AMD Pays GF to get out of wafer commitment

[Idontcare]

This sounds horrible. Give me one reason to believe that AMD will not be bankrupt this time next year.

Mubadala

Mubadala Development Co. extended its stake in AMD to about 19 percent including warrants and took an extra board seat last month, making it more likely that the investment arm of the government of Abu Dhabi would bail out the company if cash reserves dipped too low, said Cody Acree, an analyst at Williams Financial Group.

 

[mrmt]

This sounds horrible. Give me one reason to believe that AMD will not be bankrupt this time next year.

Because get out of a business isn't the same thing as bankruptcy. I expect AMD to get out of the x86 mainstream business in 2014 or 2015 because their products simply aren't competitive. They are going to stay in niches like embedded or whatever they can do using Kabini cores, but nothing more.

If you look at AMD COGS now and what they signed with GLF this week (~250 million per quarter), you'll see that AMD commitments to GLF is less than half of what it was in 2010 (some 650 million per quarter). Extrapolating from their current margins today, they are going from generating 700 million dollars per quarter today to something like 325-350 million, and two years ago this same business generated some 1 billion in sales per quarter. It is clear that they are shrinking, slowly folding their mainstream x86 business.

But at the same time, the Brazos business is pretty healthy. They are making some cash here, and that might sustain the company, albeit in a smaller shape than now. AMD is also able to tap some of its patent portfolio if they need cash badly, that would give them an extra year or two trying to pan out some new business.

Unless we see Brazos experiencing a plunge in sales similar to those Bulldozer is having now AND deeper plunges in Bulldozer sales to the point of breaching this new WSA it is unlikely to have AMD filling for bankruptcy next year.

 

[IntelUser2000]

So yes, it is understandable to go that route once you are the underdog, but why go that route when it is already clear you are going to have problems?

It's never "clear", unless its "clear as mud"(reference from some famed editor).

You've got egos going around too. They must have been confident that they could have succeeded despite everything going against them. I sometimes think that the whole company is basically made out of "enthusiast-level" guys that became engineers that can never drop that mentality. They don't get the market.

The whole shebang with JF-something, the little e-mails reviewers got when there was a 6-series chipset recall. It's all cute and cuddly and all, but if you products don't hold up to the reputation it does not matter in the long run.

Remember how many of us believed Bulldozer will bring AMD back in the game?

 

[tviceman]

Mubadala

If you had billions of dollars, would you be investing in this company thinking it's going to turn around it's operations, get back on a profitable growth track, and remain relevant OR would you investing elsewhere? What do they see that everyone else does not? Are they buying shares in anticipation/hope of a buyout by another comapny, in which they'd get a premium return on their stock? They are trading at near all-time lows, so even though the general tech public see's a company like RIMM failing (or becoming entirely irrelevant in a few short years), their stock has rebounded quite a bit from their all-time lows they hit earlier this year - paying off well with short-term investors who cash out ahead of BB10's launch (I'm assuming BB10 will be about as big as webOS, which I love and miss dearly).

 

[Idontcare]

If you had billions of dollars, would you be investing in this company thinking it's going to turn around it's operations, get back on a profitable growth track, and remain relevant OR would you investing elsewhere? What do they see that everyone else does not? Are they buying shares in anticipation/hope of a buyout by another comapny, in which they'd get a premium return on their stock? They are trading at near all-time lows, so even though the general tech public see's a company like RIMM failing (or becoming entirely irrelevant in a few short years), their stock has rebounded quite a bit from their all-time lows they hit earlier this year - paying off well with short-term investors who cash out ahead of BB10's launch (I'm assuming BB10 will be about as big as webOS, which I love and miss dearly).

Honestly I do not get that part to mubadala. Maybe they are brilliant where I am simply average or perhaps below average in investment prowess, but my strategy if I had billions would be to invest in Intel (market leader for the segment) and perhaps Nvidia.

I would not invest in the 2nd best for both sectors.

From my perspective it is a bit of a long-shot investment to add cash to AMD. That said, they are already invested to 15%, so adding 5% is more of a measure of salvaging the existing 15% more than a vote of confidence for the company.

(see: throwing good money after bad)

 

[mrmt]

From my perspective it is a bit of a long-shot investment to add cash to AMD. That said, they are already invested to 15%, so adding 5% is more of a measure of salvaging the existing 15% more than a vote of confidence for the company.

Mubadala is ATIC owner since 2011, so you not see the two as different entities now. And once you put things in perspective, GLF alone between promised and committed funds already costed over 6 billion to ATIC. To Mubadala spend 65-75 million to keep the only relevant customer of ATIC's 6 billion baby on its feet is argent de poche.

 

[beginner99]

Mubadala is ATIC owner since 2011, so you not see the two as different entities now. And once you put things in perspective, GLF alone between promised and committed funds already costed over 6 billion to ATIC. To Mubadala spend 65-75 million to keep the only relevant customer of ATIC's 6 billion baby on its feet is argent de poche.

Just shows that AMD and GF are only on paper separate entities. Financially GF needs AMD and fro AMD it stil costs a ton of money they can't afford just as before. All in all both together are a huge black hole for money. Personally I never got the point to invest in AMDs fabs in the first place. If i had billions I would have rather let it rot on a n account than invest in something that guarantees a loss.

 

[nforce4max]

The optics for Global Foundries are terrible. If you're considering them as a vendor, sure as heck doesn't give you the warm and fuzzies.

I agree, AMD will sink if they don't dump GLF and their crappy IBM SOI process then move on. 28nm TSMC would be great even for crapdozer, improved yeilds and lower power consumption plus they can use some of the design tools that are only being used for their GCN line fairly quickly. IBM's SOI is extremely difficult to work with efficiently and is why AMD for all these years have been slipping behind and with high power consumption figures.

 

[mrmt]

Just shows that AMD and GF are only on paper separate entities. Financially GF needs AMD and fro AMD it stil costs a ton of money they can't afford just as before. All in all both together are a huge black hole for money. Personally I never got the point to invest in AMDs fabs in the first place. If i had billions I would have rather let it rot on a n account than invest in something that guarantees a loss.

The real business is GLF, not AMD. It is GLF that has the biggest potential of the two. If Mubadala has to burn AMD to the ground to make GLF work they will.

 

[mrmt]

I agree, AMD will sink if they don't dump GLF and their crappy IBM SOI process then move on.

They did the latter but the former, well, have a look at the other thread. The new HSA has no exclusivity waiver, so they have to manufacture all their CPUs and APUs at GLF, plus part of their GPU portfolio. They are more entangled to GLF than ever.

 

[nforce4max]

They did the latter but the former, well, have a look at the other thread. The new HSA has no exclusivity waiver, so they have to manufacture all their CPUs and APUs at GLF, plus part of their GPU portfolio. They are more entangled to GLF than ever.

The only way out for AMD is to be bought up or find enough cash to pay for voiding the contracts, if they don't then they will in due time will be completely broke. Their IP is the only thing of value they have and for some time it appears to be deliberate by the management that AMD is in such drastic decline. As some often put it "circling the drain", if that doesn't do them in the economy will.

 

[tviceman]

Honestly I do not get that part to mubadala. Maybe they are brilliant where I am simply average or perhaps below average in investment prowess, but my strategy if I had billions would be to invest in Intel (market leader for the segment) and perhaps Nvidia.

I would not invest in the 2nd best for both sectors.

From my perspective it is a bit of a long-shot investment to add cash to AMD. That said, they are already invested to 15%, so adding 5% is more of a measure of salvaging the existing 15% more than a vote of confidence for the company.

(see: throwing good money after bad)

Sorry my post was a bit incoherent (I was sleepy). I was trying to compare AMD to RIMM, in that while RIMM has somewhat bounced back from the all-time stock lows they hit earlier this year, I believe they are ultimately doomed even if BB10 is amazing. Similarly, AMD's stock has hit at or near all-time lows recently, and even if their new focus ultimately delivers what the products they are aiming to make, I think their still ultimately doomed and will be too far removed to claw their way back. So I think it's plausible that mubadala is expecting a short-term stock bounce back and/or anticipating a buyout to earn a premium off their stock.

 

[Idontcare]

Sorry my post was a bit incoherent (I was sleepy). I was trying to compare AMD to RIMM, in that while RIMM has somewhat bounced back from the all-time stock lows they hit earlier this year, I believe they are ultimately doomed even if BB10 is amazing. Similarly, AMD's stock has hit at or near all-time lows recently, and even if their new focus ultimately delivers what the products they are aiming to make, I think their still ultimately doomed and will be too far removed to claw their way back. So I think it's plausible that mubadala is expecting a short-term stock bounce back and/or anticipating a buyout to earn a premium off their stock.

I think you are making a keypoint here - mubadala has a large enough stake in AMD, as well as seats on the board, such that they would be far more privy to internal activity of an actionable nature as far as the investment value of AMD is concerned (sort of like legitimate insider trading, like employees moving their 401k weightings around between company stock vs bond/cash when they see fab loadings going up or down)...so we probably should look at mubadala's actions as a good direction to trade the stock. Buy when they buy, sell when they sell.

 

[Av9114]

Has this ever happened between TSMC and one of it's customers?

No. But "take or pay" is common in the industry. What is uncommon is the forced dependence that AMD signed itself into with the exclusivity contract to GloFo. No one does that, well no one but AMD that is.

Can someone help me catch up here? So the actual report is that there was a "take or pay" clause in the contract between AMD and GF and AMD is choosing to pay. Is there also a contract that requires/required that AMD not use any other foundry as well? I guess what I don't understand is if "take or pay" is common in the industry what is unusual about this case other than the fact that AMD has chosen to pay? Is it the size of the penalty relative to the order?

 

[gammaray]

your graphic does not make justice to how bad this company is, go back 10 years when the price was at 40$....

from 40$ to 2$...great job AMD

 

[mrmt]

. Is there also a contract that requires/required that AMD not use any other foundry as well?

The WSA between GLF and AMD had an exclusivity clause. AMD could not use any other foundry to manufacture CPUs or APUs.

Until last week AMD had a waiver for this clause, so it could maufacture Brazos on TSMC. Now they can't.

I guess what I don't understand is if "take or pay" is common in the industry what is unusual about this case other than the fact that AMD has chosen to pay?

When you enter in a take-or-pay contract you usually establish the "take" value reasonably below your consumption/sales forecasts, so the fact you had to pay means something went wrong with your sales/consumption predictions.

What was impressive on AMD, on average it was entitled to take some 350/400 million dollars in wafers per quarter, and this quarter they are taking only 115 million, so about a third of what they were supposed to take. Their sales will crater.

 

[ShintaiDK]

http://www.xbitlabs.com/news/cpu/di...e_Not_Interested_in_Low_Volume_Customers.html

Mr. Read wants AMD to only use two process technologies. Moreover, he wants to reduce the number of metal layers that contain logic elements within microprocessors to “industry standard” levels, which naturally means compromises in performance.

If anyone should still be in doubt about high performance parts.

 

[Idontcare]

The WSA between GLF and AMD had an exclusivity clause. AMD could not use any other foundry to manufacture CPUs or APUs.

Until last week AMD had a waiver for this clause, so it could maufacture Brazos on TSMC. Now they can't.



When you enter in a take-or-pay contract you usually establish the "take" value reasonably below your consumption/sales forecasts, so the fact you had to pay means something went wrong with your sales/consumption predictions.

What was impressive on AMD, on average it was entitled to take some 350/400 million dollars in wafers per quarter, and this quarter they are taking only 115 million, so about a third of what they were supposed to take. Their sales will crater.

This.

(1) the exclusivity agreement is toxic, it would be toxic for any fabless company to tie themselves to the success/failure of just one specific foundry...it defeats the value of being fabless and having the flexibility to shop for the best deals from competing foundries in terms of wafer prices and process node technology capabilities (gate first was a bad idea and everyone knew it, but AMD got stuck with it because IBM wanted it).

GloFo has a guaranteed captured audience in AMD with the exclusivity clause, and with a take-or-pay contract they are guaranteed to get paid too. Kinda defeats the competitive spirit that is supposed to rally GloFo employees to best TSMC at anything, let alone best them at everything as they ought to be attempting.

(2) the $500m Q4 take-or-pay contract was known WELL in advance, it has been a "known cost" to AMD for more than a year now and yet they waited until mid-quarter to drop this equity-robbing bombshell onto AMD's shareholders. That ain't right. It goes to show that absolutely no one at AMD, from the executive office to the BoD, has AMD's public shareholder's best interests in mind. They are negotiating behind closed doors in ways that are literally resulting in AMD equity being drained away and put into GloFo.

What I find to be startling for AMD is that they knew this $500m Q4 take-or-pay was coming and yet they didn't take any measure years ago to get their GPU production shifted (or shared) on GloFo's 28nm in advance of this. Had they done that then they could have used more of that wasted money on building GPU parts that they could sell for revenue.

Instead they just walked into this situation and then all of a sudden are like "whoa! where did that come from!? if only we had some heads-up, we might have anticipated and moved designs around a year or two ago..."

They might be taking $115m of wafers but the wafers didn't cost them $115m, the wafers they are buying cost them $360m. For a fabless business, they have locked in the worst price/wafer ever.

 

[Idontcare]

http://www.xbitlabs.com/news/cpu/di...e_Not_Interested_in_Low_Volume_Customers.html



If anyone should still be in doubt about high performance parts.

I had no idea they managed to crater their GPU marketshare so deftly D:

There is something toxic about AMD management, it doesn't seem to matter how golden of a thing they are given (or purchase). As a company they seem to manage to find the one perfect way to mismanage the asset and leave all the profits and marketshare to their competitors.

Can this company do anything right?

 

[ShintaiDK]

This aint helping either:

http://www.xbitlabs.com/news/other/...ate_Half_of_Semiconductor_Market_Players.html

“I do think there are a number of transitions coming up in this decade. We’ve got transition to 450mm [wafers] at some point; we’ve got a transition to EUV at some point. Both are going to be expensive and are going to require scale. In general, every time there has been a wafer-size change in the industry, only half of [design and production] market players [survived],” said Paul Otellini, CEO of Intel, at Sanford Bernstein technology conference.

 

[laputan machine]

I had no idea they managed to crater their GPU marketshare so deftly D:

There is something toxic about AMD management, it doesn't seem to matter how golden of a thing they are given (or purchase). As a company they seem to manage to find the one perfect way to mismanage the asset and leave all the profits and marketshare to their competitors.

Can this company do anything right?

Yikes, the decline has been that big? I don't remember any big scandals as far as gpus goes and they have been about par with nvidia as far as performance goes this time, seems strange they'd lose so much marketshare. Could it be that OEM's are getting more hesistant to use amd gpus simply because of the failings of the other parts of the company?

 

[Genx87]

I had no idea they managed to crater their GPU marketshare so deftly D:

There is something toxic about AMD management, it doesn't seem to matter how golden of a thing they are given (or purchase). As a company they seem to manage to find the one perfect way to mismanage the asset and leave all the profits and marketshare to their competitors.

Can this company do anything right?

As Alec Baldwin said in Glengarry Glen Ross "They should fire their effing ass. Because a loser is a loser".

 

[Joepublic2]

your graphic does not make justice to how bad this company is, go back 10 years when the price was at 40$....

from 40$ to 2$...great job AMD

They must have extended their moar coars strategy into their accounting dept. as well. They should outsource it:

AMD Financial Services - making cents of your dollars since 2002.

 

[exar333]

Yikes, the decline has been that big? I don't remember any big scandals as far as gpus goes and they have been about par with nvidia as far as performance goes this time, seems strange they'd lose so much marketshare. Could it be that OEM's are getting more hesistant to use amd gpus simply because of the failings of the other parts of the company?

I think if you split mobile and Desktop apart, it would look 'pretty good' for AMD on desktop and just horrible on mobile. Optimus has just KILLED AMD over the last 18 months or so. AMD mobile discrete solutions are just few and far between compared to NV.

 

[Genx87]

Probably not much different on the desktop parts. I think we have been back to the usual 60/40 split in that segment for months.