https://www.fool.com/investing/2017/10/04/intels-strategy-to-fight-nvidia-is-now-unclear.aspx
7 replies (most recent on top)
@nrz - 'shoestring' refers to the post-product support and industry development, not silicon dev costs, as you correctly point out. Software counts in HPC.
Shoestring budget? $hundreds of millions wasted on uncompetitive architecture and bloated dev costs. Intel wants to play a "no holes" defense. Not gonna work anymore to do a lot of average or worse things. Time to pick. Intel culture makes picking near impossible. Too many cooks, too many marketers in charge, and poor relationships and trust of and by customers. Time to be a foundry while there's still time.
Like most new projects, Xeon Phi is destined to fold because Intel does not have the knowledge nor the heart to develop anything fundamentally new. Nvidia spent many years cultivating the HPC SW industry - Intel needed to cultivate the same but gave the HPC division a shoestring budget. There were really good people in there, but the user community ALWAYS hedged their bets with NVIDIA products.
Ashraf reports nothing new here - I'll repeat my assessment - Nice Guy. Abysmal writer.
Intel has the one trick pony NVDA sorrounded from all sides. Mobile Eye, Movidius, Altera, Xeon Phi. NVDA has no where to run.
Qualcomm has shot itself in the foot anyway.
Only AMD has some chances. But likelihood of them having missteps in execution is always very high.
?
Intel has failed at everything but ther monopolistic x86 and now AMD and ARMy with process parity have scaled the walled garden and next year will have process superiority.
What is the next ACT 🤡
Fool wrote this