlong: AI PCs are a new breed of personal computers designed to run advanced artificial intelligence features directly on the device, thanks to specialised hardware like neural processing units (NPUs). These machines can do things like image generation, real-time speech translation and smarter search directly on the device, rather than relying on cloud computing. That can offer responsiveness, privacy and battery efficiency that hasn’t been achieved before.
Copilot+ PCs are a specific Microsoft-certified category of AI PCs running Windows 11, equipped with powerful NPUs (capable of 40+ trillion operations per second), at least 16GB RAM and dedicated AI-driven experiences exclusive to this hardware. The aim is to move everyday productivity, creativity and collaboration tasks off the cloud and onto the local device for the first time.
GF: Edge AI is the deployment of artificial intelligence algorithms directly on local devices, such as sensors, cameras, smartphones and laptops. It enables real-time data processing and decision-making right where the data is generated, rather than sending it to a central cloud server.
Recent advances in AI have set us up for the next wave of computing – what I call hybrid AI. It’s where you have the ability to run powerful AI workloads on your device, while sending some tasks to the cloud. It’s a highly efficient way of putting AI to work, because it allows models to operate swiftly and autonomously, delivering immediate insights even without reliable cloud connectivity.
GF: ARM, or advanced RISC machines, refers to the architecture and company behind the world’s most widely used family of low-power processor designs. These processors have a different architecture than Intel and AMD, the computer chip makers that power devices worldwide. ARM processors have been in smartphones and tablets for years and more recently in laptops and desktops.
They allow for smaller, thinner and lighter devices with less power consumption. Many ARM processors also feature an NPU, allowing real-time AI inference for tasks like speech recognition, image analysis and generative AI features. Hardware from the likes of Intel and AMD is also advancing and has performance and cost benefits. But ARM-based devices are getting a lot of attention for their suitability to run AI workloads efficiently.
GF: A key benefit is security. When your most sensitive data doesn’t have to leave your device, that greatly reduces the chances of it being stolen or used inappropriately. I believe that factor alone will increase uptake of AI. Beyond security, edge AI offers ultra-low latency by analysing data where it's generated so tasks can be completed faster.
Edge AI continues to function even in areas with intermittent or no internet connectivity, so is ideal for field deployments. By avoiding large-scale data uploads to the cloud, edge AI cuts network usage and ongoing cloud service costs, making deployment more affordable and efficient.
GF: Windows 11 is rapidly emerging as the foundation operating system for edge AI on devices equipped with NPUs for fast, local processing. Microsoft has achieved deep integration between AI hardware and its services. While it is early days in this space, you can already use AI features on newer Windows 11 devices such as Recall (timeline search), Copilot Vision (screen content analysis) and to generate images in Paint.
GF: My opinion is that within the next three to five years, somewhere between 30 and 50 per cent of Windows PCs will be running ARM silicon. There’s currently a slight price premium paid for these devices, which will decline over time. The availability of a broader range of third-party applications drawing on the NPU and edge AI will drive uptake of these devices, and the compact formats and great battery life they offer.
GF: Compatibility gaps are shrinking quickly, but persist in a few areas, such as specialist software, some games and legacy hardware drivers. Windows 11 on ARM offers robust emulation for 32-bit and 64-bit Intel/AMD apps, meaning most legacy Windows software will run, though some performance penalty and battery drain can occur for heavy workloads.
Many core applications, including Microsoft 365, Teams, web browsers (Edge, Chrome, Firefox) and most of Adobe Creative Cloud now run natively and efficiently on ARM PCs. We run a wide range of applications at Datacom and have been impressed with their performance on ARM-based PCs. The emulator is invisible to the user, which is how it should be.
GF: That’s a decision that often isn’t just based on the available technology, but factors in cost, support and existing relationships with hardware vendors. But my recommendation to organisations looking to refresh hardware is to upgrade to devices with NPUs in them, unless there are compelling reasons not to. If you believe, as I do, that we’ll be running AI at the edge extensively in the next three to four years, it makes sense to future-proof your hardware now.
GF: Our businesses can unlock new levels of efficiency, innovation and AI capability with ARM hardware. By drawing on NPUs, they can enable on-device AI for fast, secure and privacy-preserving applications, for applications like automation, predictive analytics and real-time monitoring. That can enhance productivity and enable new business models, so it's worth exploring. The risks at the moment include lingering compatibility issues, and the cost of transitioning to ARM-based hardware.
GF: Two waves of change are coming. The first is the device format changes that ARM-based platforms allow in the form of thinner, lighter devices that use less power. That is going to see more versatile devices hit the market. The second wave is the rise of so-called XR (extended reality) technology, which blends augmented reality with AI.
XR enables users to experience environments that are fully virtual, enhance the real world with digital overlays, or merge real and virtual elements for deeper interaction. It has huge potential in areas like healthcare, education and in how we collaborate and communicate. The new generation of AI-enabled devices will see incredible innovation take place that will make the experience of using a computer feel very different in the years to come.
