Dive Brief:
- AI-capable PCs and the chips that power them took the spotlight at the annual Consumer Electronics Show in Las Vegas on Monday, as AMD, Intel, Nvidia and HP rolled out their latest product lines. Hardware manufacturers are banking on AI to stoke enterprise demand for high-capacity compute laptops, workstations and devices
- “AI PCs are starting to deliver real value across a wide range of everyday tasks, from content creation and productivity to intelligent personal assistants,” AMD Chair and CEO Lisa Su said during a Monday keynote. Su showed off AMD's latest Ryzen AI PC processors, the Instinct MI440X 8-GPU cluster for on-prem enterprise AI workloads and the Ryzen AI Halo Developer Platform for small form factor desktops.
- AMD competitor Nvidia countered with upgrades for its GeForce RTX, RTX Pro and DGX Spark devices and announced the imminent delivery of the Nvidia Rubin AI processing platform, while Intel rolled out its Series 3 processors. Intel’s Core Ultra Series 3 chips will power more than 200 AI devices, including an AI-capable notebook HP unveiled along with a compact Copilot-enabled PC.
Dive Insight:
As the race to embed AI capabilities across product lines heats up, the line between AI PCs and traditional workstations is blurring.
Nearly one-third of PCs shipped last year were equipped with AI-optimized NPUs, according to Gartner. The firm expects AI PC to account for more than half the market this year and become the norm by 2029.
Enterprises and the service providers that help integrate and manage their IT estates are moving to AI PCs for more than just raw processing power, Kieren Jessop, PC and tablet research manager at Omdia, told Channel Dive. Devices are part of a broader AI adoption strategy.
“Enterprises aren't buying AI-capable PCs exclusively for the NPU, since many of the AI features being marketed aren't exclusive to devices with NPUs,” Jessop said. “The center of gravity for enterprise AI is still in the cloud or on-premises servers.”
More than half of organizations running AI workloads on-prem plan to add AI-capable devices to the mix in the next year, according to Omdia research. Among firms using on-device infrastructure for AI, nearly two-thirds aim to shift additional workloads in that direction as momentum shifts to the edge, Jessop said.
The PC industry got a boost last year from Microsoft, which ended support for Windows 10 in October, helping to trigger an enterprise refresh. Global PC shipments increased nearly 9% year over year in the third quarter, according to IDC’s market analysis.
IDC expects demand for Windows 11 PCs to carry over into 2026 but raised the specter of a memory chip shortage slowing market momentum in the months ahead. AI usage bumps up PC memory requirements as well as price points.
“Just as the industry is seeing a need to add more RAM, it has become prohibitively expensive to do so, even if they can get supply,” IDC said in the December post.
While semiconductor giants touted the compute power of the newest GPU and NPU configurations at CES, some vendors are already tweaking the narrative, according to Jessop.
“Apple and AMD are highlighting unified memory architectures as a key differentiator for AI workloads, with configurations in the hundreds of gigabytes of available memory,” Jessop said. “These systems can run much larger models locally than traditional discrete GPU architectures allow, at comparable prices.”
The pivot shows that memory matters more than compute speed for many enterprise use cases, added Jessop.
Disclosure: Informa, which owns a controlling stake in Informa TechTarget, the publisher behind Channel Dive, is also invested in Omdia. Informa has no influence over Channel Dive’s coverage.
