Apple, Nvidia Lead AI PC Sectors, Enterprise Market up for Grabs

Apple and Nvidia are frontrunners in the creative and gaming AI PC sectors, while the top spot in the enterprise market remains up for grabs, according to new research from Omdia.

Omdia’s AI PCs: An Emerging Device Class report states that AI PCs are readily available with the market set to expand.

AI capabilities in PCs have been around since at least 2019, but hardware firms are increasingly looking to capitalize on this emerging market opportunity, according to the report.

AMD, Intel, Nvidia, Apple and Qualcomm have already developed or are developing solutions optimized for PCs. Omdia’s report suggests to get ahead, developers should tailor those solutions to meet user needs.

For business users, Omdia said AI PCs need to be suitable for the "Corporate Warrior" persona, a business user who makes heavy use of collaboration tools, is less technically savvy and works both from home and in the office requiring portable hardware.

The report suggests that the business PC user is the biggest and least well-served by Apple and Nvidia.

Intel with its new Core Ultra CPU is heavily targeting neglected business users, according to Omdia. The report suggests, however, that the "Corporate Warrior" user requires a “just works” experience, with flashy experiences likely not top of their list of demands for a PC.

Related:AI PCs Are Going Mainstream Says AMD’s Jason Banta

Credit: Omdia

“User personas are crucial to understanding this space,” according to the report. “Many users will have minimal patience for reduced battery life or more weight to heave around. Their needs must be understood holistically, including things such as form factor, interaction design and APIs.”

Small Models Prove Enticing

Omdia’s report highlights another area that PC developers could target to attract new customers, small language models.

Small language models are smaller-scale AI systems that can perform generation tasks that large language models do, but using a minuscule amount of memory.

Omdia suggests PC developers tap into the likes of TinyLlama and Microsoft’s new Phi-3-mini to entice developers into picking up new PCs.

Small language models like these could, for example, power agents on an AI PC to execute user requests, the report suggests

"You can't underestimate the importance of unified memory to AI on the PC," said Alexander Harrowell, Omdia's principal analyst for advanced computing.

"The limiting factor for the kind of AI model you can even run is VRAM – the very popular seven billion parameter class of AI models, such as Mistral-7B, need at least 8GB – and the limiting factor for performance is memory bandwidth. Arm's investment in its unified memory architecture going back as far as the 1980s is a major advantage."

Related:Small Language Models Gaining Ground at Enterprises