GeeCool
GeeCoolThe 155HL is the embedded arm of Meteor Lake: 16 cores/22 threads, 4.8GHz boost and 24MB of L3 give it a proper CPU presence while AI Boost stays near 13 TOPS. Inside an edge device that inverted split is the right answer: the 16 cores run acquisition and protocol work, and AI is left to first-pass defect screening and anomaly hints — inference light enough that the NPU stays resident without touching the power budget a controller needs.
Its buyers are system integrators building boxes, not DIY users: such devices run for years, AI is one feature rather than the star, and 13 TOPS sits exactly at the embedded floor. A 155HL turning up in consumer channels isn't a bargain — it lacks the muscle for chat models, and serious AI compute still belongs on an H-series or a GPU-equipped machine.