How to take 50% market share in three years.

Page 5 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

maddie

Platinum Member
Jul 18, 2010
2,736
697
136
Since AMD has more money now and uses TSMC and GloFo, which means they need to adapt the masks and design to a different process anyways, I thin a selerat desktop version is very likely - server tsmc, desktop glofo
The increase in R&D sums will mainly begin to be seen next year at the earliest, and only for the shortest lead time projects.

Adapting a design to 2 processes is a far cry from a different design altogether. It also allows more robustness into their risk models. Same design at 2 fabs allows major failures without sinking them, even if they would still need to validate the GloFlo one, which they might slowly start to do anyhow as CYA project. AMD still is and will be for quite some time in an unstable position. Any major mistake can remove all gains made.
 

moinmoin

Senior member
Jun 1, 2017
922
406
106
More rumors about Intel on Semiaccurate.

Intel has no chance in servers and they know it

https://semiaccurate.com/2018/08/07/intel-has-no-chance-in-servers-and-they-know-it/
"Once Rome comes out in Q2 of 2019 or so, AMD never has a performance lead of less than 25% per socket, and that is the worst case for AMD vs the best case for Intel, fudged a lot in Intel’s favor on top of that."
"Intel knows their position. One of the documents says in no uncertain terms that the company understands they will not be competitive in the server market until AFTER Sapphire Rapids, the 2022 server part. AMD has a clear run in Intel’s core market for at least 4 years."

It's interesting that with Rome AMD is essentially turning Intel's market sequence on its head. Whereas Intel serves the mobile -> desktop -> server markets in that order with each generation, AMD already serves mobile after desktop, and with Rome Zen 2 will arrive to the server market ahead of both.

Unless Intel changes the sequence as well AMD will hurt them most in the server market and we will see Intel turn around the fastest in the mobile and/or desktop market (whenever they solve their 10nm node).

Edit: On Intel dropping prices and lowering their margins now that they have serious competition:
"Intel can’t drop prices, their ~3x price increase from Broadwell-EP to Purley is the only reason their financials are so solid. The ~$7000 price gain from Purley to Cascade, coupled with a steep rise in the percentage of MSRP that Tier-1 customers pay, says Intel knows they can’t touch margins."
 
Last edited:
Dec 28, 2013
53
0
66
After the SA article from Charlie how fucked Intel is I am more willing to believe in the 64c version of Rome - however I am still extremly sceptical about the uncore version an it still seems far fetched that AMD can pull that if from Gen1 to Gen2
AMD has 'leapfrogging' Zen designs teams. The Zen team and the Zen 2 team started at the same time. It's easily possible Zen 2 was started based on Zen and then switched two years on to an uncore design, likely already in early design for Zen 4, to double the dies AMD could get from a 7nm wafer in anticipation of the increased demand resulting from Intel's massive and continuing 10nm screw up.
 
Last edited:

itsmydamnation

Golden Member
Feb 6, 2011
1,862
255
136
uncore still mokes zero sense........

Why increase latency, when you can decrease it by just using a more advanced memory request tracking. Look at intels scailable data fabric, by making the home agents distributed they reduce bottlenecks on system throughput. keeping memory controllers local to a chip is better for so many reasons. There partners already have the design IP needed to scale the PHY's and double bandwidth on the PCI/GMI links.
 

Gideon

Senior member
Nov 27, 2007
489
136
136
A bit of a sanity check to all the Intel gloom&doom here. AMD only increased market share to 1.3% in Q2 or about 0.3% from Q1.

I disagree with the notion in the article that the growth will continue to be linear, but it does make reaching 5% this year quite hard.

I think we will see much quicker adoption with EPYC2, but yeah 20-25% would be a very very good result in 2-3 years.
 

beginner99

Diamond Member
Jun 2, 2009
4,196
248
126
AMD physically can't produce enough volume to hit 50% server market share lmao
Intel is called Chipzilla for a reason, and TSMC would need to ramp their operations HUGELY to be able to produce 50% of that while also supplying everyone else.
Exactly. This has always been the issue for AMD even in the Athlon vs P4 days.

Besides that even getting there will be hard. Simply said "No one ever got fired for buying intel". A lot of people that will make these decisions are clueless. Even technical IT people. We greatly overestimate how much the average even programmer or other IT person actually knows about hardware or CPUs. Most in fact do not.

As for OP, price is nearly irrelevant. Performance/watt matters a lot more and total ownership cost of the server. Companies buy servers not CPUs. Performance/watt however looks good for AMD.
 
Feb 23, 2017
640
586
106
Slight caveat on that 1.3% figure noted above; it is a revenue figure and not a volume figure, and since AMD pricing is lower than Intel, it suggests that their volume share is a little higher.
It's an interesting distinction simply because as AMD's volume share increases, the total revenue for the entire market decreases. AMD does not need to get to 50% revenue share in order for Intel to lose 50% of it's current server market revenue; the damage would be done well before AMD gets to a 50% revenue share.
 

Husky55

Junior Member
Jun 3, 2003
19
0
81
I have been an Intel server and desktop user for a very long time. I did use some AMD cpu for cheap NAS builds in the past but 95% of my builds were Intel. So with the newer products from AMD, be they are Ryzen, threadripper or EPYC I am really happy for the competition. I agree chipzilla is a tough competitor and a very rich one at that, but I love it when we have some real competition going on. What does it matter what the % of market share is as long as we all benefit from better products at lower prices!!!
 
Jul 19, 2016
135
0
71
A bit of a sanity check to all the Intel gloom&doom here. AMD only increased market share to 1.3% in Q2 or about 0.3% to 1.3%.
Funny man you are, that's a 400% increase.

If the increase 100% each q in 2018 they pretty much get the 5%
 
Sep 22, 2007
33,195
68
126
Exactly. This has always been the issue for AMD even in the Athlon vs P4 days.

Besides that even getting there will be hard. Simply said "No one ever got fired for buying intel". A lot of people that will make these decisions are clueless. Even technical IT people. We greatly overestimate how much the average even programmer or other IT person actually knows about hardware or CPUs. Most in fact do not.

As for OP, price is nearly irrelevant. Performance/watt matters a lot more and total ownership cost of the server. Companies buy servers not CPUs. Performance/watt however looks good for AMD.
This. The OP definitely doesn’t understand corporate IT management pratices. AMD will get some big wins, but they won’t hit 50% market share within 3 years.
 
Feb 23, 2017
640
586
106
1% to 1.3% is an increase of 30% not 400%.
It didn't increase by 1% FROM 0.3% TO 1.3%.
It increased by 0.3% FROM 1% TO 1.3%.
I do wonder where people learn to read, or indeed how to do math.
 

ondma

Senior member
Mar 18, 2018
472
146
61
And in other news, the Cleveland Browns are going to win the Super Bowl and Hell just froze over.
 

maddie

Platinum Member
Jul 18, 2010
2,736
697
136
This. The OP definitely doesn’t understand corporate IT management pratices. AMD will get some big wins, but they won’t hit 50% market share within 3 years.
This older model is not quite the same today. Servers as an individual company IT resource is not the only driver of sales anymore. We have a few very large cloud system providers worldwide with China hosting some significant, rapidly expanding ones. I'm fairly certain that the Chinese companies will go AMD wholesale and soon. The processor advantages are too significant to ignore.

Its one thing to know history, but more importantly, is to understand why what happened did, and more importantly, what, if anything is different today. A blind following of the past is not very wise.
 

jpiniero

Diamond Member
Oct 1, 2010
6,579
317
126
Sep 22, 2007
33,195
68
126
This older model is not quite the same today. Servers as an individual company IT resource is not the only driver of sales anymore. We have a few very large cloud system providers worldwide with China hosting some significant, rapidly expanding ones. I'm fairly certain that the Chinese companies will go AMD wholesale and soon. The processor advantages are too significant to ignore.

Its one thing to know history, but more importantly, is to understand why what happened did, and more importantly, what, if anything is different today. A blind following of the past is not very wise.
I suggest you reread my post.
 
Jul 19, 2016
135
0
71
1% to 1.3% is an increase of 30% not 400%.
It didn't increase by 1% FROM 0.3% TO 1.3%.
It increased by 0.3% FROM 1% TO 1.3%.
I do wonder where people learn to read, or indeed how to do math.
Yeah I misread Gideons post as if it increased from 0.3 to 1.3% - oops
 
Feb 2, 2009
13,000
286
126
Here my thoughts,

1. 50% of Server Revenue in 3 years time is out of question. No matter how good your product is, Intel still can sell more than you and at higher ASP. But a 20-25% server market share would be enough for AMD to really make a lot of profit and erase its current loan completely.

2. Capacity would be enough for 50% market share (revenue) using TSMC's fabs.
For those that quoted AMDs capacity constrains in the past, that time AMD only used its own Fab (what is currently GloFo.). Not only that but at that time AMD only had Fab 30 in Dresden Germany, GloFo today also has an 80K Wafer capacity new Fab 1 in New York.

3. Since EPYC 2 (ZEN 2) is build at TSMC's 7nm, then Ryzen 3 (ZEN 2) could be a completely different die for the Desktop/Mobile at GloFo's 7nm.

4. 4xCore CCX at 7nm for server die is not happening. Either 6x or 8x Core CCX would be my bet for Server dies.

5. Personally for ZEN 2, I believe that Server dies will be made with 8x Core CCX and Desktop/mobile dies with 8x Core CCX (single CCX). But there could also be a low price/high volume 4x Core CCX die (APU) for both Mobile and Desktop.
 

Vattila

Senior member
Oct 22, 2004
419
185
136
4xCore CCX at 7nm for server die is not happening.
Why not?

If they go with a chiplet design as rumoured, then an 8-core CCX chiplet will be larger, yield worse and be more costly than a 4-core CCX chiplet. And it would be less reusable in the consumer space.

My bet is on 4 x 4-core CCXs with a more sophisticated topology, interconnect and coherency protocol between the CCXs to bring down average latency between any two cores in the 16-core CCX cluster.
  • A 4-core CCX chiplet would be tiny on 7nm (25-50 mm²), and hence reduce cost and increase yield and volume on the new and expensive 7nm processes.
  • A relatively small 200 mm² active interposer on the perfected 28nm process would be dirt cheap.
  • A 200 mm² die (the interposer with chiplets on top) would fit into the current packaging scheme with few changes: 4 interposers for EPYC and high-core-count Threadripper WX, 2 interposers for low-core-count Threadripper X and 1 interposer for mainstream Ryzen.
What's not to like?

https://forums.anandtech.com/threads/speculation-the-ccx-in-zen-2.2513648/page-7#post-39528340
 
Last edited:


ASK THE COMMUNITY