Qualcomm's AI250 Attacks the AI Inference Memory Bottleneck | Durga Malladi Interview
The TWIML AI Podcast with Sam Charrington · 14:50
Qualcomm is moving into rack-scale data-center AI inference with two product generations. The AI200 is a direct liquid-cooled rack system built on its Hexagon NPU, sampling in 2026. The AI250 adds a near-memory comput...