Datacentric business reached an all time high, approaching 50 percent of revenue. For every iPhone/Android smartphone you buy, for every Netflix and Amazon Prime capable player or TV, you need additional servers. This is what Intel is good at, and, until Epyc, there was almost no competition in that space for years.
When it comes to the non X86 competition IBM has a large customer base there and Qualcomm is trying to get customes excited about its ARM based Centriq 2400 server part too.
Fudzilla has been saying this for years, more phones, more internet-based businesses equals better server sales.
As Mike Magee thinks, PC is something that Intel is abandoning, slowly but surely. There's been no growth in PC for years now, it has been shrinking but it will never completely disappear. This is was to be expected, but again, enterprises mainly do work on PC based devices.
Now for the painful part, the 10nm shrink is something that Intel is having real trouble fashioning. It did a great job optimizing things in 14nm having the last four PC generations on the same 14nm process as it could not get the reasonable yields at 10nm.
10nm starts in 2018, volume is 2019
Intel’s CEO has now told investors to expect low volume 10nm in 2018. This is no surprise, and this is what Intel always did when moving to a new manufacturing node. It starts with some small chip, something like Core Y, and then it moves the bigger cores and server chips to the new manufacturing nodes too. This is yet another way of admitting that things are delayed from the original plan.
Now, as for 10nm higher volume, Intel's CEO Bryan Krzanich said that this can be expected in 2019. Intel was more focused on fixing its security problems as this was the highest possible priority for them. Damn Google researchers making all that trouble for Intel!
Still, Intel is optimistic, it had a record first quarter and claims it is on track for Intel’s best year ever. Go figure, even without big bang with IoT, it still managed to make a record year, it claims.