AMD Pays GF to get out of wafer commitment

[-Slacker-]

Did AMD really have to sell it's fab in 2009? I mean REAAALY had to?

 

[Rvenger]

Their drivers are flawed both in crossfire and single gpu config, its evident gaming is smoother on nvidia cards and u can find the research done on it in v&c.

I have owned more Nvidia and AMD cards than the months in the year. You are entitled to your opinion but this is not true especially if you are a single card user.

 

[Idontcare]

PS - on ATI, its not so much that AMD overpaid but that they couldn't afford it the way they structured the deal.

The proof that AMD overpaid is in the fact that they have had to write-down sizable portions of the supposed value of ATI's assets post-acquisition.

Did AMD really have to sell it's fab in 2009? I mean REAAALY had to?

No. It was about unlocking shareholder value.

Did AMD really have to spinoff their NOR flash business to form Spansion? Nope, but shareholder value was unlocked.

Did ATI have to sell out to AMD? No, but they sure unlocked a wealth of shareholder value in doing so, ATI shareholders made out like bandits on that deal.

 

[mrmt]

I view it as being a means of creative shenanigans in the sense that Dirk (who was Sr.VP at the time, planned to take over for Hector because the deal was engineered to have Hector move over to GloFo) and Hector swapped an existing liability that was on the books (debt) in exchange for a long-term liability that was neither on the books nor divulged to the shareholders (a take-or-pay contract extending some 15 yrs).

I think we are saying the same thing but in different ways.

A potential liability isn't supposed to be in the balance sheet unless there is a material chance of it becoming a real liability. The liability then should be properly provisioned.

The WSA at the time of its creation was exactly what I described above, a potential liability, so it shouldn't be in the books. Had the conditions premises in the WSA stayed true AMD would never had the kind of problems they are having now, it would be business as usual. But who would predict the kind of disaster that happened with AMD big core lines in 2009?

But AMD is guilty here, but not by accounting shenanigans, but of lack of transparency. They should have divulged the WSA to investors and let them decide how they would value that, instead what we got is a kinder egg agreement where every time you look inside the deal you have a new surprise. AMD indeed misled its investors, but not through an accounting trick, but by *not disclosing known risk to its shareholders*. But what to say from a company that cannot even be forthcoming with its sales forecasts? Transparency and ethics with its shareholders isn't really AMD strong points.

Now these were the same masterminds who engineered forking over billions of shareholder equity to ATI shareholders, so it is not really of any surprise that the GloFo deal was rigged with special little equity-destroying nuggets. It does appear to have been their forte.

Yeah, but those are two very different situations. In the former they *chose* to enter in the deal, in the latter they *had* to do the deal.

Put yourself in ATIC's place. What would you do if you were to absorb the assets and debt of an almost bankrupt company? In the first "no" you got in the negotiation table you would get up and let AMD limp back a few days later agreeing with everything but the most absurd demands. You would leave AMD with just the bare minimum for them to survive, and nothing more. It isn't really negotiation.

 

[Torn Mind]

Look at how they're selling. They're not priced well. The demand is extremely low. If they were priced well, demand would be high.

Priced well based on a their prices vs the Intel offering they're competing against, not from a margins made on sales standpoint or volume standpoint. They tried to charge more than Intel's rival offerings when the war was Bulldozer vs Sandy Bridge because they hyped up Bulldozer the next great thing they made. Now, they have to cut their envisioned prices and thus sacrifice their margins because people won't buy an AMD processor anymore except when they are at or below the rival Intel processor price.

 

[pablo87]

The proof that AMD overpaid is in the fact that they have had to write-down sizable portions of the supposed value of ATI's assets post-acquisition.

I don't buy that narrative.

ATI was trading for $16 in Q2 2006, AMD paid ~$20 so not an unreasonable premium and thus no overpayment.

As a comparison, NVDA was also worth around $4B at the time and it is now worth $7.5B. What has changed is NV has made inroads in tablets with Tegra3, while ATI has numerous design wins for gaming consoles, they've also been distracted by having to focus on APUs. AMD even sold off Imageon for peanuts - to give an idea, Imagination's market cap today is $1.7B.

Had ATI remained a separate entity (or if AMD had simply invested in them and left the company alone), it would be worth IMO anywhere between $4B to $8B today, taking into account the market valuation in 2006, the fact the business would be run normally and not subject to annual layoffs, lack of mgmt focus and bad decisions, adding 6 years of accumulated profits ($500MM to $1B right there), numerous console design wins, possible royalties on AMD APUs, and last but not least, the smartphone GPU business (another $1B wag right there). And who knows, maybe ATI is also in the SOC business.

On the balance of probabilities, a reasonable investment.

***********************************

Now is a more plausible narrative that AMD didn't have the financial resources to buy ATI (they borrowed $2.5B to do the deal remember?) and then drove the ATI business into the ground? Absolutely. Just as an example, they sold off Imageon for $65MM when it should have been developed and would probably be worth a heck of a lot more today.

Also, writing down the ATI acquisition is just AMD mgmt finding a scapegoat to explain away AMD's usual poor performance and treat it as a "one time" charge so as not affect the sp. SOP at AMD, IMO (that we can agree with).

 

[Idontcare]

nvm

 

[pablo87]

Well, must be true since you say so

But back on earth, where I live, just because you pay someone their asking price doesn't mean you aren't over-paying for what you are buying.

Do you ever shop? I don't buy a loaf of bread for $5 just because it is on sale for $5, unless that loaf of bread is really worth $5.

And if it is worth $5 you won't find me taking write-downs on impaired assets a year later.

Here on earth we have a whole host of homebuyers who overpaid for their homes, currently living with under-water mortgages.

But I suppose that narrative is just a crock, to you, as well. How could those homebuyers have overpaid for their homes? They paid the asking price, and their neighbors homes were equally priced as well...

I apologize for the lack of respect. it was uncalled for.

 

[podspi]

I think it is pretty hard to argue that ATI wasn't overpriced. If AMD had used that same $5b in R&D (fab and CPU dept), where do you think they would be today?

They still could have pursued HSA, since the whole point of HSA is to integrate different IP and have it all play together nicely.

 

[mrmt]

I don't buy that narrative.

ATI was trading for $16 in Q2 2006, AMD paid ~$20 so not an unreasonable premium and thus no overpayment.

As a comparison, NVDA was also worth around $4B at the time and it is now worth $7.5B. W

You should check your assumptions here. NVDA stock was quoted around 16 at the time, which gives a market cap around 10 billion. It would have been too asinine, even for AMD management, to not grab NVDA for 4 billions and instead go after 5 billion ATI.

 

[pablo87]

You should check your assumptions here. NVDA stock was quoted around 16 at the time, which gives a market cap around 10 billion. It would have been too asinine, even for AMD management, to not grab NVDA for 4 billions and instead go after 5 billion ATI.

My bad but so are you. 385MM shares fully diluted (nv 10Q) at say $18 (yahoofinance price history) = $7B (current valuation $8B fully diluted)

So AMD did get a discount on ATI and under independent management, would probably still be worth close to what they paid for it, just like NVDA is today. Though the opportunities for ATI were probably greater (SOC but also Imageon, console and APU).

opportunity cost? yes

overextended themselves? yes

mis-managed ATI? yes

overpaid? Nope

 

[lyssword]

And that's why girlfriends are more expensive than...
Oh, wait

 

[ShintaiDK]

AMD overpayed on a factor of 5 to 10 for ATI depending how you view it. They bought a bubble.

 

[BD231]

XFire I have no idea but their drivers have worked perfectly well in single card configs for me and have done for all the AMD cards I have used. The only driver problems I have ever had were with nvidia ones ironically, considering how everyone raves about nV drivers...

If you read the article you'd know your reply has nothing to do with what im talking about.

 

[mrmt]

My bad but so are you. 385MM shares fully diluted (nv 10Q) at say $18 (yahoofinance price history) = $7B (current valuation $8B fully diluted)

Thank you for the clarification.

So AMD did get a discount on ATI and under independent management, would probably still be worth close to what they paid for it, just like NVDA is today. Though the opportunities for ATI were probably greater (SOC but also Imageon, console and APU).

opportunity cost? yes

overextended themselves? yes

mis-managed ATI? yes

overpaid? Nope

How do you say that they didn't overpaid?

From an asset POV they overpaid, that's why they had to incur in impairment charges in the next quarter.

From a valuation POV they also overpaid, as ATI brought back almost nothing of the cash they paid for the acquisition.

No assets worth 6 billion and no cash flows worth 6 billion, then how are you defending that they didn't overpaid? Because of Nvidia valuation? The companies weren't exactly equal. Nvidia generated a lot of cash because of their bigger share in the consumer GPU market, and they had Quadro, a product line that sustained the companies in the bad times from 2008 through 2010, and btw didn't have a correspondent in ATI product line.

ATI was bubbled, Nvidia was not, simple as that.

 

[Idontcare]

I apologize for the lack of respect. it was uncalled for.

And I apologize for being so dismissive, that was equally uncalled for on my behalf. :rose:

Since you were kind enough to augment your post, I have removed the inflammatory commentary that my own added to the dialogue.

My bad but so are you. 385MM shares fully diluted (nv 10Q) at say $18 (yahoofinance price history) = $7B (current valuation $8B fully diluted)

So AMD did get a discount on ATI and under independent management, would probably still be worth close to what they paid for it, just like NVDA is today. Though the opportunities for ATI were probably greater (SOC but also Imageon, console and APU).

opportunity cost? yes

overextended themselves? yes

mis-managed ATI? yes

overpaid? Nope

If AMD didn't over-pay then they would not have had to take write-downs on it afterwards.

ATI was over-valued, their shareholders were over-valuing the company at the time AMD made a bid on it.

To be sure AMD could not have purchased ATI for less than they did at the time they purchased ATI.

But to be sure they paid more than it was worth, as proved by the fact they had to write-down the value afterwards.

 

[Idontcare]

I saw this at EETimes, full of the usual stuff we'd expect, but contained one nugget that confirms what we have long suspected:

AMD recasts foundry deal to save cash

In addition, AMD said it will reduce R&D payments to Globalfoundries by approximately $20 million per quarter over the next several years as it moves off proprietary AMD manufacturing processes and onto standard Globalfoundries' CMOS process technologies at 28 nm and below.

source

That means no more SOI. Probably a good thing.

 

[meloz]

but contained one nugget that confirms what we have long suspected:

Have we not known this for a long time already? Isn't this what Common Platform is all about?

 

[lothar]

Must be one of those subscriber-only articles? Regardless, those numbers are dead awful. AMD really got hamstrung by Dirk's and Hector's creative accounting shenanigans when they spun off GloFo.

Search Google with the exact title of the article and click on the first link.
That's how to read all WSJ articles for free.
http://www.google.com/search?q=AMD+...&sugexp=chrome,mod=3&sourceid=chrome&ie=UTF-8

 

[TekDemon]

I'd like to see their numbers

You'd think they could do something else, like sell them $1 over cost, or bundle them with GPUs - buy a 7970 GE and get a 8350 for $100.

My best guess is that they're trying to avoid having to dump the CPUs super cheaply because that would dilute what value they have left in their chips-if people get too used to paying $30 for AMD cpus they won't be able to charge more when they want to later on.

Still....this is a very costly way to go about it.

 

[mrmt]

My best guess is that they're trying to avoid having to dump the CPUs super cheaply because that would dilute what value they have left in their chips-if people get too used to paying $30 for AMD cpus they won't be able to charge more when they want to later on.

They are past that point already. For how much do you think they are going to sell those Llanos that took the inventory markdown charge last quarter? As IDC pointed out, they are in the point of not being able to sell processors at all.

 

[IntelUser2000]

IMO its understandable why AMD went the Bulldozer route. It was a gamble of their own. We now know it wasn't worth the risk, but that's the whole point of it. You never truly know the outcome until you take the risk. Enhanced Thuban might have turned out better but they weren't expecting that. And even then it would still be behind. So they opted to going Bulldozer which had a slim chance of being a wild success.

But that's about the end of something that was really out of their control. Whatever Rory seems to be doing since he rose to power at AMD is all wrong. They need to cut the crap and consolidate their product lines for one thing. Seperate ARM core division in a deeply x86-rooted company? RIDICULOUS! They really think licensing ARM cores will give them an advantage when there's lot more manufacturers there? Do they still have the mentality of going "against Intel" and ARM is somehow their solution out? Why not use the same resources to advance their low power core?

I'm not sure they have any way out anymore.

 

[mrmt]

IMO its understandable why AMD went the Bulldozer route. It was a gamble of their own. We now know it wasn't worth the risk, but that's the whole point of it. You never truly know the outcome until you take the risk.

As they say fool me once, shame on you, fool me twice, shame on me.

What was launched in 2011 wasn't the first generation Bulldozer, but the second one. The first one was supposed to be built on 45nm and was scrapped by Dirk in 2009.

At that point it should have been clear some of the weakness of the architecture, such as the high dependency on high clocks, the large die size, the lower IPC (you know, that useless metric that JF-AMD kept saying that it wouldn't go down with a straight face) and the atrocious cache performance, and even so they proceeded with the project that resulted in Orochi.

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?

A lot of checks and balance didn't work at AMD. What probably happened at AMD was lots of egos tied to the project, lack critical analysis of the evidence gathered in the development and a lot of wild assumptions regarding the performance of a process node that wasn't even their own development in the first place, and more important, it was not a failure of a single individual or group of individuals, but failures of individuals coupled with a failure of the structures that should prevented or at least reacted to those things a failure of AMD as an institution, that allowed a failure of such a magnitude to develop on this huge time frame.

But that's about the end of something that was really out of their control. Whatever Rory seems to be doing since he rose to power at AMD is all wrong.

For a lot of reasons I think the guy is clueless. He is falling back from a market with only one big competitor to get to a market with a lot of big competitors + their old nemesis and spreading thin their scarce development resources, when he should fight less fights but bring a bigger stick on each of them. But still, after Hector Ruiz, he would have to fight hard to get the title of worst AMD CEO ever.

 

[OILFIELDTRASH]

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

 

[Lepton87]

Brutal. Almost 'binary'. Either you lead or you die. No middle ground. Hopefully the 8000 series GPU will help the cashflow.

With AMD poor brand reputation and nVidia strong and viral marketing it does not matter. Even if they have both cheaper and faster cards nVidia still sells a hell of a lot more cards.