This is fine, because a NPU is memory bandwidth and compute (compute performance is still based on memory bandwidth) in disguise:
Quote from: 48 and 64 GB RAM nice on January 23, 2026, 11:19:25The Ryzen 350:
Quote from: www.amd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-7-350.htmlOverall TOPS
Up to 66 TOPS (I think it's 8-bit / INT8)
NPU TOPS
Up to 50 TOPS (same)
Quote from: nvidia.com/en-us/geforce/laptops/compareGeForce RTX 5050 Laptop GPU: 440 AI TOPS (4-bit, scammy NGREEDIA, so it's half that -- 220 -- in 8-bit)
GeForce RTX 4050-Laptop-GPU: 194 AI TOPS (8-bit)
194/66 = ~3 times, so it's 3 times slower.
3dmark.com/search:
4050 (notebook): Average score: 8288
Ryzen AI 350' 860M iGPU: Average score: 2885
8288/2885 = ~3 times, which is the same 3 times.
-> Looks like Ryzen AI 350' NPU has to be mainly understood as its iGPU, really. An iGPU is still an ASIC, the most power efficient way. Maybe a NPU is just marketing, instead of just saying it the way NVIDIA says it ("AI TOPS", no mention of a NPU).
Which tells us that it has been all along about what I said in my previous comment ("it's all about memory size, memory bandwidth and the usually, out of it, resulting GPU performance") ;)
And for running various AI LLM models locally and privately (see article: notebookchat.com/index.php?topic=315954.0 ("Your own ChatGPT, offline: AI without the cloud on your laptop")), a NPU does not need to be used anyway.
PS: It's almost like the comment from January either predicted or gave AMD an idea to release non-AI Ryzen. Tho, binning chips is not new and why throw these perfectly fine chips away, just because of NPU PR reasons? Certainly not.
PPS: Let me predict another thing: "Ryzen AI" will remove its "AI" in its product name later. It will still be AI capable, but as proven in the quote, its -- and generally speaking as well -- AI capabilities come from its iGPU's compute and that, in turn, comes from it's memory bandwidth. Memory size is the other one.