AMD Q3 2017 earnings

[dark zero]

Ryzen refreshes on "12nm" are quite obvious, but Vega?
I can't quite see any point in throwing even more money at a non-salvageable design.

Maybe Vega Nano might be the savior of that design... just like Fiji/Fury Nano... and they might save it along the mobile Vega in order to avoid rebadges while is preparing Navi.

And makes me wonder...what will happen if Fury nano got a node shrink to 12 nm....

 

[Yotsugi]

Maybe Vega Nano might be the savior of that design...

The software is the savior of that design.

And makes me wonder...what will happen if Fury nano got a node shrink to 12 nm....

Vega (in its current state) but clocks lower.

 

[Qwertilot]

AMD's goal with Vega 10 on 12LP is to be competitive with GV104.

That's going to be so hard though - they're kind of barely competing with GP104 as is (performance wise they're OK, not on any other metric) and GV104 is getting a full redesign and a reasonably real node shift with it. Probably actually their bigger concern should be getting a viable product into the 570/80 's markets. If they don't manage to upgrade those before/around the time that GV104/6 rolls out they'll have huge problems.

 

[Yotsugi]

That's going to be so hard though

A drived update would be enough.

GV104 is getting a full redesign

Moar ALUs - moar L1 is not a full redesign.

and a reasonably real node shift with it

About as reasonble as 12LP, being not at all.

If they don't manage to upgrade those before/around the time that GV104/6 rolls out they'll have huge problems.

Vega11 should arrive earlier than both.

 

[Cloudfire777]

imo Vega in its current form is a product with significant bugs. AMD have not even been able to get HBM2 pseudo channel to work on Vega. Vega regresses in raw and effective bandwidth from even Fury X.

https://forums.anandtech.com/thread...on-oc-uc-beyond-3d-suite-ms-dx11-sdk.2510875/
https://www.reddit.com/r/Amd/comments/6mocjw/vega_fe_has_a_serious_regression_in/

The drivers for Vega have not enabled most of the advanced new features like primitive shader.

https://forum.beyond3d.com/threads/amd-vega-hardware-reviews.60246/page-59#post-1997699

AMD has a lot of work to do to get Vega running upto its full potential. Hopefully they can get Vega's advanced features to work properly with their yearly December Crimson driver release. The next refresh of Vega 12LP should fix hardware bugs and get the memory controller working upto its maximum potential efficiency. HBM2 yields should improve allowing AMD to clock HBM2 at 2 Ghz speeds. AMD's goal with Vega 10 on 12LP is to be competitive with GV104. If they fail to do that they will get thrashed and bleed market share. If they succeed to do that then AMD will be able to stay around 30% GPU market share until Navi arrives in 2019 and AMD can try to compete across the entire range upto the highest end GPUs like the Gx100 series.

You have been here for quite some time now raghu.
As usual, the Gx104 chip will be 20-30% faster than the previous flagship, this time the Titan X.

With the Titan X being 30% faster than RX Vega 64, you really expect a node refresh will get the Vega 10 to be 50-60% faster than the RX Vega to match the GV104 card?
That sounds like wishful thinking for me.

The only way for AMD to close the gap is through new architecture AND new node. Which leads me to my next question: With Vega being so painfully late compared to Pascal, what makes you even believe Navi will be here on time?

A lot of ifs and buts here. I think the odds against AMD is just too big to count on AMD in the GPU space for years to come. Unless they allocate a ton of resources for GPUs. Which Nvidia is already spending right now and have been for years.
ELI5: like a race between a turtle and a rabbit. AMD have been using a turtle for years now while Nvidia have been racing with a rabbit. Then suddenly AMD is also switching out the turtle to a rabbit. What makes you think the rabbit will close the huge gap Nvidias rabbit have?

 

[IntelUser2000]

How could this not make the stock go higher?
I can tell you why. Investors, the same morons that told people to sell AMD stocks when it was down to $2 (almost everyone had a «sell rating» back then), heard Lisa Su yesterday say that Q4 will go down compared to Q3. That was all it took. These individuals, which have almost none experience with hardware, cant grasp the fact that historically Q4 usually been lower than Q3, so they freaked out. And like any mindless sheep, they do their everyday business by looking at sell and buy walls of stocks, meaning they follow the rest when a sell wall appears. Which short the stock value.

I don't know why you are so harsh on the average investors. Imagine if you invested $50-100k worth of AMD shares. You will be very careful(and paranoid).

Even at $12, its 5-6x worth its value just a few years ago. Current pricing already has future projections baked in. AMD by revenue is in no way worth 6x few years ago. I think its very possible that they can go down another 20% from now.

Do you think the alternative is better? AMD being way overpriced at $30 per share right now? They have been basically considered bankrupt by most folks for over a decade. Such opinions won't change overnight. The last peak was $40 before Conroe ruined the party. It wasn't until after the release of Athlon 64, and few years later when Intel was at its worst AMD reached that point.

A company as small and as volatile as AMD will have wildly varying share prices. Which is why institutional investors and those more serious in long term investing go for big guys like bank stocks, and if technology, more stable ones like Google, Intel, Microsoft.

 

[Dribble]

For share price you need to take a step back - just because you have 1 quarter where you actually make a little money with slightly less bad margins does not suddenly make you a great investment. Investors want consistent profit and growth, and the margins need to be at least competitive in the market they are in. AMD has a way to go to show that yet.

 

[StrangerGuy]

For share price you need to take a step back - just because you have 1 quarter where you actually make a little money with slightly less bad margins does not suddenly make you a great investment. Investors want consistent profit and growth, and the margins need to be at least competitive in the market they are in. AMD has a way to go to show that yet.

Yup, $71M profit is pretty awful considering they didn't had to fight CFL/Volta, peak seasonal console demand and riding the back of the god-knows-what-will happen-next mining craze. It's like some guy who did tons of overtime bragging how he finally earned more than his coworker who also got a pay cut for AWOL.

 

[rainy]

Yup, $71M profit is pretty awful considering they didn't had to fight CFL/Volta, peak seasonal console demand and riding the back of the god-knows-what-will happen-next mining craze. It's like some guy who did tons of overtime bragging how he finally earned more than his coworker who also got a pay cut for AWOL.

AMD have increased significantly their spending on R&D this year, especially in Q3:
https://overclock3d.net/news/cpu_ma...ir_r_d_budget_by_almost_22_in_the_past_year/1
https://ycharts.com/companies/AMD/r_and_d_expense

 

[frozentundra123456]

I don't know why you are so harsh on the average investors. Imagine if you invested $50-100k worth of AMD shares. You will be very careful(and paranoid).

Even at $12, its 5-6x worth its value just a few years ago. Current pricing already has future projections baked in. AMD by revenue is in no way worth 6x few years ago. I think its very possible that they can go down another 20% from now.

Do you think the alternative is better? AMD being way overpriced at $30 per share right now? They have been basically considered bankrupt by most folks for over a decade. Such opinions won't change overnight. The last peak was $40 before Conroe ruined the party. It wasn't until after the release of Athlon 64, and few years later when Intel was at its worst AMD reached that point.

A company as small and as volatile as AMD will have wildly varying share prices. Which is why institutional investors and those more serious in long term investing go for big guys like bank stocks, and if technology, more stable ones like Google, Intel, Microsoft.

Yep, no matter how great Ryzen is for the consumer, AMD needs better margins and consistent and increasing profits. Actually, it is the opposite of what Cloudfire said. Not "how could the stock not go up", but instead "why would it go up?". To me, it was an good earnings report, but what some fail to see is that the future growth has already been priced in. Basically, the stock has gone up 6x from its low with no profit. Dont forget the gpu division is in shambles, hasnt had a competitive top end product for what, about a year, and Ryzen without an igpu is ultimately a niche product. Now I do think the APUs look very promising, if AMD's preliminary results are valid, and those and the server chips could lead to sustained long term growth.

 

[dullard]

Yep, no matter how great Ryzen is for the consumer, AMD needs better margins and consistent and increasing profits.

Too many people on these forums confuse great products with great stocks. They are not the same thing. AMD's first real profit in quite some time, and it is still quite lackluster. Even using the more generous non-GAAP $0.10/share profit, that works out to 0.7% of the $14.25 share price at the time of the announcement. In comparison, Intel's $1.01/share profit at $41.35/share at the time of the announcement works out to be 2.4%.

AMD's great products have not turned into great profits. And this was an ideal quarter for them: booming video card sales due to mining, the pre-christmas boom in console processors, most of the Ryzen-based products are out, and Intel has been stumbling for years on 10 nm. Even in that great situation, AMD's earnings per share price was still about 1/4th that of Intel.

AMD really must find a way to get better margins and to keep them high, not just in a magical quarter here and there. Low cost chips are great for consumers, they are not great for share holders.

 

[Yotsugi]

And this was an ideal quarter for them

A quarter with no EPYC or Ryzen Mobile can't be ideal; period.

 

[PeterScott]

Was there any kind of Margin breakdown?

I suspect Margins are dragged down by GPUs (expensive HBM and support) while the Margins on Ryzen are quite good.

 

[dullard]

A quarter with no EPYC or Ryzen Mobile can't be ideal; period.

I wouldn't quite say no EPYC. From AMD itself:

"Enterprise, Embedded and Semi-Custom segment revenue was $824 million, approximately flat year-over-year primarily driven by lower semi-custom SoC sales, mostly offset by IP related and EPYC processor revenue."
http://phx.corporate-ir.net/Externa...WxkSUQ9LTF8VHlwZT0z&t=1&cb=636443784420408311

So, they did have some EPYC revenue. Not a full quarter's worth, but not none either.

 

[Yotsugi]

So, they did have some EPYC revenue. Not a full quarter's worth, but not none either.

Still peanuts considering there are no racks from major OEMs currently available.

 

[FIVR]

Was there any kind of Margin breakdown?

I suspect Margins are dragged down by GPUs (expensive HBM and support) while the Margins on Ryzen are quite good.

Gross margin increased 5% QoQ from 30% to 35% and increased 30% YoY from 5% TO 35%


The stock is highly manipulated but has strong support in the 10-11$ range. Above that range it is a crapshoot and you have idiots like Cramer bashing all day while he names his own dog "nvidia".

 

[raghu78]

Still peanuts considering there are no racks from major OEMs currently available.

Exactly. Major OEMs like HP,Dell and Lenovo are going to introduce EPYC servers starting in Q4 2017. But people should not think a server ramp is a step function. Its a slow ramp but I expect we will see EPYC increase AMD's revenues substantially and affect Intel's revenues materially by end of 2018.

 

[scannall]

Seems quite good to me. I was expecting break even at best for this early quarter, before the whole product stack was out. And while teething issues etc were sorted. I'd agree with others in the thread that 10-12 a share is a reasonable price at this time. Need to see how well the rest of the product stack is received, and if Epyc can make at least a small dent in the server market. I think they will, but time will tell.

 

[Yotsugi]

EPYC increase AMD's revenues substantially and affect Intel's revenues materially by end of 2018.

It would be somewhat faster I think, Scalable Family is nothing but silly overpriced duds.

 

[jpiniero]

It would be somewhat faster I think, Scalable Family is nothing but silly overpriced duds.

The server market is...different. Cloud/Telecom is like 60% of the market (and growing, unlike the typical server market), and that's a handful of companies at best. You don't know what kind of discounts they are getting, if they have access to Cascade Lake, etc.

 

[Yotsugi]

Cloud/Telecom is like 60% of the market (and growing, unlike the typical server market), and that's a handful of companies at best. You don't know what kind of discounts they are getting, if they have access to Cascade Lake, etc

AMD has already secured 3 out of 7+1.
Should be enough for 2018.
Besides Cascade Lake is a refresh with what, Apache Pass (Optane NVDIMMs) support?

 

[raghu78]

It would be somewhat faster I think, Scalable Family is nothing but silly overpriced duds.

Yes. I think AMD will get to 10% server market share by end of 2018.

 

[sandorski]

Will be interesting if sometime next year AMD cab break $2billion Revenue in 1 quarter. Seems achievable and would be quite the milestone.

 

[raghu78]

Will be interesting if sometime next year AMD cab break $2billion Revenue in 1 quarter. Seems achievable and would be quite the milestone.

I think its a good bet that AMD will break USD 2 billion in Q3 2018 (which is seasonally their best quarter). But will AMD be able to hit USD 2+ billion every quarter. I think that might take 2019 to happen.

 

[jpiniero]

AMD has already secured 3 out of 7+1.
Should be enough for 2018.

Yeah, but are they really interested or are they just interested in scaring a better deal from Intel? Can't say the answer to that yet.

Besides Cascade Lake is a refresh with what, Apache Pass (Optane NVDIMMs) support?

Higher clocks I imagine, but yeah.