Industry observers note that now, PCs with dedicated GPUs will have access to those functions. An under-the-hood change in Windows seems to signal the further deterioration of Microsoft’s Copilot+ branding, which, at least historically, depended solely on NPUs as the engine of local PC AI.
According to the latest update, the Windows App SDK 2.2 governs functions like text summarization, text rewriting and editing, code generation, and others. It might also allow the Microsoft Photos app’s upscaling or Super Resolution capability to run on PCs with a moderately powerful GPU, as well as enabling “erase” and object extraction using AI. An experimental drop of the Windows App SDK on Github now allows certain AI-specific functions to run on Nvidia RTX GPUs, rather than solely depending on an integrated NPU.
In a fresh development, but Microsoft’s Copilot+ PCs originally looked first for the NPU and virtually ignored the GPU when supporting those functions. That’s changing, apparently. For over two years, we’ve known that GPUs generate the most AI TOPS within the PCs, even if the NPU is the most efficient AI engine for functions like Windows Lab Effects, Paint’s Cocreator, and more.
As part of the ongoing story, you’ll also need a PC running a Windows Insider build. You’ve got to enable Developer Mode, too. That’s dedication. For now, Microsoft’s “support” for GPUs requires downloading the experimental version of the WinAppSDK 2.2 Experimental 9 and pairing it with an Nvidia GeForce RTX 30 series GPU with at least 6GB of VRAM.
In a fresh development, still, as Windows Current points out, it’s less about what it enables and more about what it signals: Microsoft continues to chip away at what was a monolithic approach to AI–the NPU or bust.
In a fresh development, likewise, Microsoft’s experimental Windows App signals that it’s trying to bring its AI capabilities to a broader range of Windows 11 devices, a more common-sense approach. In an interview with Andrew Hill, Microsoft’s corporate vice president of Surface, at Microsoft’s Build conference last week, my colleague Alaina Yee noted how the publisher was becoming increasingly open to running local AI models where it made sense and relying on the cloud for other AI tasks when needed.
The report highlights that millions of older PCs lack an NPU but have access to a robust GPU. A move like this is a no-brainer, even if it may take a few months to become more broadly available to the Windows user base at large. As much as some segments of the internet tend to frown on generative AI as a replacement for creation, the ability to use AI as a tool seems to be much more broadly accepted.
The report highlights that he has authored over 3,500 articles for PCWorld alone, covering PC microprocessors, peripherals, and Microsoft Windows, among other topics. Mark has written for publications including PC Magazine, Byte, eWEEK, Highly adopted Science and Electronic Buyers' News, where he shared a Jesse H. Neal Award for breaking news. He recently handed over a collection of several dozen Thunderbolt docks and USB-C hubs because his office simply has no more room. Mark has written for PCWorld for the last decade, with 30 years of experience covering technology.