Consumer

Microsoft Copilot+ and The New AI PC

Microsoft has announced Copilot+ PCs, but are they all ARM? Don’t forget about existing Intel and AMD AI PC owners.

Microsoft has unveiled Copilot+ PCs, a new category of computers equipped with AI-driven capabilities designed to enhance productivity and creativity. These PCs feature deep integration with Windows 11 and Microsoft 365 Copilot+, enabling advanced AI functions like personalized assistance, real-time collaboration, and intelligent insights. Microsoft emphasizes that Copilot+ PCs are optimized for both individual and professional use, aiming to transform the computing experience with powerful, user-centric AI tools.

 

When Intel and AMD announced  “AI PCs” for Windows late last year, it was not clear what the NPU would be doing. At the 2023 Lenovo Tech World in Austin, Lenovo said they were working on a model that would run everything locally. It was presented as a “Personal AI” where data never exited the PC. In December, Intel started shipping the first CPU’s for “AI PCs”, the Core Ultra “Meteor Lake” CPUs featuring built-in Neural Processing Units (NPUs). We reviewed some of these, but we (and some of our friends) struggled to find a compelling use for the NPU. The anticipation of the local version of Copilot was a big selling point for these AMD and Intel NPU-enabled chips.

Here Be Dragons

The upcoming Copilot+ PCs from major brands like Acer, Asus, Dell, HP, and Lenovo will exclusively feature Qualcomm Snapdragon X series CPUs. These processors require Windows on Arm, which is still rapidly evolving. It has come a long way, but still has more to go. Microsoft continues enhancing Windows on Arm, including native support for both Microsoft and third-party apps and a new Prism emulator for x86 app compatibility.

Processor Options Description
Qualcomm® Snapdragon® X Elite X1E-80-100 12 cores up to 3.4 GHz, Dual-Core Boost up to 4.0 GHz, NPU up to 45 TOPS
Qualcomm® Snapdragon® X Plus X1P-64-100 10 cores up to 3.4GHz, NPU up to 45 TOPS
AI Accelerator Neural Processing Unit (NPU) with over 40 TOPS on NPU alone.

We have reviewed and traveled with the Lenovo X13S. This is the last-gen ARM laptop that has the Qualcomm Snapdragon 8cx Gen 3. It is long in the tooth these days and just a tad too small. We are looking forward to the new lineup, like the ones announced this week at Dell Technologies World, and have high expectations for the new Snapdragon Elite.

Closing Thought and Feelings

For those of us who invested in “AI PCs,” it is immediately concerning. Current AI PC models, like those we use with Intel Core Ultra and 8000G series AMD with built-in NPU, don’t meet the new performance requirement for Copilot+ PC certification. The certification requires 40+ TOPS (trillion operations per second). Intel can almost reach 10 TOPS. However, it is unclear if Copilot+ will run on the NPU, but won’t meet the certification for the branding of “Copilot+ PC.”

We reached out to and spoke with several people from Intel, AMD, and Dell. The gist of the feedback we got was inconsistent about the compatibility of the new local Copilot AI models with the current lineup of NPU’s and Copilot+ branding.  Though not completely clear, the responses we were able to obtain fell into two categories.

Update: We recieved clarification that Copilot+ features like Recall among several others will require a 40TOPS or beter NPU. Once we have a complete list we will update it here.

  • First, that Copilot+ is basically just a branding certification, and that everything announced by Microsoft will still run, just slower, lack the certification, and that your shiny new AI PC you have right now will be useful still.
  • Second, that the local models will run but some features may not be available or able to run (I was pointed to lack of on-NPU memory as the problem in this case).

One of these is very reassuring for those who have bought in early to the Core Ultra NPU, and AMD 8000G APUs, the other less so. Either way, cause for concern is warranted. Stay tuned, we will keep looking at this and post an update once it is clear which direction this is going to take.

Engage with StorageReview

Newsletter | YouTube | Podcast iTunes/Spotify | Instagram | Twitter | TikTok | RSS Feed

Jordan Ranous

AI Specialist; navigating you through the world of Enterprise AI. Writer and Analyst for Storage Review, coming from a background of Financial Big Data Analytics, Datacenter Ops/DevOps, and CX Analytics. Pilot, Astrophotographer, LTO Tape Guru, and Battery/Solar Enthusiast.

Recent Posts

Micron 9550 SSDs Feature Speed and Power Efficiency

With its top-tier performance and power efficiency, Micron's PCIe Gen5 9550 SSD is expected to set new standards for AI…

4 days ago

IBM Expands Enterprise AI Offerings with Mistral Large on watsonx.ai

IBM has announced the addition of Mistral Large to its watsonx.ai platform, reinforcing its commitment to making high-quality, performant foundation…

1 week ago

AWS Graviton4-Powered EC2 R8g Instances Go GA

AWS EC2 R8g instances powered by the latest Graviton4 processors are GA! (more…)

1 week ago

VAST Data Platform Earns High-Performance Storage Solution Certification for NVIDIA Partner Network Cloud Partners

VAST Data achieves certification as a high-performance storage solution for NVIDIA Partner Network cloud partners. (more…)

1 week ago

Giga Computing Unveils Advanced AI and Cooling Solutions

Giga Computing, a subsidiary of GIGABYTE, recently introduced solutions to tackle complex AI workloads and improve energy efficiency through advanced…

1 week ago

Azure Stack HCI: Microsoft’s Hybrid Cloud Infrastructure Solution, Release 23H2

Azure Stack HCI represents Microsoft's cutting-edge approach to hyperconverged infrastructure (HCI), offering a seamless blend of on-premises computing power with…

1 week ago