While we know many Windows 11 24H2 rumors include upcoming AI features, namely via Microsoft’s Copilot assistant, there hasn’t been much concrete information from Microsoft on its plans for AI in Windows. However, Intel just revealed a few basic requirements from Microsoft for OEMs to build an AI PC.
In an Intel press briefing with The Verge, system requirements mentioned for a Windows PC to be classified as an “AI PC” include a Neural Processing Unit, the latest CPUs and GPUs, access to Copilot, and — shockingly — a dedicated Copilot key on the keyboard.
Todd Lewellen, head of the PC ecosystem at Intel, told The Verge that “From an Intel perspective our AI PC has Core Ultra and it has an integrated NPU because it is unlocking all kinds of new capabilities and functions in the AI space. We have great alignment with Microsoft, but there are going to be some systems out there that may not have the physical key on it but it does have our integrated NPU.”
So Intel believes that laptops equipped with Intel’s new Core Ultra chip and an integrated NPU are AI PCs, regardless of whether they have the dedicated Microsoft Copilot key, but Microsoft’s definition would say these laptops aren’t AI PCs. As long as a PC has Copilot access, it seems frivolous to require a dedicated key to classify it as an AI PC.
Future AI PCs might require 40 TOPS of NPU performance
While The Verge learned basic requirements for current-gen AI PCs in the Intel press briefing, Tom’s Hardware chatted with Intel at its AI Summit in Taipei about how the AI PC requirements might change over time, with next-gen AI PCs possibly requiring 40 TOPS of NPU performance to classify.
Intel’s Todd Lewellen told Tom’s Hardware that “there’s going to be a continuum or an evolution, where then we’re going to go to the next-gen AI PC with a 40 TOPS requirement in the NPU,” followed up by this statement: “We have our next-gen product that’s coming that will be in that category.”
This 40 TOPS requirement will allow many of Copilot’s key capabilities to run locally (rather than in the cloud), which will result in lower latency, better performance, and benefits to privacy. However, many current NPUs fall shy of that 40 TOPS requirement. Intel says its next-gen product will be in that 40 TOPS category, but its current Meteor Lake chip only offers up to 10 TOPS for the NPU.
AMD‘s Ryzen Hawk Point platform, for comparison, offers up to 16 TOPS for its NPU. That’s slightly better than Intel’s current-gen offering, but both fall short of the 40 TOPS next-gen requirement.
Qualcomm’s upcoming Snapdragon X Elite chip, releasing in laptops as soon as mid-2024, will supposedly support a whopping 45 TOPS in its NPU, surpassing the next-gen AI PC requirement with a current-gen product.
According to The Verge, Intel is expanding its new AI PC acceleration program to help software developers create AI-powered features for apps to utilize the full capabilities of Intel’s AI-powered processor and Microsoft’s Copilot assistant. In theory, the more optimized apps and hardware for Intel’s AI PCs before Qualcomm’s Snapdragon X Elite laptops roll out, the better (for Intel).
But with rumored performance of the Snapdragon X Elite, even compared to Apple’s current-gen M3 chip or next-gen M4 chip, optimized apps and hardware for Intel’s AI PCs might not be enough to keep Qualcomm from leading the competition.
