| From: captainfreedom2 | 9/15/2019 5:35:24 PM | | | | | | I want to ask the financial experts here a question that's being bothering me.
I was looking through AMD's earnings call and I noticed that Lisa said that in 1Q 2019 EPYC revenue increased quarter over quarter. Yet, if you look at Mercury research numbers it shows EPYC market share dropping by 0.3%. How can this be? Either she is not telling the truth, or ASP increased hugely.
This is exactly what she said:
"Growing HPC and regional cloud service provider deployments resulted in EPYC
processor channel sales increasing sequentially." |
| | AMD, ARMH, INTC, NVDA | Stock Discussion ForumsShare | KeepReplyMark as Last ReadRead Replies (1) |
|
| To: captainfreedom2 who wrote (32618) | 9/15/2019 5:51:09 PM | | From: neolib | | | | Epyc sales did not increase QoQ in Q1. That is something we have discussed a number of times here. From the Q1 earnings press release, under segment performance:
Enterprise, Embedded and Semi-Custom segment revenue was $441 million, down 17 percent year-over-year and up 2 percent sequentially. The year-over-year revenue decrease was primarily due to lower semi-custom product revenue, partially offset by higher server sales. The quarter-over-quarter increase was primarily driven by higher semi-custom revenue.
There was a 2% QoQ increase in EESC in Q1 vs Q4, and that increase per the above is attributed to higher semi-custom revenue. Any actual increase in Epyc sales if there was some, would have been negligible.
Wrt to your quote from the CC, that quote specifically mentions channel sales, which would pertain to smaller enterprise type end markets. AMD has repeatedly stated that current server sales are driven by the large cloud providers, those are clients AMD is working directly with, they don't buy through the channel. So that comment would have been referenced to the fact that perhaps they were seeing some pickup in the much smaller channel sales of Epyc, which in the long run is good, but its not a very large section of their sales currently.
However, I should note that if the overall market grew some, than Mercury could show AMD's share dropping 0.3% even though AMD might have increased sales too. Just not as fast as the market.
FWIW, the only real datapoint we have from AMD is that 4Q18 Epyc sales were about $115M and Lisa would later state that 1Q19 was about the same. We know 2Q19 has grown a little, but I would eyeball no more than to $140M given the % shifts they have hinted at on the console side. And the TAM that they consider their server market to be is about 20M sockets/year. AMD's 4Q18 at $115M was not really 5% of that TAM, which would be about 250K/qtr. If it were, then their ASP was only $460 or so. In reality, their ASP is likely more like $780 or so, and their volumes would then be 150K so more like 3%. |
| | AMD, ARMH, INTC, NVDA | Stock Discussion ForumsShare | KeepReplyMark as Last ReadRead Replies (1) |
|
| To: neolib who wrote (32619) | 9/15/2019 7:02:39 PM | | From: captainfreedom2 | | | | Thanks for the answer neolib, it makes sense. So, I would say Lisa's statement is a bit misleading, but anyway I'll let it go.
> If it were, then their ASP was only $460 or so
Actually, that sounds about right for Naples. I wonder where you get the $780 figure. That would seem on the high side to me. |
| | AMD, ARMH, INTC, NVDA | Stock Discussion ForumsShare | KeepReplyMark as Last ReadRead Replies (1) |
|
| To: captainfreedom2 who wrote (32621) | 9/15/2019 8:02:15 PM | | From: neolib | | | | That higher figure was quoted from either an analyst or one of the industry stat outfits and posted on this board some time back. Don't recall exactly when. The surprising thing was that the source claimed AMD's ASP was a bit higher than Intel's server ASP. I think they had AMD at like 795 or so and Intel below 750, but don't recall the exact numbers.
There is a sanity check that the number is close however, in that some independent stat outfit claimed AMD's 4Q18 was more like 3-3.5% and when coupled with their $115M that gives an ASP in about the right place.
AMD really needs to get more transparent in their reporting categories:
1) Client CPU
2) Client GPU
3) Datacenter CPU
4) Datacenter GPU
5) Console
6) Other semicustom
That would be actual info. What they give instead leaves everyone guessing.
Nvidia has done a pretty good job of being more transparent by how they do their reporting categories. |
| | AMD, ARMH, INTC, NVDA | Stock Discussion ForumsShare | KeepReplyMark as Last ReadRead Replies (1) |
|
| To: THE WATSONYOUTH who wrote (32593) | 9/15/2019 10:29:50 PM | | From: ph57 | | | | Neolib,
You mentioned no pundits were writing about what low interest rates are doing to the world.
Here are some articles that discuss potential bubbles building lately.
Articles by William D. Cohan:
Note first and third articles are one year apart with similar warnings.
Article explaining warning from Michael Burry (predicted 2008):
|
| | AMD, ARMH, INTC, NVDA | Stock Discussion ForumsShare | KeepReplyMark as Last Read |
|
| To: rzborusa who wrote (32626) | 9/15/2019 10:33:27 PM | | From: J_F_Shepard | | | | | Have you suggested the same thing to Watson? I don't know the guy but for 20 years he's been trashing me ....I don't know what his problem is but I won't sit still for the trashing and the lies...... You can google me and get my record, not so with him. |
| | AMD, ARMH, INTC, NVDA | Stock Discussion ForumsShare | KeepReplyMark as Last ReadRead Replies (2) |
|
| |
