I don't really hard-disagree with any of that; we're all guessing. The co. sending chiplets where they fetch the most also doesn't seem outlandish as a factor,

I don't really hard-disagree with any of that; we're all guessing. The co. sending chiplets where they fetch the most also doesn't seem outlandish as a factor, though. FWIW, here's the price-per-core chart for the whole second-gen server line: https://www.servethehome.com/wp-content/uploads/2019/09/AMD-EPYC-7402P-in-7002-Series-Cost-Per-Core.jpg https://www.servethehome.com/wp-content/uploads/2019/09/AMD-... (Doesn't factor in voltage/freq needs for different SKUs, some of them needing fully-working chiplets, etc. Still.) The low pricing on the 7282 and a few others is interesting, too. Wonder if it's a factor that it can use partially-working chiplets since the server I/O die can take up to eight chiplets and the client die only two, or if that's totally unrelated.