First Ars article I’ve downvoted in a while. They’re not doing NEARLY as thorough of a job on this as L1 or GN.
There is a LOT more going on here than what Intel’s official line here implies, and the additional context makes this whole situation look like Intel isn’t being upfront and forthcoming with some pretty serious details on the failure. The way Intel is playing this is starting to look a lot like CYA, but this issue has been so bad that business partners who would normally remain silent are starting to talk to people like GN and L1 about lithography issues (supposedly fixed since then, but Intel has so far refused to specify date or serial ranges potentially affected, so who knows really) in addition to potential voltage issues. They’re only starting to release info on this now because these two channels are digging in WAY harder than I think Intel expected anyone to bother with.