Edge AI Applications Set to Boom, Driven by AI PC Expansion

TrendForce predicts a substantial surge in edge AI applications in 2024, building upon the foundations laid by AI servers and extending into AI PCs and other terminal devices.
Cisco AI Assistant for SecurityThe global AI server market, covering AI Training and AI Inference, is forecasted to surpass 1.6 million units, experiencing an impressive growth rate of 40 percent. Concurrently, Cloud Service Providers (CSPs) are expected to escalate their involvement in this burgeoning sector.

As edge AI applications evolve, there is a discernible shift towards AI PCs, aiming to decentralize AI server workloads and diversify the spectrum of AI use cases. TrendForce asserts that AI PCs are projected to meet Microsoft’s benchmark of 40 TOPS (tera operations per second) in computational power.

The arrival of new products meeting this threshold, expected to hit the market in late 2024, is poised to fuel significant growth in 2025, particularly following Intel’s planned release of its Lunar Lake CPU by the end of the current year.

The driving forces behind the burgeoning AI PC market are twofold.

Firstly, there is a substantial demand for terminal applications, predominantly steered by Microsoft’s Windows OS and Office suite. Microsoft’s integration of Copilot into the next generation of Windows establishes Copilot as an essential requirement for AI PCs.

Secondly, Intel, a prominent CPU manufacturer, advocates for AI PCs that integrate CPU, GPU, and NPU architectures, facilitating a range of terminal AI applications.

Qualcomm’s Snapdragon X Elite platform, anticipated to meet Copilot standards, is poised to be the pioneer in this domain, with shipments expected in the second half of 2024, delivering around 45 TOPS.

Following closely, AMD’s Ryzen 8000 series (Strix Point) is also expected to meet these requirements. Intel’s Meteor Lake, launched in December 2023 with a combined CPU+GPU+NPU power of 34 TOPS, falls slightly short of Microsoft’s standards. However, Intel’s forthcoming Lunar Lake might surpass the 40 TOPS threshold by the year-end.

The competition among Qualcomm, Intel, and AMD in the AI PC market is intensifying the rivalry between the x86 and Arm CPU architectures in the Edge AI market. Qualcomm’s early adherence to Microsoft’s requirements positions it to seize the initial wave of AI PC opportunities, challenging the x86 camp as major PC OEMs like Dell, HPE, Lenovo, ASUS, and Acer gear up to unveil Qualcomm CPU-equipped models in 2024.

Beyond the immediate impact on hardware, AI PCs are expected to drive growth in average PC memory capacity and elevate the proportion of LPDDR (Low Power Double Data Rate) in PC DRAM.

Microsoft has set the baseline for DRAM in AI PCs at 16 GB. In the long term, TrendForce projects that AI PCs will stimulate an increase in annual demand for PC DRAM bits, with consumer upgrade trends further amplifying this demand.

A noteworthy aspect of CPUs meeting Microsoft’s 40 TOPS requirement for NPUs is the adoption of LPDDR5x over the current mainstream DDR SO-DIMM modules. This strategic choice is driven by the imperative need for faster data transmission.

While DDR5 speeds range from 4800–5600 Mbps, LPDDR5x achieves a significantly higher speed of 7500–8533 Mbps. This choice is crucial for AI PCs, which demand swift language processing and responsiveness.

Consequently, TrendForce anticipates that LPDDR will account for approximately 30–35 percent of PC DRAM demand this year, with future increases propelled by the specifications mandated by AI PC CPU manufacturers, leading to a further surge in LPDDR adoption.

Related News

Latest News

Latest News