Key Takeaways
- Microsoft defines AI PCs with local Copilot capability and 40 TOPS NPU performance standard.
- AI PCs require dedicated hardware for AI processing, setting the stage for Microsoft’s AI-powered Windows.
- NPU manufacturers may compete in a TOPS war to meet Microsoft’s requirements for AI PCs and future Windows OS.
We know that companies have been working away at supplying the public with AI PCs, but it hasn’t been easy to tell exactly what an “AI PC” is. Microsoft has stepped up with its own standards, declaring that all AI PCs running Windows must come with a Copilot key. Now, more details on what Microsoft considers an “AI PC” have come to light, including allowing Copilot to run locally on the computer’s hardware and an NPU that can hit 40 TOPS.
Google Pixel 8 Pro review: A giant leap for smartphone AI
The Google Pixel 8 Pro’s generative AI capabilities are legit, and I can’t put the phone down
Microsoft begins to pull back the curtain on what an “AI PC” is
As spotted by Tom’s Hardware, Microsoft is beginning to lay the foundations on what it means to be an AI PC. First of all, the computer must have the capability to run Copilot on local hardware. Right now, if you talk to Copilot, it sends your query back to its servers, which processes your request and sends the response back. With AI PCs coming with special hardware dedicated to running AI-based processes, your local hardware can handle Copilot’s processes. As Todd Lewellen, the VP of Intel’s Client Computing Group said:
“[…]And as we go to that next gen, it’s just going to enable us to run more things locally, just like they will run Copilot with more elements of Copilot running locally on the client. That may not mean that everything in Copilot is running local, but you’ll get a lot of key capabilities that will show up running on the NPU.”
Second, AI PCs are going to need NPUs that can put out 40 TOPS of performance. The NPU is the Neural Processing Unit, a special chip dedicated to AI-based processes that you’ll find in AI PCs. “TOPS” is the current unit of measurement for how fast the NPU can go; it stands for “Trillions of Operations per Second” and the bigger the number, the faster the chip can go. It’s not the be-all-end-all way of comparing NPUs, as you can also measure how much power it uses while processing (measured in TOPS per watt, or TOPS/w), but it’s a good way to set a benchmark for the minimum amount of power an NPU should have.
Microsoft is building toward an AI-powered Windows
With Microsoft setting so many benchmarks for AI PCs, it’s likely establishing the base work for its planned AI-powered Windows operating system. Microsoft needs people to have the hardware ready to go before it releases its new operating system; if it goes too fast, people won’t have the proper hardware to take advantage of the new features. And now that we have a baseline, expect a TOPS war to break out between NPU manufacturers to be the best in the market.
[ad_2]
