[Beyond3D Article] AMD Q1 2007 Warning Analysis

Referring to the other thread about 'real men have their own fab': this is probably a textbook case about the dangers of owning a fab. They may be better for margins when times are great, but they are a disaster when times are not so good: it's hard to reduce carrying costs of a fab without selling it, whereas you can just reduce orders in a fabless model. (As 3D pointed out, they don't apparently have a choice because of contractual reasons...)

Well, the corollary of "Real Men own Fabs" is "Real men also know how to run the rest of the company to maximize the benefits of owning fabs."
 
Up to $2.2 billion @ "B" rating? The interest rate will be insane. The rumors of private equity buyout must have been unfounded, because this is a huge "We are desperate for cash at any cost" move.

Can we now officially conclude that ATI deal, the way it was structured, was a boon for ATI shareholders and a disaster for AMD, or do we need even more evidence?
 
Up to $2.2 billion @ "B" rating? The interest rate will be insane. The rumors of private equity buyout must have been unfounded, because this is a huge "We are desperate for cash at any cost" move.
Indeed. It will be interesting to watch the "Interest Expense" category in Q2/Q3...

Either way, even with high interest rates, this should provide some short-term relief, so they will now be able to last well into the Barcelona generation. On the other hand, this will force them to have higher operating profits to compensate, so it does raise expectations on their new architecture even further...
 
For the desktop, Agena or Kuma is not likely to beat Penryn, and unlikely to beat Conroe. Barcelona's big changes deal with FP and SSE processing and quad-core optimizations, but the integer side is not radically different from K8.

http://www.fudzilla.com/index.php?option=com_content&task=view&id=663&Itemid=1

Fudo said:
In Sandra multimedia benchmark Agena / Barcelona K10 2.5 GHz scores over 400.000 it/s (instructions / second) in integer test and about 300.000 it/s in floating point test.

Intel's fastest QX6800 Quad core at 3 GHz scores about ~300.000 it/s in integer and ~180.000 it/s in floating point test.
 
That is largely irrelevant for the desktop market if those higher scores are achieved through better quad-core scaling, however. Aggregate performance is important, of course, but so is per-core performance... If true, it does bode well for servers, however.

Also, another factor to consider, especially but not exclusively for desktops: even if some applications are quad-core-ready, that doesn't mean they use all 4 cores 100% of the time. What happens if it's single-threaded for 50% of the quad-threaded for the other 50%? This could happen even if the two run in parallel, if the threaded part of the program finishes before the serial one... (and yes, this is a direct result of Amdahl's Law - of course, the algorithm can often be improved to minimize its impact, but that requires more programming effort, which you are unlikely to see in the first PC games capable of exploiting many-cores architectures imo)
 
Sandra's int benchmark actually uses integer SSE ops, so it's not the integer side of the core I was originally addressing.

Most of the intensive work is still going through the SSE pipelines.

This is a possible good sign that Barcelona should do well on SSE workloads that scale as well as Sandra does to multiple cores, though I'd take a few sentences about a Sandra benchmark from Fuad with a large grain of salt.
 
Advanced Micro Devices Inc. (AMD) priced $2 billion of 6% convertible senior notes due in 2015 in a private placement to qualified institutional buyers.
The Sunnyvale, Calif., provider of processing systems in the computing, graphics and electronics markets granted the initial purchasers a 30-day option to buy up to $200 million of additional notes to cover overallotments.
Advanced Micro also entered into a capped call transaction in connection with the offering which is capped at $42.12 a share.
The company intends to use a portion of the net proceeds to pay the cost of the capped call transaction. Advanced Micro said if the initial purchasers exercise their option to purchase additional notes, it expects to use a portion of the net proceeds to enter into an additional capped call transaction.
The company expects to use at least $500 million of the remaining net proceeds to repay a portion of the term loan it entered into with Morgan Stanley Senior Funding Inc. to finance a portion of the acquisition of ATI Technologies Inc.
Shares of Advanced Micro were down 19 cents, or 1.4%, to $13.85, in recent trading.
-Veronica Dagher; 201-938-5400; AskNewswires@dowjones.com > Dow Jones Newswires
04-24-07 0940ET
Copyright (c) 2007 Dow Jones & Company, Inc.

Can someone explain this one to me?
 
Sounds like some kind of corporate bond issuance.

edit:

The capped call (from wikipedia) seems to reference allowing investors the option to buy the right to buy AMD stock. If it goes up in price, the investors can purchase the stock at the price they paid for the call, and if it goes down, they only lose the value of the premium they paid.
 
Up to $2.2 billion @ "B" rating? The interest rate will be insane. The rumors of private equity buyout must have been unfounded, because this is a huge "We are desperate for cash at any cost" move.

Can we now officially conclude that ATI deal, the way it was structured, was a boon for ATI shareholders and a disaster for AMD, or do we need even more evidence?

AMD should never have purchased ATI at the time that they did. They should have focused on getting Barcelona to market, then witnessed ATI's performance on getting R600 to market, and then make a decision after that. Now they are stuck with two delayed product lines, and boatloads of debt. On top of that, by combining ATI with AMD, they have made big competitors Intel and NVIDIA even more aggressive and eager to stay on top of their game. It's really unfortunate.

And ATI should never have tackled so many custom GPU projects for outside vendors like MSFT and Nintendo. ATI was in a clear leadership position with R300, but they were forced to devote many talented engineers to highly custom projects like the Xbox 360. On the other hand, NVIDIA just recycled their 7 series architecture for the PS3 without having to do a highly customized architecture. After all, it's not the video card that sells consoles, it's the other features and games. As a result, ATI had to delay one architecture after another while NV keeps beating them to the punch after R300 generation.
 
Last edited by a moderator:
Sounds like some kind of corporate bond issuance.

edit:

The capped call (from wikipedia) seems to reference allowing investors the option to buy the right to buy AMD stock. If it goes up in price, the investors can purchase the stock at the price they paid for the call, and if it goes down, they only lose the value of the premium they paid.

The convertible is the bond issuance. AMD or investors have the right to convert it to stock if the price of AMD's stock goes above $28 by a certain time.

The capped call is separate, and is basically a call option that AMD is buying on its own stock which will kick in if the stock price goes above $28. But unlike a regular option, the potential gains are capped at $42. This lowers the option premium that AMD must pay. A<D is saying this is a hedge against dilution if the covertible gets excercised = ie the proceeds from the option could be used to buy back shares.

a very odd move in my opinion. it is very rare for a company to make such a direct bet on its own stock price, and my might particularly question the wisdom of buying calls when finances are stretched. mayeb it was driven by a desire to appease current investors. If AMD stock get back to $28 again AMD management and investors should just be thankful.

not that there is any impropriety here, but there is something about it that seems enronian, if i may coin that adjective.
 
Isn't there a time limit that the stock has to reach this price though?

edit ah you already said that lol, that sounds very risky.
 
humm

I play the stock market and amd just looks screwed.

my biggest concern is what will amd do w/ all that inventory as intel keeps on cutting prices?

going into next quarter i see massive losses and a massive inventory write off.
 
This is ugly and in all my years following AMD the worst I have seen it. Back in the K6 days they were losing like 10s of millions a qtr, not nearly 3/4th of a billion.

The ATI deal was a terrible idea and done at the edge of a cliff. They are saddled with a delayed R600 and a delayed Barcelona and not enough cash to do anything with them.

I wouldnt buy their stock nor bonds until they can deliver even a glimmer of hope.
Intels pricing structure through the summer is going to kill AMDs already pitiful margins. And most likely the R600s will also see depreciation unless they can seriously outperform the 8800 series.
 
http://biz.yahoo.com/bw/070427/20070427005623.html?.v=2

Whee, that was damn quick. Suggests pretty strongly that institutional buyers still have a lot of confidence in the long term viability of the company.

SUNNYVALE, Calif--(BUSINESS WIRE)--Advanced Micro Devices, Inc. (NYSE:AMD - News) today announced the closing of its offering of $2.2 billion aggregate principal amount of 6.00% Convertible Senior Notes due 2015, including $200 million of notes that were issued in connection with the exercise in full of the initial purchasers' over-allotment option. The notes were privately offered to qualified institutional buyers pursuant to Rule 144A under the Securities Act of 1933, as amended (the "Securities Act").

In connection with the offering, AMD entered into capped call transactions with an affiliate of one of the initial purchasers. The capped call transactions are intended to reduce the potential dilution to AMD's stockholders upon any future conversion of the notes. The capped call transaction effectively will increase the conversion price of the convertible notes to $42.12 per share of AMD's common stock, representing a 300% premium relative to the last reported sale price of $14.04 per share of the common stock on April 23, 2007.

AMD estimates that the net proceeds from the offering, will be approximately $2,169 million, after deducting discounts, commissions and estimated offering expenses. AMD used approximately $182 million of the net proceeds of the offering to fund the cost of the capped call transactions. AMD used $500 million of the remaining net proceeds to repay a portion of the term loan AMD entered into with Morgan Stanley Senior Funding, Inc. to finance a portion of the purchase price of, and expenses related to, the acquisition of ATI Technologies Inc. AMD will use the remaining amount for general corporate purposes, including working capital and capital expenditures.
 
With 6% AND first dibs on all assets if anything goes horribly wrong, it's not a very risky proposition. This move however will cap their share price below $28 for years and year.
 
http://www.eetimes.com/news/semi/showArticle.jhtml?articleID=199202315
Intel in March acknowledged that for three and a half months after AMD filed its suit, a small number of employees whose e-mails were considered potential evidence failed to move all messages to their hard drives, which means they would have been purged automatically from Intel's system. In addition, "a few" employees believed erroneously that Intel's IT group was automatically saving their e-mails.

The disclosure brought an angry response from AMD. While not accusing Intel of intentionally destroying evidence, the company questioned the effectiveness of the procedures Intel put in place to protect potential evidence. Farnan ordered Intel to determine what was lost and whether there was any damage to potential evidence.

Determining whether Intel was negligent is important because of the punishment the company could face. If Farnan decides Intel failed to take proper steps, he could fine the company million of dollars. Worse, the judge could decide during the trial to instruct the jury that they should assume that the lost e-mails would have been detrimental to Intel's defense. Such a move could play a role in swaying the jury toward AMD.
 
AMD Financing Terms: Even Stranger Than Expected (AMD)

The investment bankers were probably laughing over and over on this one. It is a true voodoo financing and one reminiscent of past blow-ups. We aren’t just trying to kick the company while it is down, but this one takes the cake. It really must be true that the more things change, the more they remain the same.
Shares are down 2% today to $13.54 and are all the way right back down to levels right before this financing was announced (and priced the following morning). A conspiracy theorist would probably say that the financing was to try to wrestle control or to further hold AMD hostage if things don’t get better. We’ll stop short of that because it is just a head-scratcher all the way.

"Voodoo Financing"
 
Mercury News is reporting that AMD has cut 430 jobs (2.6% of it's workforce).

San Jose Mercury News said:
[...]
In addition to the local job cuts, AMD will pare 80 positions in Austin, Texas, and 50 in Markham, Ontario, where the company's ATI Technologies unit is located. No jobs were lost in Dresden, Germany, where AMD has chip making plants.

The Sunnyvale company said most of the jobs came from sales, marketing and administration. A few engineering positions were cut, but no engineering projects were affected, said spokesman Michael Silverman. The company had 16,823 employees worldwide at the end of April.
[...]
 
Back
Top