When the world is being rewoven around Artificial Intelligence, can our ubiquitous PC be left behind. So as expected it is right out there. Having missed out on the mobility tech bus quite sometime back, Microsoft has been the first mover in any new technology hitting the market, and AI is the apex of it all. The OpenAI Microsoft connect we all know a bit too well by now. The birth and sustenance of CoPilot we are all aware of. Now it is the bold vision to completely rebuild Windows PCs around AI.
The debate as to where should AI reside; cloud or your device, did not rage for Microsoft, as the tech giant has quietly chosen both through a unique hybrid AI approach. Windows is betting on a hybrid AI approach that blends both local and cloud together. Corporate VP of Windows and Devices Davuluri says what is critical is the ability to compute and provide context appropriately for the underlying experience, whether it be client based, cloud based or both. AI compute needs to be seamlessly available.
The approach has been that tech should adapt to the user and not the vice versa. To make this vision the reality, Microsoft raised the bar for what was possible to run sustained AI workloads oin a device. This thought process was clearly visible when Microsoft introduced Copilot+ PCs last year and with it established a 40+ TOPS NPU as the new performance benchmark for AI PCs. Microsoft has a fair idea of the trajectory of how AI and AI-compute silicon were evolving. The crux of the matter is to have a requirement that was scalable and still push what was possible on the client silicon.
With on-device NPUs can run locally and “eliminate some subscription fees while delivering enhanced privacy and performance.” There can be no denying the fact that small language models, SLMs, will continue to advance, Copilot+ PCs shift users toward highly capable AI they truly own. Renting through cloud services should be an exception rather than a norm. How does it happen? Microsoft’s Copilot PCs integrate next-gen NPUs from AMD, Intel, and Qualcomm. They are engineered for a clear cut straight purpose – engineered to offload and accelerate complex AI tasks locally.
IF MOST OF THE AI COMPUTE IS UNDER YOUR COMMAND, THE WORK AUTOMATICALLY BECOMES SEAMLESS.
Sanjay Sahay
Have a nice evening.