this post was submitted on 09 Nov 2023
4 points (100.0% liked)
Hardware
33 readers
1 users here now
A place for quality hardware news, reviews, and intelligent discussion.
founded 11 months ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Server side from almost non-existence to close to 25% in about 5 years is actually very impressive, given the specific market tends to be rather conservative and has a lot of inertia to steer for your favor.
The server market is the fattest segment, Intel needs to get its $hit together, this is very alarming.
And considering AMD has roughly double the number of cores per CPU as Intel, AMD probably has a much higher share of CPU cores and of revenue than the 20-ish % market share in sockets sold.
Fat in what way?
This year intel has sold around $6-8B in client chips and $4B in server chips per quarter, give or take some hundreds of millions. At best years they were about equal at the time when intel was able to charge whatever they wanted for server chips. On AMD server chips now outsell client chips if you don't count gaming segment stuff ($1.6B vs $1.5B respectively last quarter), but that is mostly because of very anemic client sales.
When you look at the best offerings of the two companies, AMD is offering their 96 core 3d v cache EPYC at almost a third less price than Intel's 56 core sapphire rapids, and everything starts to make sense
Then you realize few customers actually buy either of those. Most sales are in the ~30ish core products. And it starts to make sense why AMD doesn't just capture the entire market but is still sitting at 25% even though the 96 core chip is so great.
The problem with this line of reasoning is that we don't buy CPUs in datacenter - we buy full servers from 3rd party suppliers. Look at a company such as CDW.
Most server customers aren't massive hyperscalers that need to maximize computer per rack-U. Xeon servers are plentiful, can be found for cheaper than their CPU MSRP's would suggest, and if we're looking for a 16 - 24 core model to spin up a branch office, a lot of times we don't even really care whether it's SPR or Epyc. There're other factors like "Is my LOB app certified by the vendor to run on Epyc"? etc.
A lot of times, when I need to order, a lot of the Epyc servers may be on backorder, or may be comparable pricing, or maybe I specifically need Xeon because I already have a Xeon Hyper-V server and want this server to be able to work in the event of a failover (best to keep your VMs on the same platform).
Hell, in Windows Server, Epyc only got support for nested virtualization with Server 2022, so it wasn't even a consideration when we did a big refresh a few years back.