Review Index:

AMD Q1 2012 Earnings Analysis: Looking Back and Looking Forward

Author: Josh Walrath
Subject: Editorial
Manufacturer: AMD

Acquisitions and Charges

There were a number of charges for the first quarter that did have an overall negative impact, but in the long run can be viewed as positive. The first charges were those related to the acquisition of server company SeaMicro. This company specializes in high density, low powered units, which rely on a unique fabric interconnect technology that is very efficient at what it does. Currently these servers rely on Intel Atom based processors, but we can assume that they will transition to Brazos 2.0 and Trinity based units. This transition is a win/win for both SeaMicro and AMD. SeaMicro will have near unlimited access to these processors at a lower price, achieve better 64 bit performance over Atom, and also have a robust integrated graphics portion with OpenCL support. AMD obviously benefits here by moving more low power parts into the server market and by providing unique products to the market that simply are unmatched by other outfits. This is an aggressive acquisition for AMD, and hopefully one that will pay off for both companies. 

View Full Size
The large and unpleasant charge that we see this quarter is that of the wafer agreement change with GLOBALFOUNDRIES. AMD has now divested itself fully from GF, and that is reflected in the negative results for this quarter. AMD owns no part of GF, and that is most likely a good thing. Losses incurred by GF are no longer reflected in AMD’s quarterly results. This also has resulted in slightly more flexibility for AMD in negotiating wafer agreements with GF. Now AMD can bring up the specter of cheaper production with TSMC and UMC to potentially allow for better deals with GF. Unfortunately, these new deals will not match the one-time sweetheart deal that AMD was able to get while still being a joint owner of GF. The upside again is that GF and ATIC can absorb any potential losses and missteps that are endemic to the foundry industry, and AMD will be unaffected by this.
There has been some confusion to the second part of the major loss incurred by the $703 million charge credited towards GLOBALFOUNDRIES (made up of both cash and the rest of AMD’s stake in the company). Some have taken this as a negative and that it is essentially AMD buying themselves out of the exclusive agreement to use GF’s 28 nm process. My understanding is that it is quite the opposite. In the conference call Rory Read explicitly stated that AMD will be producing 28 nm parts through GF in 2H 2012, and those parts will be released to market in early 2013. AMD in fact bought exclusivity on GF’s 28 nm process. Currently TSMC is maxed out on 28 nm production and cannot offer a significant amount of space to any new customer. So for AMD to try to produce CPUs on TSMC’s process, they would have to redesign for a gate first manufacturing technique, and they would have to cut down their graphic’s orders to be able to get any CPU based wafers out. This is obviously not a good solution for AMD, so they are tapping GF to make these next generation 28 nm parts based on the Piledriver CPU architecture and the GCN graphics architecture.
This is a good trade off. AMD is taking a charge now, but the potential advantages of having exclusive use of GF’s 28 nm process is quite big. Unlike GF’s current 32 nm SOI (Silicon on Insulator) process, there will be plenty of customers lining up to use the bulk 28 nm HKMG line. AMD is moving away from SOI wafers for now, but that does not necessarily mean that they will not adopt fully depleted SOI in the future. The current 32 nm production does use partially depleted SOI wafers, but the jump to 22 nm might in fact include fully depleted SOI. This is still up in the air, and we will not know what AMD’s and GF’s plans for FDSOI will be until well into next year.
GF has really hit its stride with 32 nm, and this will have a direct effect on 28 nm production later this year. Getting rid of SOI wafers will lower the complexity of manufacturing, and decrease the cost per wafer. While GF has been behind in getting 28 nm out the door, they do look to be firmly on track for ramping production aggressively for the second half of this year. Once production starts up, AMD will be lowering orders on 32 nm, and GF will then be able to start swapping over that line space to 28 nm.
View Full Size
Trinity in the flesh!
Going to 28 nm production is a very good step for AMD. Obviously they are adjusting their designs to run on bulk silicon rather than SOI. 28 nm HKMG is still chock full of material advances which will give switching performance that should be on par with the current 32 nm SOI process, and transistor leakage again should be very comparable between the two products. What AMD is really gaining here is the smaller die size with much the same power and thermal characteristics of the slightly larger 32 nm SOI process. We can assume around a 15% to 20% decrease in die size as compared to the older process. Using Trinity as an example, we see that at 32 nm it is around 240 mm squared in size. If we go with the upper limit of a potential shrink of 20% we have a die that is now 192 nm squared. On a 300 mm wafer we would see approximately 253 gross dies per wafer (assuming a 2mm edge exclusion and counting incomplete dies) for the 32 nm version, but we jump to 313 gross die. This is a pretty hefty jump in both gross die and (assuming similar yields) good die. We also must factor in the lower cost of bulk silicon wafers vs. SOI. The GCN architecture also appears more efficient per mm squared than VLIW4, so the assumption is that the die space needed for the graphics portion will not grow significantly from what is currently being seen in Trinity. 
These factors will help AMD be more competitive with Intel and their 22 nm process that the current Ivy Bridge chips are being manufactured on. AMD will still be at a disadvantage, but it will not be nearly as bad as when relying on 32 nm SOI process technology exclusively. If AMD can effectively utilize half node jumps for their CPU products, they can cut off anywhere from 6 to 8 months of Intel’s overall process lead. When we consider that Intel is around 16 months ahead of where AMD is now, that is a significant win for AMD. So basically when Intel makes the leap to the next process node (in this case 22 nm) AMD will make the jump to 28 nm around 7 to 8 months later. This is a much better scenario for AMD rather than waiting another 18 months to go to 22 nm. Eventually they will hit 22 nm about 9 to 10 months before Intel goes to 14 nm. This again helps keep AMD more competitive with its larger rival.


April 25, 2012 | 08:07 PM - Posted by tbone (not verified)

good read josh ;)

id like to add the design wins for all next gen consoles will also be a big plus for AMD

looking forward to seeing Trinity in the ultrathin/laptop market, but I have to say Kaveri looks to be pretty cool on the 28nm process, should be a powerhouse!

April 26, 2012 | 10:45 AM - Posted by Josh Walrath

Those design wins are a big plus for AMD, but we have not seen any money from them yet (afaik). Also, production is still a bit in the future, but hopefully the WiiU money will start rolling in for AMD. Q1 though is typically a down time for console royalties, and I think they mentioned that in the conference call.

Vishera will give us a good idea where AMD will sit with a more IPC/performance optimized Bulldozer redesign. The first iteration of Piledriver in Trinity was more about fixing thermal issues and some latency problems in the design (from what I am hearing, and what the performance leaks are showing). Vishera is a much more aggressive redesign that is "supposed" to address more performance issues along with thermals. That particular product will likely have the same IPC characterstics of Kaveri.

There are still a lot of hurdles that AMD has to overcome to get CPU performance up there to reasonable levels against what Intel currently offers.

April 25, 2012 | 08:58 PM - Posted by Meh (not verified)

You paint an overly optimistic picture of AMD.

They have no control over their manufacturing and process anymore.
They have lost investor confidence.
Their entire marketing of Bulldozer and propaganda ahead of it and damage control after launched FAILED.
Parachute payments to execs, shareholder value eroded.

What is exactly that you see, other than regurgitating their propaganda and powerpoint slides?

April 25, 2012 | 10:04 PM - Posted by Josh Walrath

There is a lot of upside to AMD, but as I believe I stated in the article, it still has a lot of hurdles in front of it before we can consider them successful against Intel. Then again, they are the only competitor to Intel and have lasted for some 40 years now.

Intel currently is the only semiconductor manufacturer (other than the memory guys, IBM, and Samsung) that has control over their manufacturing and process. NVIDIA, Qualcomm, Ti (who has closed down most of their fabs), Apple, Motorola, and countless other semi firms that all utilize foundries. I would say that we have definitely seen the shift to a fabless model for the vast majority of companies out there. With the capital behind GF, and the proven track record of TSMC and UMC, I don't see them falling all that much further behind Intel. Intel will always have better margins due to them owning their own fabs, but that does not mean that the other companies will not be able to pull a profit from that model.

Yes, investor confidence is down, and a lot of people have panicked with all of the upper management losses. Then again, look at the last two months and see how many upgrades the stock has received (from sell to hold, and hold to buy). There is a positive trend going there.

Bulldozer is bad. I have not seen any damage control from AMD other than they stopped sending out review products. I see a few basic ads, but it seems more like AMD is turning a blind eye to Bulldozer and focusing instead on Llano, Brazos, and the upcoming Trinity.

Shareholder value eroded... you mean share price going from $2.50 a share 9 months ago up to around $8 a share today? I would say the new management team has stopped the bleeding.

I see a lot of value in AMD's graphics technology. Bobcat is a very good architecture, and much more impressive than Intel's current Atom lineup. Trinity, while not a CPU powerhouse, looks to have fixed the power issues with Bulldozer and integrates the most powerful graphics processor on a CPU. If they can fix the TDP issues of Bulldozer, and improve the IPC further with Vishera (as they are rumored to have done... Vishera is another step above what they have done with Trinity), then we will see more competitive parts in the desktop and server space. Will they surpass Intel in performance? I really doubt it. Will it have good price/performance and a competitive TDP? I think they have a much better chance of pulling that off.

There is a lot of upside with AMD, and I don't feel that I am overly optimistic. Nowhere did I claim that AMD would retake the CPU performance crown, or that they would have a whole portfolio of products that will overturn Intel, but I did state that they have the best integrated GPU in the business from the low end to the midrange. They have very competitive products in the GPU market. Unlike NVIDIA, they are actually delivering a lot of products from top to bottom.

It really does look like AMD is reinvigorated, but again time will tell if this lasts. The foundation is there, they just need to execute and focus on smaller, more meaningful steps in terms of technology and architecture. Gone are the days of massive, sweeping changes in CPU architecture. We will see a lot of refinement in the CPU, but the biggest catalyst for growth will be the GPU. Even Intel sees that, and the jumps we have seen from them in the past 5 years in graphics has been impressive. They still have a lot to learn though, and that is an area that AMD can exploit.

April 26, 2012 | 02:36 AM - Posted by Matt Smith

I can't comment on the desktop space, but AMD's competitiveness in the mobile market has basically vanished.

Bobcat's actual processor performance is on par with Atom. The graphics portion is good for playing 1080p video and some older 3D games - and that's about it. It's obviously better than Atom's IGP in regards to 3D performance, but it's not much better than Atom + Ion.

In mainstream components, AMD's best mobile Fusion APU is well behind Intel HD 4000 in performance. And processor performance is so far behind Intel it's sad.

This is a poor state of affairs. I do not want to see Intel be a monopoly. But what hope is there? AMD could double processor performance tomorrow and still be in trouble.

April 26, 2012 | 03:17 AM - Posted by GettCouped (not verified)

Considering AMD's biggest success last year was in the mobile market with Brazos, and the fact that they have Brazos 2.0 and Trinity (which has a record number of design wins and will offer cheaper ultrabooks), I don't know how you can validate any of your statements

I am concerned about your agenda.

April 26, 2012 | 12:03 PM - Posted by Matt Smith

Rather than being concerned with my agenda, maybe you should spend time reading my reviews?

April 26, 2012 | 12:15 PM - Posted by tbone (not verified)

"AMD's best mobile Fusion APU is well behind Intel HD 4000 in performance."

really link me to a review that shows this? lol

April 26, 2012 | 04:35 AM - Posted by dragosmp (not verified)

To begin with I'd just like to say to say this is a nice overview. Maybe a bit optimistic, but if I didn't want AMD to recover maybe I wouldn't have been too interested in the article anyway. You can just see, actually is quite obvious, that Josh welcomes a possible AMD recovery, and why not? They have been worse, when they launched K7 it was laughable to spend more than 120$ for an AMD processor as they simply couldn't touch PIIs' performance. At the time they only had a desktop Celeron competitor which pretty much is the case now with Llano and Bulldozer. Now they're not quite that low as they have a server CPU and a mild presence in laptops.

About Brazos vs Atom+Ion, I'm not so sure. From a performance standpoint I agree that Brazos offers little over Atom+Ion, but Atom is limited by Intel @7-10 inch laptops, while my feeling is that most Brazos CPUs sell in 300-400 13-15" laptops (have a friend that is perfectly happy with a 15" Tosh based on E-450 & Momentus XT). So while Brazos doesn't bring much technical advantage, it is positioned differently in the market and this makes it more competitive than it would otherwise be.

Llano vs HD4000 - the only preview I've seen is the one @anandtech and I'm looking forward for your view Matt on an IB laptop. In that review the HD4000 is really competitive with Llano's iGPU, not to mention IB's CPU cores run circles around the K8 era Llano. However the HD4000 is only competitive, not really superior to the HD60something in Llano. And this is before considering drivers - I am not convinced at all that Intel provides decent graphics quality and I don't mean Medium/High settings. I mean something like the Intel IGP rendering a blue surface red, or jaggies/tears and pixels running about on the screen. I would love to see a detailed review of the HD4000 as a graphics processor with FPS, but also checking weather the games are properly rendered. For one I was hoping to game on the HD3000, but the drivers were a letdown (or buggy GPU?), now I wonder if the HD4000 is a 3DMark rallycar or a real competitor to a real GPU like Llano's.

April 26, 2012 | 09:42 AM - Posted by Josh Walrath

Some extra fuel for the fire:

April 26, 2012 | 12:06 PM - Posted by Tbone (not verified)

I believe Matt is smoking crack....

Intel HD 4000 doesnt beat llano igp

And Brazod ripped apart whatever Intel atom had in market share.

April 26, 2012 | 09:50 AM - Posted by Anonymous (not verified)

cant add much except why dump the 28nm brazos?

have long thought it would be a killer

I see little evidence that more than 2 cores achieves much

brazos would be my first choice apu for 28nm - not dumping it
A huge step. 40nm >28nm - 30% better power & heat in theory

thats 13~ watts from current 18

hard to believe intel/ion would have an answer - it would be a slam dunk for amd in all day ultra mobile or light devices

a 15" thin, light brazos 28nm would sure tempt me

one theory is they feel bad after all those design wins & then not being able to supply the oems who comitted to it - give them a chance to recoup for a while.

another guess is they couldnt resist a few tweaks to brazos using the easier current node & then doing the step change. Unusual to keep u roadmap secret, but it often makes sense.

April 26, 2012 | 10:27 AM - Posted by Josh Walrath

My guess is that Rory took a look at all the current projects, looked at their manpower and budget, and cut the ones he thought made the least amount of sense. Quad core Brazos, while nice in theory, really does not have as much of a market as perhaps a dual core. When going dual core Brazos on 28 nm, I wonder if they were just looking at a die size that was almost too small for the number of pads needed? Could be some other issues there, such as the theoretical TDP was not going to be as low as they were hoping, and it made more sense to revise/optimize the 40 nm product? A good question though, and one we can hopefully get an answer to one day.

Also they are probably still constrained with TSMC's 28 nm process. Would it make more sense to order Brazos wafers and potentially lose out on higher margin products in the GPU space? I'm betting the long and short of it is that AMD felt they could not meet 28 nm Brazos demand with the supply of wafers they had to deal with, all the while cutting down on the GPU supply... which would be bad considering that NVIDIA is even more supply constrained and AMD wants to take advantage of that.

April 26, 2012 | 09:53 AM - Posted by Anonymous (not verified)

No mention of drivers or chipsets

r not amd chipsets more integrated? & if not faster, more generous with port options?


April 26, 2012 | 10:52 AM - Posted by Josh Walrath

Drivers certainly are an advantage for AMD. Intel's graphics drivers are nothing to write home about. Now, compare AMD's recent attempts at drivers vs. NVIDIA, and we have some issues. AMD still has a lot of room for improvement there.

AMD chipsets do give more SATA 6G ports and they have integrated USB 3.0 for the FM1 platform since last year. Too bad they never integrated USB 3.0 for AM3+, and it certainly seems like the 1090FX chipset will never see the light of day. So no PCI-E 3.0 for AM3+ anytime soon, or USB 3.0. AMD is sort of shafting those individuals who desire that particular platform. Maybe if Vishera comes out better than expected they will update the platform, but I think that particular project is dead and buried.

April 28, 2012 | 03:03 AM - Posted by Anonymous (not verified)

ta for the replies josh

as i suspected

I hear reasonably priced intel boards are niggardly w/ ports, which can be an expensive hassle down the road

amds integration should yield benefits - i doubt many discrete chipsets are 32nm - let alone on die - esp usb3 - it handles some serious traffic

seems to me, in the am3+ market, they need all the space they can get to make the cpu keep up w/ intel (bulldozer a bit of a dud, but I hear is making some traction in servers that like lotsa cores cheap - which is some big servers)


May 12, 2012 | 04:05 AM - Posted by Anonymous (not an intel troll) (not verified)

I'm REALLY not an Intel troll (and I most certainly don't have Steve jobs posters in my bedroom.)

AMD is a useless company. Bulldozer is epic fail. Trinity is going to suck anyway. llano is bad.
EVERYONE KNOWS ALL THIS. proof not. needed. Because the truth is indisputable.

Anyway, anyone who buys cheap junk (anything without fruit logos) does not deserve to live. huff ...

Just an objective and broad-minded observation.

Yours sincerely
Mr.Not an Intel troll

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.