AMD Q414 results

[AnandThenMan]

AMD has value but it isnt in x86.

Consoles? Also their APUs are actually extremely good they just need a node shrink and some tweaks. In fact the best thing AMD could do is forget about desktop completely and put all their resources into a lean and node shrunk mobile APU for notebooks and tablets.

The graphics side of their business is good except for some extremely stupid marketing practices like releasing stock cooling that sucks and reference cards that get used in benches for years by unscrupulous sites. AMD is also way way too nice about sharing their tech, being the nice guy leaves too much money and competitive advantage on the table.

 

[krumme]

If Qualcom buys AMD. I dont see x86 chips being sold by Qualcom. They will use GPU IP to build their mobile business.

At this point I wonder if it makes sense to spinoff the GPU division into a seperate company. Focus on gaming, professional, HPC, and mobile designs using ARM as the CPU. Let the x86 business wind down and call it a day. AMD has value but it isnt in x86.

What is GF worth if amd is sold?
What is mubadala interest in "selling" amd?

- and btw it also shows how idiotic this prediction of amd going bankruptcy is. Its like saying the need for oil disappear. Actually amd have never been more secure than now - but also stuck in a bad place.

- buying amd shares is the same as giving money to mubadala.

 

[Genx87]

In order to spin off the GPU division AMD would have to either raise equity on the stock market (unlikely), split its cash reserves among the old AMD and the new GPU company (impossible) or find someone to outright buy the new company (hard to do). That, and they would have to somehow retain access to the new GPU company IP but it would also have to pay for this IP, which would further erode AMD's prospects of survival. The spin off is very unlikely for the given reasons.

There's value in AMD's CPU business but only as a service provider for the semi-custom business, as a consumer business the value is negative (the division is bleeding money, money that could be used to prop up other profitable businesses). Will AMD double down its bet that K12 and Zen will be viable competitors in the consumer/server market, and spend R&D money on them or will AMD kill/trim scope of one or the two projects and refocus the company faster in the semi-custom business? That's the question Lisa still has to answer.

I think a scope trimming/project killing was the way chosen but the changes weren't made public yet. The evidences for that are:

- Lisa's refusal to provide guidance in terms of the product roadmap (she said that these questions will be answered in the Financial Analyst Day.

- R&D plunged 15% in Q414 from Q314 and should fall some 5% more in Q115. Unless Lisa found a way to make AMD R&D teams work 20% better, something has to be trimmed/killed in the R&D pipeline. Delaying a project is out of question, given AMD's deteriorated market position.

Then I would guess an ordered bankruptcy would be their best bet to restructure the business. Being a 2nd rate provider in multiple markets is not doing well for them.

 

[mrmt]

Consoles? Also their APUs are actually extremely good they just need a node shrink and some tweaks. In fact the best thing AMD could do is forget about desktop completely and put all their resources into a lean and node shrunk mobile APU for notebooks and tablets.

The Trails are very cost optimized, I doubt a node shrink would be enough to keep AMD in the race, and the mobile market is heading towards integration and commoditization, AMD does not have nor the IP for integration neither the volumes to support commoditization.

AMD strategy of being second source for Intel only worked while the PC market was booming, because Intel would divert its focus for the high growth areas. Now that there are no high growth areas, Intel is getting all the market it can, leaving only scraps for AMD.

There are no easy choices for AMD CPU business.

Then I would guess an ordered bankruptcy would be their best bet to restructure the business. Being a 2nd rate provider in multiple markets is not doing well for them.

Bankruptcy would get rid of debt and the WSA, but then, what? The viability of AMD CPU business would still be in question, they still wouldn't be able to make money with their GPU IP and semi-custom wouldn't get more traction than now. The kind of restructuring they need is more about hard work, good ideas and deep cuts than about bankruptcy and debt forgiveness.

 

[III-V]

Bankruptcy would get rid of debt and the WSA, but then, what? The viability of AMD CPU business would still be in question, they still wouldn't be able to make money with their GPU IP and semi-custom wouldn't get more traction than now. The kind of restructuring they need is more about hard work, good ideas and deep cuts than about bankruptcy and debt forgiveness.

Getting rid of the WSA is enough incentive, in and of itself. It's in AMD's best interest to not be joined at the hip with GF.

Heck, if they were bold enough, they could even move over to Intel's foundry. That'd be delicious.

 

[pablo87]

I went back and did more research on what "ATI" might be worth today and whether AMD overpaid. I did so by reviewing NVIDIA's performance.

What I found grosso modo is: sales have doubled; they've generated well over $6B of cash flow which they've used to pay dividends and buy back shares to the tune of over $4B; they've strengthened their cash balances by $2B; and their market cap today is $10B.

In my opinion, its not unreasonable to think had AMD left ATI alone, with almost no growth sales today would be in the $600-700MM range quarterly, they would have generated substantial cash flow over the years, and they'd still be worth at least $5B.

The narrative that AMD overpaid for ATI simply doesn't pass the smell test. At all.

Did they mismanage it? Absolutely, as only they know how.

Did that acquisition allow AMD to survive a few more years? Absolutely. Without GPU, AMD would have ceased to exist a few years ago because no APU and no console designs. Edit: they probably knew this when they bought it.

The scary part is if the old AMD is jettisoned, ATI can resume where it left off, growing the business, making money, generating cash flow, with the market cap of ?B. Two things stand in the way: the WSA and LTD holders. For LTD holders, if somebody makes the case to them, they won't stand in the way, why would they? Its a better business model. Global Foundries? Not so much.

 

[cbn]

HBM letting the APUs starting to really eat into huge chunks of the discrete market might end up helping them get a chunk of the market there.

Why does AMD want to eat into huge chunks of the discrete market when they already own a discrete GPU business?

Intel doesn't own a discrete graphics business so having a large iGPU makes sense for certain niches in mobile (eg, gaming laptops). However, I am still very doubtful about the benefit on desktop other than extreme niche scenarios (ie, the desktop with high graphic capability needs to be very small form factor for some reason-even if it means losing the ability to swap in a video card).

Actually I think both Intel and AMD would be better off reducing iGPU on desktop in most cases. Make the processor cheaper to make and the technology more accessible, the exception being processors for gaming laptops etc.

 

[AnandThenMan]

The narrative that AMD overpaid for ATI simply doesn't pass the smell test. At all.

It's used to fit a narrative, nothing more.

 

[Dribble]

Why does AMD want to eat into huge chunks of the discrete market when they already own a discrete GPU business?

This has been mentioned before but the fanboys just care about AMD on the box, and don't consider how much money AMD actually makes on each sale. Before AMD went with high performance on-chip graphics they effectively sold a cpu for X, and sometimes a second sale of some graphics for Y. Now they just sell a cpu for X with the graphics built in (i.e. adding the graphics didn't increase the cost over the old cpu's). They can't charge extra because Intel doesn't charge extra for embedded graphics. Hence while AMD had no choice but to move to APU's to try and stay competitive with Intel it actually lost them money.

 

[cbn]

The GPU division is far cheaper to fund and has the most promising outlook. It's far easier for AMD to fight Nvidia: They don't have a node handicap, they aren't tied to GLF there, their graphics IP might add value to the embedded business and the R&D gap between them and Nvidia is significant but not unsurmountable.

I think the real trouble is with their CPU division: They will fight either Intel or the other ARM players, the value of the CPU division is highly questionable for their embedded business and they do have a node handicap compared to the top players and the R&D GAP is really huge. And depending on the new scope of their embedded business, a small cheap x86 core and off-the-shelf ARM cores might do the trick.

So I wouldn't count on the GPU division demise now. Even if they fall behind NVidia it is far cheaper to chase them than to chase the top dogs in the CPU division. I think the real threat today is to the CPU division. A close to 20% drop in the consumer division is surely enough to harm margins, put them in the red and wipe out whatever gains they can make with their embedded business, and given how this is the R&D budget already, further cuts are likely to delay or force scope reductions (like Excavator).

I don't think AMD CPUs are so bad, but being tied (in the form of an APU) to a overly large cost adding memory bottlenecked iGPU that causes CPU throttling under load isn't helping IMO.

The Athlon x4 860Ks are much better, but I doubtful how many AMD can make at X retail price considering each one comes from a 245mm2 28nm die.

 

[cbn]

This has been mentioned before but the fanboys just care about AMD on the box, and don't consider how much money AMD actually makes on each sale. Before AMD went with high performance on-chip graphics they effectively sold a cpu for X, and sometimes a second sale of some graphics for Y. Now they just sell a cpu for X with the graphics built in (i.e. adding the graphics didn't increase the cost over the old cpu's). They can't charge extra because Intel doesn't charge extra for embedded graphics. Hence while AMD had no choice but to move to APU's to try and stay competitive with Intel it actually lost them money.

Some iGPU is fine, but I am pretty much of the opinion Intel GT2 graphics for desktop are pointless. (re: GT2 is overkill for non-gaming use, but no sufficient for gaming).

Being in that middle ground of iGPU (like GT2) just adds cost without adding any meaningful benefit.

What AMD could do to better compete with Intel is not make that same mistake, and instead scale back iGPU (except for processors targeting gaming notebooks, etc.).

 

[mrmt]

I went back and did more research on what "ATI" might be worth today and whether AMD overpaid. I did so by reviewing NVIDIA's performance.

What I found grosso modo is: sales have doubled; they've generated well over $6B of cash flow which they've used to pay dividends and buy back shares to the tune of over $4B; they've strengthened their cash balances by $2B; and their market cap today is $10B.

In my opinion, its not unreasonable to think had AMD left ATI alone, with almost no growth sales today would be in the $600-700MM range quarterly, they would have generated substantial cash flow over the years, and they'd still be worth at least $5B.

The narrative that AMD overpaid for ATI simply doesn't pass the smell test. At all.

Did they mismanage it? Absolutely, as only they know how.

Did that acquisition allow AMD to survive a few more years? Absolutely. Without GPU, AMD would have ceased to exist a few years ago because no APU and no console designs. Edit: they probably knew this when they bought it.

The scary part is if the old AMD is jettisoned, ATI can resume where it left off, growing the business, making money, generating cash flow, with the market cap of ?B. Two things stand in the way: the WSA and LTD holders. For LTD holders, if somebody makes the case to them, they won't stand in the way, why would they? Its a better business model. Global Foundries? Not so much.

Pablo,

ATI would be worth 5 billion today, after growintg, not in 2006, when AMD bought them. ATI didn't generate 5 billion for AMD since 2006 after wacc, so yes, AMD overpaid for ATI.

 

[AnandThenMan]

so yes, AMD overpaid for ATI.

Saying it over and over again doesn't make it so.

 

[mrmt]

Saying it over and over again doesn't make it so.

Read the 5 billion discounted at wacc part and it will make it so.

 

[AnandThenMan]

I went and looked at your post history to try and get a better sense of where you are coming from. Do you realize that 99% of your posts are in AMD related threads? Do you own shares in AMD? I genuinely curious why the keen interest in the company.

 

[Erenhardt]

I hope amd goes belly up so all those who don't appreciate the perf/$ amd offers can get me my future upgrade parts. They don't care about money, so they can get me free upgrades. 0 perf/$ for them, but they don't care about perf/$

 

[frozentundra123456]

Saying it over and over again doesn't make it so.

And saying they didnt over and over doesnt make that true either. The true cost of ATI is the opportunity cost of what AMD could have done with 5B dollars.

If that would have let them keep their fabs and not entered into the disastrous WSA with GF, it could have let them stay competitive in cpus. But really nobody knows what would have happened if they had not purchased ATI. In a way it doesnt really matter. Personally, I think it is obvious they overpaid, especially relative to the assets of AMD. It also seems obvious from the current state of the company that the superior igpu of AMDs apus relative to intel is not a compelling purchase feature.

 

[rainy]

Do you realize that 99% of your posts are in AMD related threads? Do you own shares in AMD? I genuinely curious why the keen interest in the company.

Don't expect to see the answer to those questions.
Btw, he has got a permanent ban (as I presume) on SemiAccurate forum for that what he's doing here - trolling.
http://semiaccurate.com/forums/member.php?u=1281

 

[AnandThenMan]

...let them stay competitive in cpus.

Wrong. How many CPUs today have no GPU tech? And don't say AMD could just license the tech, from who? And even if they did you still need a team to do it, and you need software engineers. This doesn't happen just because it takes a dedicated and concerted effort.

But that won't stop you and others from saying AMD paid too much, or should have never gone after ATI. So continue.

 

[AnandThenMan]

Don't expect to see the answer to those questions.

I expect answers. You don't have a post count that is 99% in AMD threads (this is not an exaggeration take a look) unless you have a financial interest in the company, or you are a troll. If it's the latter then it needs to be taken care of.

 

[pablo87]

Pablo,

ATI would be worth 5 billion today, after growintg, not in 2006, when AMD bought them. ATI didn't generate 5 billion for AMD since 2006 after wacc, so yes, AMD overpaid for ATI.

NVIDIA I believe was also worth the same or similar in 2006/2007 that it is now. Point being, they invested $5B, what would they have gotten out of it had they managed it properly? A lot.

Cash flow, intellectual property without which AMD was doomed, and they could still get their capital back today, had they managed it properly.

As far as spin off is concerned, its been done many times before so where there's a will there's a way. They would value what is being spun off, agreements would be reached, money would be raised. However, this probably can't be done under the current structure - they need to prepare the groundwork internally and deal with Global Foundries' WSA. The longer they wait though, the more devalued the ATI asset becomes. Customers would certainly be better off. AMD is no longer competitive with Intel in performance x86, and the revenues in Consumer APU's can't justify their operations. On the other hand, competition in GPU's is starting to suffer because of AMD's issues.

This WSA has been such an albatross. Its not the only issue but its certainly the elephant in the room and its like a cancer, spreading everywhere, affecting decisions. When I first invested in AMD, Bobcat benchmarks were being leaked, Bulldozer was being hyped and ATI was on a nice cadence. I thought this is going to be a sweet deal. Then I noticed the elusive WSA and that raised a red flag. I managed to get out without losing and have kept updated in case the situation improves. Even if I have a lot of faith in Lisa Su but if she doesn't address the structural problems, no amount of good mgmt can fix things. In the words of Warren Buffett, when good Mgmt meets a bad business model, its the reputation of Management that suffers.

 

[frozentundra123456]

and they are competitive in cpus now on 28 nm and joined at the hip to global foundaries?

 

[AnandThenMan]

and they are competitive in cpus now on 28 nm and joined at the hip to global foundaries?

Compared to have NO GPU tech in processors yes. Without it they would be no competition AT ALL. You seem to have a lot of trouble separating IF AMD should have purchased ATI, and how they handled things going forward.

 

[III-V]

Wrong. How many CPUs today have no GPU tech?

Ever heard of licensing?

 

[AnandThenMan]

Ever heard of licensing?

Already covered that.