Catenaa, Sunday, November 02, 2025- Qualcomm announced plans to launch new artificial intelligence accelerator chips for data centers, challenging Nvidia and AMD in the rapidly growing AI hardware market.
The announcement sent Qualcomm shares up 11%.
The AI200, slated for 2026, and the AI250, planned for 2027, are designed to operate in liquid-cooled, full-rack systems capable of hosting multiple chips as a single computational unit.
The chips are optimized for AI inference, running models efficiently, rather than AI training, which involves massive data processing.
Qualcomm’s data center strategy leverages Hexagon neural processing units, the AI components already embedded in its smartphone chips, scaled for high-performance computing. The company emphasized flexibility, allowing clients to use full rack systems or mix and match components such as central processing units (CPUs) and AI cards.
Rack-scale systems are designed to support 768 gigabytes of memory, exceeding comparable offerings from Nvidia and AMD, while lowering power consumption and operational costs for cloud providers.
Durga Malladi, Qualcomm’s general manager for data center and edge, said the move reflects the company’s gradual entry into larger computing domains after proving itself in mobile and edge AI markets.
Qualcomm also confirmed partnerships with hyperscale customers, including Saudi Arabia’s Humain, which plans to deploy the chips in large-scale data centers.
The AI hardware sector is expected to see nearly $6.7 trillion in capital expenditures on data centers through 2030, with most investment directed toward AI chips.
Nvidia continues to dominate with over 90% market share, but rising demand from AI labs, cloud providers, and startups has created openings for competitors.
Qualcomm’s launch signals intensifying competition in AI infrastructure as companies seek cost-efficient and scalable solutions for inference workloads.
Qualcomm unveils AI200 and AI250 chips for data centers, targeting AI inference workloads and challenging Nvidia and AMD in the growing AI hardware market.
