AMD Q3 results: even worse than revised expectations

[tviceman]

My post #51 was in response to this post #30. Did you read post #30 and then post #51? How is my post trolling when I provided an counter-argument with facts?

Ocre's post was more rambling and musing than anything. He didn't say anything about Northern Islands consuming too much power and he didn't say anything about Steam. He made comments about their out-of-the gate pricing (which was bad), he made comments about Nvidia waiting to what Nvidia had, and he made comments about 40nm inventory levels.

I don't really see the correlation at all with how you indirectly responded. And honestly, you have linked that market share report several times, but you always conveniently leave out that Nvidia's consumer division made more money than it's previous quarter despite losing market share. Market share is great and all until it's realized that gobbling up more market share isn't translating into making more money.

 

[RussianSensation]

AMD clearly had more GPU share and GPU profits in the past.

http://news.cnet.com/8301-13924_3-20012025-64.html

I never said AMD never had more GPU market share. But the article you linked says nothing about profits or compared profits of HD5000 series to HD7000 series for example. Again, you haven't and cannot prove that desktop HD7000 series is what's destroying the company. The reason for that is it's financially inaccurate. You can continue to pretend AMD is largely a GPU company but the reality is it's largely a CPU company. Therefore, based on the breakdown of AMD as a firm, it's the aggregate performance of all segments that's dragging the company down. Putting the blame on the profitable GPU business is simply misleading as I have stated continuously. However, it keeps being repeated over and over.

 

[notty22]

Ocre's post was more rambling and musing than anything. He didn't say anything about Northern Islands consuming too much power and he didn't say anything about Steam. He made comments about their out-of-the gate pricing (which was bad), he made comments about Nvidia waiting to what Nvidia had, and he made comments about 40nm inventory levels.

I don't really see the correlation at all with how you indirectly responded. And honestly, you have linked that market share report several times, but you always conveniently leave out that Nvidia's consumer division made more money than it's previous quarter despite losing market share. Market share is great and all until it's realized that gobbling up more market share isn't translating into making more money.

Exactly.
Jon Peddie Revises Q1 2012 GPU Sales Results; NVIDIA Share Loss Reduced May 26 2012

However, the number of shipped units does not mean a lot when we take a look at profitability. After all, AMD APU and GPUs gained share over NVIDIA and the company's GPU division still managed to lose money in vast majority of the past five quarters.

 

[SirPauly]

I never said AMD never had more GPU market share. But the article you linked says nothing about profits or compared profits of HD5000 series to HD7000 series for example. Again, you haven't and cannot prove that desktop HD7000 series is what's destroying the company. The reason for that is it's financially inaccurate. You can continue to pretend AMD is largely a GPU company but the reality is it's largely a CPU company. Therefore, based on the breakdown of AMD as a firm, it's the aggregate performance of all segments that's dragging the company down. Putting the blame on the profitable GPU business is simply misleading as I have stated continuously. However, it keeps being repeated over and over.

Offered many AMD quarters to you that offered more profits.

 

[RussianSensation]

Ocre's post was more rambling and musing than anything.

I don't really see the correlation at all with how you indirectly responded. And honestly, you have linked that market share report several times, but you always conveniently leave out that Nvidia's consumer division made more money than it's previous quarter despite losing market share. Market share is great and all until it's realized that gobbling up more market share isn't translating into making more money.

Are you joking? You must have not read his post at all. tviceman, you need to read Post #30 and Post #51 again.

I never said anything about power consumption of Northern Islands in my post. Where are you getting this from?

Here I'll quote his post by parts:

"We keep hearing how "nvidia is late" on these forums but if you havent figured it out just yet there is probably a lot more to it. Although a few people will not want to agree, there is reasons to hold off launching products sometimes especially when there are boat loads of previous generation stock in the warehouses. As a company you want to prevent having to write off a lot of stock as its just throwing away money. Amd had to write off a lot and it really hurt them. There are better ways to do things."

^ This is 100% wrong. AMD's graphics division was more profitable in Q1 2012 than in Q1 2011, which means launching first was better than holding off the launch. The implication that AMD had to write down inventory due to GPUs not selling well is also 100% wrong as inventory writedowns are related to AMD's computing unit. I addressed both of these points in my post #51.

He continued with his opinion that has nothing to do with facts:

"As far as AMD. I think they expected nvidia not to be able to compete and tried to capitalize on this. They dropped their 28nm line-up early to try to take advantage of what they thought would be a great opportunity. Having many older generation cards in the pipeline didnt matter too them because they planned to offset this with their premium. The plan all went a bust and they got stuck with a bad outcome. They waited as long as they could to drop prices because they really couldnt afford too. With older products on the shelves and no way to get that money back, they have a huge loss to take. The bullet to bite. The plan backfired. "

^ None of this is accurate and I addressed it directly. The write-down has little to do with "older HD6000 series" products as his keeps insinuating.

"AMD was first to have the full 28nm lineup out in the market but theu failed to turn it into profits. You cannot leave so many of your last gen cards out that your burning the candle at both ends. This is why their newer 28nm cards were less of a value than their 40nm cards. You could get faster 40nm AMD cards for cheaper than their 28nm cards. Having so much stock left in the pipeline really can hurt. There are better ways to do things."

^ 100% wrong again. There is no evidence to support the opinion that AMD had millions of 40nm series waiting to be sold last quarter and because of that AMD took a huge writedown for inventory. This is grasping for straws making up facts!

 

[SirPauly]

but you always conveniently leave out that Nvidia's consumer division made more money than it's previous quarter despite losing market share.

Indeed --15 percent more.

 

[RussianSensation]

but you always conveniently leave out that Nvidia's consumer division made more money than it's previous quarter despite losing market share. Market share is great and all until it's realized that gobbling up more market share isn't translating into making more money.

How much NV made last quarter is irrelevant to showing that HD7000 series did well. If both companies achieved higher margins on GPUs, then they are both profitable. How do you not understand this? How do you not understand that AMD could have negative aggregate cash flows despite the graphics division making $?

How can a profitable unit of the firm be destroying AMD when only 8% of that units cash flow is attributable to the value of the entire firm?

It's hilarious hearing these remarks from so many of you who never went to business school or worked in finance a day in your lives. I feel I am totally wasting my time since the majority of you believe you are actually knowledgeable about finance, or a company's financial position because you set on a conference call.

When terms such as "making money" are thrown around, I bet you guys don't even know what it actually means because the vast majority of you don't even understand the difference between net income, operating income and cash flows! So how can you know what "making money" actually means in finance?

What are unlevered free cash flows of AMD's Graphics unit?
What is the split between desktop discrete and mobile discrete?
What is the expected future growth rate for the free cash flows of AMD's Graphics Unit for desktop and mobile GPUs?
Do we need to apply a different discount rate to AMD's desktop discrete GPU cash flows relative to their mobile GPU cash flows?
How much is the value of a stand-alone unit of AMD Graphics?
What are the real SG&A expenses associated with AMD Graphics? Do we need to adjust for any non-recurring/extraordinary expenses?

If people can't answer these questions, they shouldn't even try to attempt to figure out how much money AMD's GPU division is making and what's it's worth as a fraction of AMD's stock price. Until then, you are just hardware enthusiasts trying to correlate AMD's woes and its stock price and pin it on desktop HD7000 series based on your opinion. The problem is the finance data is available and it shows you guys are wrong. And yet you continue to make these all-encompassing statements that HD7000 series is a disaster, and is supposedly taking AMD to the grave.

What about 80% of the company? Did you forget about it?

This is what blows me away the most.

Indeed --15 percent more.

You don't need to know at all how much $ NV's GPU division made last quarter to figure out how much of AMD's stock price the AMD graphics division is worth. The financial data simply does not support the view that's being thrown around here that the firm is losing a lot of $ mainly because of desktop discrete HD7000 series.

This about what you guys are saying. If McDonald's breakfast menu made them $ but if the rest of their product lines barely made $ despite much higher operating expenses, would you blame McDonald's breakfast for destroying the company? That's exactly what you are doing here.

 

[blastingcap]

Indeed --15 percent more.

There's a lesson somewhere in here, kids:

1) stay out of debt (unless you get a deal you can't refuse)

2) high margins in defensible markets *cough Quadro, and if NV has its way, Tesla as well* count for more than low margins in cutthroat markets *cough GeForce/Radeon* when the going gets tough

3) don't sell your foothold in fast-emerging markets like mobile for a pittance

4) don't try to go toe-to-toe with an entrenched goliath *cough Intel x86 CPUs*; you must innovate *cough Fusion, but sadly enough for AMD, Haswell looks like it will erase that advantage as well* or even leave the field

5) don't hire Rory Read, as he will fire the talent and replace them with yes-men.. the guy even LOOKS like a bean-counter. At least Dirk was an engineer.

 

[notty22]

I have to think, going so long with a temporary CEO, only let simmering problems get worse. Like not opening your bills.

 

[ShadowOfMyself]

Ive been hearing this shit since the merger back in 2006... Must be a record to stay alive for 6 years like this

What pisses me off is how AMDs terrible CPU division and marketing is dragging down the genius of ATI... I hope if AMD goes under at least ATI resurfaces somewhere else (other than NV headquarters, obviously)

 

[Siberian]

RussianSensation if graphics are only pulling in 8% of the cash flow as you say, that sounds like a huge problem.

 

[blastingcap]

RussianSensation if graphics are only pulling in 8% of the cash flow as you say, that sounds like a huge problem.

He is using flawed analysis (hint: valuation of a company's division is not based strictly on cash flow or income contribution or revenue, and Trefis is not the Bible on how to value pieces of companies), but I know if I get involved, I'll get dragged into a drawn-out discussion that I frankly find irrelevant. The topic is irrelevant because only extremists would blame AMD's problems on the ATI side of the business. ATI was doing fine before AMD. It even matched or beat NV from time to time and was a strong second player in the market, unlike AMD which beat Intel one time in 40 years and which was a weak second player for as long as I can remember, other than that one time.

By the way, don't be surprised if Intel winds up being the one to toss AMD a lifeline, if only to prop up a pretend-competitor to keep DOJ Anti-Trust from investigating Intel's de facto monopoly on desktop and server x86 CPUs. Think Microsoft giving money to Apple in the 90s. Except Rory Read is a ****, unlike Steve Jobs, who was a true leader. (I dislike Apple fanboys but have to give Jobs his due.) So don't expect AMD to pull an Apple so long as Rory Read is CEO.

 

[RussianSensation]

There's a lesson somewhere in here, kids:

Great post. :thumbsup:

I think AMD's management issues started way before Read. Can't really blame him too much to be honest because he was left with the most uncompetitive CPU, APU and server lineup in the history of AMD. These products take 3-4 years to develop. I doubt RR could have done anything to improve Bulldozer, Trinity or even Kaveri as their actual designs have been in place a long time ago. Phenom I and II were not successful CPUs for the firm. These structural and engineering problems started a long time ago but with the financial crisis and shifting consumer trends away from traditional PCs, AMD got left with no hedging strategy. At least NV has the lucrative professional Quadro/Tesla lines and Tegra. AMD has neither.

The consumer graphics business is not what it used to be even for NV. Look at their stock price as well from their glory days.

 

[blastingcap]

Great post. :thumbsup:

I think AMD's management issues started way before Read. Can't really blame him too much to be honest because he was left with the most uncompetitive CPU, APU and server lineup in the history of AMD. These products take 3-4 years to develop. I doubt RR could have done anything to improve Bulldozer, Trinity or even Kaveri as their actual designs have been in place a long time ago. Phenom I and II were not successful CPUs for the firm. These structural and engineering problems started a long time ago but with the financial crisis and shifting consumer trends away from traditional PCs, AMD got left with no hedging strategy. At least NV has the lucrative professional Quadro/Tesla lines and Tegra. AMD has neither.

The consumer graphics business is not what it used to be even for NV. Look at their stock price as well from their glory days.

I agree that products take years to develop and Read inherited a mess, but he axed a ton of people who were responsible for the ATI side's success. W.T.F.

And I agree leadership is very important. JHH may be a jerk sometimes (remember his line about opening up a can of "whoop ass" on Intel?), but he's a SMART jerk. You could tell he was trying to steer the company elsewhere since years ago, with the opening of the CUDA era. He steered away from desktop GeForce and into Tesla. Now the HPC market is booming and Quadro is as solid as ever. Tegra shows promise and is better than the absolute NOTHING that AMD has in that market segment. The one headscratcher NV has done lately is to move into cloud gaming partnerships, but I'm not sure if that was a last resort since they apparently got shut out of consoles; plus JHH may know something we don't. I also think buying Ageia was a mistake, but hindsight is 20/20 and you can't win them all.

 

[RussianSensation]

RussianSensation if graphics are only pulling in 8% of the cash flow as you say, that sounds like a huge problem.

What did you expect? 8% is for Desktop Discrete graphics. Mobile is about 11.2%. Together it's 19.2%.

Strictly from a financial sense, AMD is primarily a CPU company not a GPU company. The CPU division is what kept them reasonably afloat all these years. Whatever $ they made from desktop discrete GPUs was nice but not critical to their survival. The bread and butter of AMD have been CPUs & servers. That all started to dry up as Phenom I failed, then II, then Bulldozer, etc. Most importantly the high profit margin server market is where AMD started to lose more market share. Those profits are extremely important. Also, the budget/low-end market sectors where Intel wasn't as competitive allowed AMD to make some $ but AMD is starting to lose foothold in that space to Pentiums and Core i3s.

NV's desktop discrete GPU cash flows are estimate at about 17% of the stock price.

Here is the kicker:

AMD has $1.5 Billion of long-term debt, NV has $20 million. That debt carries an interest component that some of the cash flow has to cover.

Essentially whatever cash flow AMD makes has to be used to keep paying the debt obligation NV doesn't have. By definition, until that debt is cleared, NV graphics unit will always make more $ than AMD's assuming they have similar gross margins in the GPU division, similar SG&A expenses, the same 50/50% GPU market share split. Why? Because NV has no debt obligation, which burdens the firm's cash flow.

However, we know NV has higher gross margins to begin with since it sells the GPUs at a premium and Kepler is likely a more cost effective series for NV than HD7000 series is for AMD. This though doesn't affect the main point I am making - that AMD is failing mostly because of terrible APU, CPU and server side, while the cash flows from their GPU side have eroded to small levels a long time ago, not just with HD7000 series. HD7000 series is "failing" not much worse than HD4000-6000 series did.

It's that other 80% of the business that's killing AMD slowly. Phenom II was worse than Nehalem/Lynnfield but apparently better than Bulldozer is today.

Graphics holds steady as CPU business feels the pain - July 12, 2012
http://www.theinquirer.net/inquirer/news/2193212/amd-profits-fall-27-per-cent-due-to-lower-demand

AMD posted a $570 million loss in the first quarter of 2012 but its GPU business was profitable. You get the picture now why the firm is failing? It's not because of graphics. Do you think AMD's GPU division made more cash flows during HD4000-6000 days? It wasn't. The difference is AMD's CPU/Processor unit (which includes APUs) was a lot more profitable so it masked whatever little cash flow the GPU division made.

Trinity and Piledriver delays, as well as 5% market share in CPU server space, combined with generally weakened global PC space are exacerbating AMD's financial position despite AMD graphics perform reasonably well.

 

[blastingcap]

RS I have a lot of respect for you but I refuse to get dragged into this discussion, which I already say was irrelevant. Why are we even talking about this? Because some extremists are blaming AMD's woes on the ATI side of the business? That's just a bunch of stuff. I'm not going anywhere near that malarkey. I do agree on the large-picture stuff you painted though; going down to like 5% server share from a high of 30% (IIRC) has got to hurt, and of course you are right that the hefty debt load is a big factor. The lines between what counts as CPU and what counts as GPU are blurred thanks to APUs, though. And there has got to be some R&D overlap between APU and discrete and integrated GPUs (whatever is left of them).

I bet if ATI had never been sold, it would be doing fine even today, as standalone company. Maybe not great, but not as badly as AMD is doing right now, either.

What did you expect? 8% is for Desktop Discrete graphics. Mobile is about 11.2%. Together it's 19.2%.

Strictly from a financial sense, AMD is primarily a CPU company not a GPU company. The CPU division is what kept them reasonably afloat all these years. Whatever $ they made from desktop discrete GPUs was nice but not critical to their survival. The bread and butter of AMD have been CPUs &servers. That all started to dry up as Phenom I failed, then II, then Bulldozer, etc. Most importantly the high profit margin server market is where AMD started to lose more market share. Those profits are extremely important.

NV's desktop discrete GPU cash flows are estimate at about 17% of the stock price.

Here is the kicker:

AMD has $1.5 Billion of long-term debt, NV has $20 million. That debt carries an interest component that some of the cash flow has to cover.

Essentially whatever cash flow AMD makes has to be used to keep paying debt obligation NV doesn't have. By definition NV graphics unit will always make more $ than AMD's assuming they have the same gross margins, the same 50/50% market share split. Why? Because NV has no debt obligation, which burdens the firm's cash flow.

 

[RussianSensation]

And I agree leadership is very important. JHH may be a jerk sometimes (remember his line about opening up a can of "whoop ass" on Intel?), but he's a SMART jerk. You could tell he was trying to steer the company elsewhere since years ago, with the opening of the CUDA era. He steered away from desktop GeForce and into Tesla.

I very much admire JHH. He has a good vision for the company and is excellent at managing it overall (bumpgate notwithstanding). It took a lot of risk on his part to start with Tegra and it has evolved to Tegra 3 with finally notable design wins. Also, he listened to feedback from customers and refocused on performance/watt with Kepler which is admirable, especially with how quickly they turned the Fermi power consumption into a positive with Kepler which is essentially just a massaged Fermi architecture.

Generally speaking, AMD graphics has issues, no question because even when they sold GPUs on price/performance (HD4000-6000) series or had market leadership (5870 for 6 months, HD7970 Ghz this round), NV still continues to have maintain an overall desktop discrete GPU market share lead.

RS I have a lot of respect for you but I refuse to get dragged into this discussion, which I already say was irrelevant. Why are we even talking about this? Because some extremists are blaming AMD's woes on the ATI side of the business? That's just a bunch of stuff. I'm not going anywhere near that malarkey. .

Exactly!

Same for you wrt to mutual respect. I am sorry, didn't mean to drag you down into this discussion. My main point was that I disagree with the claims that were made in this thread (and most of this year) that AMD's woes are primarily related to ATI's side of business, specifically the desktop discrete side. NV won a lot of mobile design wins which shows poor execution of AMD's mobile discrete side, but this distinction is never made on our forum. That was basically the main point I was trying to make with all those examples and balded sections, quotes and links. If someone is going to blame AMD's 60% decline in share price, Q3 miss, and lack of profitability, it cannot be blamed on AMD's desktop HD7000 series because financially it doesn't even add up.

Since this thread addresses general reasons for AMD's Q3 revised results, I didn't agree with the view that AMD's revenue decline of 10%, its $100 million write-down, its inventory issues were mostly related to desktop HD7000 series or somehow millions of 40nm GPU parts that were laying around waiting to be sold (Post #30). That is all.

Sources tell S/A that one division of AMD that's making money, graphics, is going to have a lot of engineering outsourced.

That's pretty much the worst news here - the division that's actually the strongest at AMD and one that's actually making $ is going to be hit the most by the layoffs. That explains why Eric Demers booked it early.

 

[AnandThenMan]

The reason AMD is in such dire straights today is simple. They poured billions into their fabs in Dresden, but were never able fully to realize the fruits of that investment. Intel has bled AMD dry through every means at their disposal, legal or not. At times, mostly illegal.

Obviously not all of AMD's problems can be blamed on their #1 competitor, some critical mistakes were made at critical times. But every company makes mistakes, Intel has made some colossal ones over the years, like RAMBUS, Pentium 4, Larrabee etc. But when you hold such a large marketshare lead over your competition, you can take the hits and come out mostly unscathed. If you look at the return on R&D that AMD gets, they are actually a highly efficient company compared to say Intel. Problem is they simply don't sell the volumes to make the research sustainable.

Unless something drastically changes for AMD and soon, the company as we know today is finished. BTW, the ATI aka graphics division of AMD is their one bright spot, world class and giving Nvidia all they can handle and more.

 

[RussianSensation]

BusinessWeek reports 2,340 people or about 20% of the workforce.

http://www.businessweek.com/news/20...ces-said-to-plan-to-cut-as-many-as-2-340-jobs

"AMD is striving to trim expenses to help cope with sagging demand for personal computers that rely on its processors. The blog All Things Digital said the measure would affect workers in engineering and sales."

"In the third quarter of this year, total global PC shipments fell 8.3 percent from a year earlier to 87.5 million, market-research firm Gartner Inc. said earlier this week."

 

[Red Hawk]

Unless something drastically changes for AMD and soon, the company as we know today is finished. BTW, the ATI aka graphics division of AMD is their one bright spot, world class and giving Nvidia all they can handle and more.

While AMD is certainly giving Nvidia decent competition in the GPU market -- more than they can say in the CPU market -- I wouldn't go as far as to say it's "all Nvidia can handle". Let's be real. If Nvidia had wanted to release a bigger desktop GPU than GK104, they would have.

 

[cotak13]

The management's the issue. When I read Read is hiring consultants that to me is the warning bell.

Management hires consults when they are out of ideas. And consultants are well.. if they are so great at running companies why the F aren't those guys doing it? Because they don't want to be successful? Don't want to have a legacy like Steve Jobs?

Read's out of ideas, AMD's out of money and the consultants knows jack all about what new thing could save the company. Not sure where else this can go but down.

 

[AnandThenMan]

Let's be real. If Nvidia had wanted to release a bigger desktop GPU than GK104, they would have.

Meaning? Nvidia could have had a faster GPU right now, but they chose not to? Why would they do that?

The management's the issue. When I read Read is hiring consultants that to me is the warning bell.

Sounds like a pure desperation move.

 

[chimaxi83]

While AMD is certainly giving Nvidia decent competition in the GPU market -- more than they can say in the CPU market -- I wouldn't go as far as to say it's "all Nvidia can handle". Let's be real. If Nvidia had the ability to release a bigger desktop GPU than GK104, they would have.

Fixed.

 

[OCGuy]

"The chipmaker has also been plagued by execution setbacks that have led to a loss of market share to Intel and to graphic chip rival Nvidia"

http://news.yahoo.com/amd-plans-layoffs-recent-warning-source-001204692--sector.html

We are watching the downfall of the only real rival to Intel........sad.

 

[OCGuy]

You do realize they are talking about CPU's? This has nothing to do with when they released there SI lineup or old GPU inventory. It definitely has nothing to do with what nVidia being late to market with Kepler. Just like nothing AMD did had nothing to do with GK100 being canceled, as some have tried to postulate.

It's the Bulldozer architecture that's hurting them.



While I know there are some people who really get all amped up thinking that it's nVidia kicking AMD's butt, it's not.

Keep your head in the sand......it is comforting in there!