
Intel and Qualcomm have a lot in common: >Large tech company with minimal exposure to AI chips >Not their 1st attempt at it (Nervana, Gaudi, AI 100) >Announced new chips for 2026 with lots of LPDDR (Crescent Island 160GB, AI 200 768GB) with hardly any details or performance targets >Tries to stay relevant by scaling up IP from legacy businesses (PC iGPU, Mobile NPU) >Wants to compete on TCO with LPDDR because they've missed the boat with HBM















