Market up on China optimism. TSMC survived lower guidance and actually all stocks up.
New article from SmartAnalyst.
Baird Analyst Tristan Gerra doubt that AMD can reach 10% CPU server market in 2019.
“It takes three quarters for server customers to evaluate a new architecture before ramping incremental volume. Given many of these customers started to use EPYC in 4Q18, market share gains at existing customers will likely be gradual starting in 2H19, leading us to believe a doubling in EPYC volumes in 4Q19 versus 4Q18 may be difficult to achieve.”
This is stupid. Customers started validating with Naples a long time ago. It took a long time because of brand new chip architecture. Now the adoption rate with Rome will be a lot faster. Lisa Su specifically stated that the 2nd 5% market share will be reached much faster than the first 5%.
Furthermore, “We have received feedback some EPYC-based platforms were less stable versus Xeon Scalable.”
“While EPYC’s performance exceeded that of Intel in comparable implementations last year, testing of Xeon Scalable nextgen platform so far suggests performance ahead of EPYC, per our field research.”
Unfounded. Is it ahead of Epyc2 (Rome) or Naples?
I will automatically dismiss this “analysis”, but will many other readers who think they are doing “research”?
According to this article by Steve Burke, AMD may be launching X570 motherboards in June/Computex, at the same time as Ryzen 3000 CPU’s, “reportedly confirmed by multiple manufacturers”. Also, AMD is dropping chipset made by ASMedia, designing the chipset on its own.
This seems likely, because Lisa Su already stated to expect Ryzen 3rd gen and Epyc 2 by mid year.