Qualcomm and Apple's ARM-based chips, which have leapt in computing power, will bring severe challenges to the x86 system in the AI ​​PC field.

Qualcomm and Apple have some very compelling ARM-based solutions that will give AMD and Intel a run for their money in the AI ​​PC market in 2024. We've seen a recent industry-wide shift toward ARM-based solutions, especially with the potential for emerging competitors to adopt ARM.

As Apple increases its share of the ARM architecture in the laptop sector, we may see widespread adoption of this chip architecture in the PC industry in the future, mainly because ARM is able to close the performance gap with x86 architecture and is more power efficient.

ARM already dominates the mobile industry, with Qualcomm and MediaTek using the standard for quite some time.

However, ARM's market share in the PC industry will rise rapidly in the coming years, as Qualcomm has launched its own PCSoC, the "SnapdragonXElite", which boasts impressive performance and is expected to hit shelves in mid-2024. In addition, it is rumored that NVIDIA and AMD may also launch ARM-based CPUs in 2025, which means that the x86 architecture will face fierce competition.

Currently, the Hexagon processor unit used by Qualcomm in SnapdragonXElite has reached an astonishing 45TOPS. Intel, on the other hand, seems to be playing hard to catch up and is not expected to reach comparable performance levels until the launch of the Lunar Lake platform in 2025.

As OEMs are keen on launching AI roadmaps on their devices, ARM processor-based PC makers have an opportunity to grab a piece of this emerging segment. However, Intel still believes that in most use cases, their current product designs are sufficient to meet user needs for artificial intelligence tasks.

In addition, Apple’s M3 chip reached 18TOPS, which is only slightly higher than the M2’s 15.8TOPS. Although lower than the data of Qualcomm and AMD, Apple has a higher degree of control over its software and hardware ecosystem, which will help the Mac series of computers provide a comparable AI experience with lower hardware specifications.

The next-generation AIPC market is undoubtedly the target that all chip manufacturers are currently eyeing. All companies include dedicated NPUs in their chips to accelerate AI workloads and bring additional capabilities through robust software stacks and ecosystems.

Qualcomm has now announced that its AI engine will deliver 75TOPS of performance on the fastest XElite SOC, while AMD has just announced that its upgraded Ryzen 8040 "HawkPoint" APU will deliver 39TOPS of performance (16TOPS from the NPU). Its successor product will be launched in the second half of 2024, code-named StrixPoint. Its AITOPS is expected to increase by 3 times, and XDNA2NPU alone can reach nearly 50TOPS.

Apple's M3 SOC is also capable of delivering around 18 TOPS, and while this number is lower than other competitors, Apple's software ecosystem running on its own operating system offsets the requirement for a powerful AINPU through optimization. Finally, Intel is also trumpeting that its NPU will be installed on the Core processor "Meteor Lake" that will be unveiled on December 14 next week.

Recently, ARM introduced the Cortex-M52 SoC, which comes with the company's "Helium" or M-Profile vector extensions to significantly improve the performance of machine learning (ML) and digital signal processing (DSP) applications. As the PC industry will be massively impacted by artificial intelligence capabilities, companies like ARM will need to deliver high performance in this area.

It will be interesting to see how the influx of ARM chips changes the dynamics of the PC industry, especially the AI ​​PC segment, as it will not only bring more diversity but also lead to more competition in the market.