An accurate (but incredibly biased by hindsight) analysis. If this were a matter of ML, I'd say you way over-fitted to the working data. It would've required f

An accurate (but incredibly biased by hindsight) analysis. If this were a matter of ML, I'd say you way over-fitted to the working data. It would've required forseeing the sea-change in the smartphone/mobile market produced by the iPhone in order to predict that it would explode the market for high performance ARM chips so wildly. This wasn't in the cards at the time. At the time, everything but the Blackberry was a joke. Anybody in a position to direct capital, with the faculty for this kind of prescience, would then know the real best use of capital in that circumstance is to direct it at Apple, not for AMD to make ARM chips. It likely would've been more profitable for AMD to sell off the entirety of its assets and buy Apple stock in 2005 than for them to try to jump on the AMD bandwagon. More to the point, AMDs brief reign as the gamer/hacker/tinkerer chip maker of choice was a byproduct of the Pentium 4 architecture, especially the Prescott gen, being such an unmitigated disaster. The moment Intel rolled back to the Core Solo/Duo series based on the Pentium III architecture, it was clear they were going to retake the lead and probably keep it. All AMD did to "succeed" was not fail as hard at chasing the megahertz demon. In particular, by choosing not to sacrifice pipeline efficiency for higher clock rate, AMDs chips were significantly more performant for virtually every use-case except SSE-heavy vector operations.