Shortcuts

New Details Emerge On AMD's CEO Dismissal

rated by 0 users
This post has 13 Replies | 2 Followers

Top 10 Contributor
Posts 25,665
Points 1,155,165
Joined: Sep 2007
ForumsAdministrator
News Posted: Thu, Jan 13 2011 9:22 PM
It's been two days since AMD's Board of Directors announced it was firing then-CEO Dirk Meyer and new information on the board's reasoning has begun to appear. As we suspected, it was Meyer's decision not to focus on a new ultra-mobile processor that incurred the board's wrath. The particular incident in question is believed to be the decision to sell AMD's Imageon technology to Qualcomm in January of 2009 for $65 million. Qualcomm took the ARM-based Imageon core, integrated it with the company's own Snapdragon SoC, and has produced a line of profitable handsets. This, apparently, cheesed the BoD at Sunnyvale right the heck off.

This information has sparked fresh conversation over whether or not Meyer did the right thing back in '09 and if the board was justified in firing him. The charts below explore this question by examining AMD's financials from several angles.  

Note: All of the charts below were generated by YCharts. In many cases, the charts only run through the end of 2009 / early 2010. While we'd prefer charts that ran through 2010, the majority of financial websites don't offer the same degree of comparison or track as many separate statistics. Since we'll be focusing on 2006-2008, the time range is adequate.



Note that this is profit margin not gross profit margin. Gross profit margin is calculated by measuring the difference between the manufacturing price and the selling price of a product. Profit margin is how much money the company makes after all other expenses are accounted for. While AMD's profit margin was negative from September of 2006 through September of 2009, the critical crunch was in September of 2008. At that point, the company was losing an average $1.24 for every product that it sold.



AMD's revenue changes from quarter-to-quarter are measured above. In this case, negative revenue growth means that sales fell during the period in question; positive revenue growth means that revenue increased. Again, we see AMD struggling badly from midway through the second quarter 2008 through Q3 2009.




The current ratio is a measure of a company's ability to pay short-term debt and current liabilities for an estimated one year period. A value of two is considered healthy. A value of one (which AMD nearly hit in September of 2008) would mean the company had one dollar of assets for every $1 in liabilities.




Here's the last graph we'll reference; it's a comparison of AMD's long-term debt to the amount of cash the company had on hand at any given point. Cash on hand reflects a company's solvency (could AMD repay all debts by selling all assets) and liquidity (can AMD meet its ongoing costs without sacrificing physical assets like equipment, fabs, etc). In September 2008, AMD had four times as much long-term debt as it had cash. That was actually a ten year record—but the company broke it in December of that year when the ratio rose to 4.42x.  The debt situation is much better now—the company was carrying just $2.19B in Q3 2010—but the company's cash on hand has shrunk to 620M.

From Hindsight to White Washing

Pick virtually any measure of financial health you want, and AMD's nadir was between Q3 and Q4 of 2008. What the charts don't reflect is that AMD was selling property hand-over-fist through the same period. From August to December, AMD sold 200mm fab equipment to Angestrem for $192 million, Xilleon to Broadcom for $150 million (down from an initial $192M), and decided to sell Imageon to Qualcomm for $65 million. Without the $407M in cash from those sales, the picture would've been even worse.

When Dirk Meyer took over, AMD's continued existence was factually at risk. Shanghai was still six months away and the GlobalFoundries spinoff hadn't been finalized yet. Financially, the company was a wreck. It had just written off $2.2 billion in assets, had no cash reserves to leverage, and virtually no room to maneuver on any front. The only thing Meyer had to peddle was his promise that Shanghai would deliver, he'd slash costs, sell non-core assets, and turn the company around.

The great irony of this story is that Dirk was canned for his solutions to problems Hector caused. Scroll back up and look at the same charts but eyeball early 2006 instead of mid-2008. In March of 2006, AMD abandoned its plan to aggressively pay down long term debt and instead borrowed approximately three billion to pay for ATI. Then the economy crashed, Barcelona turned out to be a sick joke, Bulldozer, Bobcat, and Fusion were all delayed, and Intel's Core 2 Duo turned the Athlon 64 into mincemeat.

If AMD had paid fair market value for ATI, the company's long-term debt would've been much lower or it would've had the cash on hand to respond more proactively to the problems that struck in 2007-2008. Even if the company's sales and launches had played out the same, its negotiating position would still have been improved. When Hector bought ATI he put a noose around the company's neck betting that the platform wouldn't drop out from under it. It's quite possibly the worst bet AMD ever made.

After checking the financial data we're even more unsettled by Meyer's 'retirement.' Even if selling Imageon was a genuine ***-up, the decision was made during the worst crisis AMD's suffered at a time when the company's future existence was in question. Worse, firing Meyer strongly implies that the Board of Directors is willing to play a twisted game of Let's Pretend with both Meyer's initial decision to sell and AMD's actual chances of whipping up a new mobile part that's market competitive anytime soon. It's not just our opinion—the finances don't back this up. 
  • | Post Points: 125
Top 10 Contributor
Posts 5,053
Points 60,700
Joined: May 2008
Location: U.S.
Moderator
3vi1 replied on Fri, Jan 14 2011 8:42 AM

I knew there had to be a reason, given that they didn't immediately have a permanent replacement. But, I didn't suspect the reason would be so bromidic. I was *so* hoping for sex scandal.

I still think this is less about what he did in the past vs. his immediate direction for the company's future. I hope the board knows what they're doing, or they're just as likely to replace him with someone that runs the place into the ground.

Oh well, I wish him the best of luck at his next company.

What part of "Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn" don't you understand?

++++++++++++[>++++>+++++++++>+++>+<<<<-]>+++.>++++++++++.-------------.+++.>---.>--.

  • | Post Points: 20
Top 100 Contributor
Posts 1,063
Points 11,520
Joined: Jul 2009
Joel H replied on Fri, Jan 14 2011 9:59 AM

3vi1,

That's actually a big part of why I question the board's decision. Assuming that AMD had a decent Q4, the company is at best back to 'completely broke.'

In the next 12 months AMD is launching two major product families in critical product segments and finalizing the 'Enhanced' Bulldozer and Bobcat processors currently set to launch in H1 2012. It's also launching two new Opteron processors--Valencia and Interlagos. AMD's Bulldozer server cores are going to have a few tricks that the mainstream/enthusiast flavors won't have, and they'll pack 6-16 cores and offer 4HT links. Let's not forget the mobile and desktop platforms, either--both mobile and desktop platforms are going to see some enhancements/changes.

Then, of course, there's the new 28nm GPU codenamed 'Southern Islands' that'll form the basis of what's tentatively called the Radeon 7000 series. That processor will be built on 28nm--it'll be the first Radeon GPU built by GlobalFoundries instead of TSMC and GlobalFoundries first 28nm high-end GPU.  Since the 6970 is based on 40nm and the 7000-series is expected to be 28nm flavors of the same, AMD may also have to tweak the chip both for the die shrink and the foundry swap.

2011 is almost certainly the busiest, most complex, most important launch year in the company's entire history. The schedule above would be respectable even if Intel were the one executing it; it's more than AMD has ever taken on. It's not that I'm underestimating AMD for being AMD—I literally see no room (and no additional funds) for the company to kick-start a new GPU or CPU/GPU SoC based on ARM technology. We know Meyer was backing Bobcat and counting on ramping it into tablets as well as netbooks; the board has stated it wasn't pleased with Meyer's mobile roadmaps. 

Planning a Bobcat refresh for H1 2011 is aggressive as it is; firing Meyer now suggests the board wants to launch a project or projects in the near term that he wasn't going to support. It's not just that there's no realistic room on the company's plate to handle yet more work--who's going to do the work, and how're they going to pay the megabucks it'd take to whip out a new architecture in less than 2-3 years?

  • | Post Points: 20
Top 10 Contributor
Posts 8,555
Points 103,030
Joined: Apr 2009
Location: Shenandoah Valley, Virginia
MembershipAdministrator
Moderator
realneil replied on Fri, Jan 14 2011 9:59 AM

3vi1:
I was *so* hoping for sex scandal

Ha-HA!

3vi1:
Oh well, I wish him the best of luck at his next company.

Me too.

Dogs are great judges of character, and if your dog doesn't like somebody being around, you shouldn't trust them.

  • | Post Points: 5
Top 200 Contributor
Posts 416
Points 3,205
Joined: Dec 2010
LLeCompte replied on Fri, Jan 14 2011 11:05 AM

Greed is why he was fired. The BoDs are getting mad that they are losing out to intel and had to take it out on someone. The tablet market may or may not blow up; i mean people are kinda starting to buy them but it could slack off. I would have ignored the tablets too and focused on the upcoming launch. Hopefully AMD can rebound and keep from being bought by intel.

  • | Post Points: 5
Top 10 Contributor
Posts 5,053
Points 60,700
Joined: May 2008
Location: U.S.
Moderator
3vi1 replied on Fri, Jan 14 2011 11:35 AM

I don't know if I wouldn't have made the exact same decision about Imageon given what was going on at the time and their need to concentrate on core business. I'm sure there's another side to the story, but it certainly does sound like the BoD's doing Monday morning quarterbacking.

Hopefully two things will happen: Dirk will find a company whose vision is more in line with his, and AMD finds someone that can continue to improve the company and give Intel and the ARM producers decent competition.

What part of "Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn" don't you understand?

++++++++++++[>++++>+++++++++>+++>+<<<<-]>+++.>++++++++++.-------------.+++.>---.>--.

  • | Post Points: 5
Top 100 Contributor
Posts 1,063
Points 11,520
Joined: Jul 2009
Joel H replied on Fri, Jan 14 2011 12:17 PM

*Imageon. Not Infineon. :P

  • | Post Points: 5
Top 10 Contributor
Posts 4,814
Points 45,635
Joined: Feb 2008
Location: Kennesaw
rapid1 replied on Fri, Jan 14 2011 12:25 PM

I think as someone said in a post they were mad (the BOD), and looking for someone to blame. One thing probably both identifiable, as well as indicative that I see is to look at who heads this board individually, and then what they did in the past.

Yes; many of them were big names at one time in computer history, but most of those big named companies also at best failed eventually as well. This looks like the blame game, and at best one by someone with bad aim. Yes; these technologies sold may have in the future made money, but at that time AMD had no future either, that is unless something they did have was sold. It also seems curious to me that a board would not have to be consulted, and approve the sale of something from a companies patent/idea book like this as well.

So at best it seems like someone looking for some recognition on tough decision making or something. I personally think this was a bad idea, and I think this because the person they let go has been at the head if not directly responsible for many of there successes, at least at the bottom line if not to a greater percentage. As stated in the article this CEO was trying to fix Hectors screw ups, largely on top of that many if not most of Hectors positive actions were led by Meyer, but at that time he was in a different position.

While I do not agree now nor did I when they sold that technology, I also know they would have had to develop there own "Snapdragon" too, for it to have been any use to anyone at AMD at that time.

OS:Win 7 Ultimate 64-bit
MB:ASUS Z87C
CPU:Intel(R) Core(TM) i7 4770 ***
GPU:Geforce GTX 770 4GB
Mem:***ingston 16384MB RAM
  • | Post Points: 5
Top 500 Contributor
Posts 102
Points 1,315
Joined: Jun 2010

Superb analysis, Joel... and your follow up comment is insightful enough to justify an additional opinion article so that it can be more readily read.

  • | Post Points: 20
Top 10 Contributor
Posts 5,053
Points 60,700
Joined: May 2008
Location: U.S.
Moderator
3vi1 replied on Sat, Jan 15 2011 4:14 PM

>> Superb analysis, Joel

/agree. Good info. Keep it up.

What part of "Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn" don't you understand?

++++++++++++[>++++>+++++++++>+++>+<<<<-]>+++.>++++++++++.-------------.+++.>---.>--.

  • | Post Points: 5
Top 25 Contributor
Posts 3,654
Points 29,000
Joined: Jul 2004
Location: United States, Texas
Drago replied on Sat, Jan 15 2011 4:19 PM

Board of Directors are all tools. They rule on stuff that is way to big for their britches. Most people on boards dont know diddly squat about the business they are in and make decisions based upon numbers and charts instead of getting their hands dirty and actually spending time working to make their billions. AMD has never been able to get into the laptop market till they got ATi and they never had anything for the mobile market so selling something that is practically useless unless a ton of money got invested into that mobile sector Dirk made money when AMD needed it.

A+ Certified PC Repair Technician
Associates Degree in Computer Science
Bachelors Degree in Computer Information Systems

DFI Lanparty UT NF3 250GB Dead.......Replacement  Abit KV-85
Learn more about Comp TIA A+ Certification.

  • | Post Points: 5
Top 100 Contributor
Posts 1,063
Points 11,520
Joined: Jul 2009
Joel H replied on Sun, Jan 16 2011 4:29 PM

One thing I want to add: AMD has one reasonably straightforward option where Bobcat is concerned *if* GlobalFoundries can deliver on it. One of the reasons AMD has fallen behind Intel in fabrication technology is because Sunnyvale can't afford to aggressively tune foundry production. Instead of targeting high speed grades, for example, AMD has concentrated on lowering its cost of goods sold by maximizing the number of CPUs it can sell per wafer.

That's the bread and butter of any CPU manufacturer, but it doesn't allow AMD as much room to focus on ultra-high performance or ultra-low power chips. (The latter is somewhat lessened by the natural tendency of CPU quality to improve over time. Check first-gen AMD CPus going back the last five years, then compare them to later chips at the same frequencies and cache sizes--you'll often find that the company knocks the fastest chips from 125W at launch to a 95W second-gen.

If AMD had enough money, the company could bring its 28nm Bobcat parts to market more quickly if it was willing to accept the corresponding increased in production costs that would come as a result from launching the chip on an early 28nm production line as opposed to waiting 6-9 months. How well this works would depend significantly on how well GF's 28nm production lines are ramping. Best case, it's a great move. Worst case, it's TSMC's 40nm all over again.

It's also possible that AMD hypothetically try to bin a lower-power port from TSMC or make some modifications to that design if it can introduce some small changes that would reduce power consumption further without requiring any sort of significant change. Ultimately though it comes to the same--there are a few production-centric changes AMD could make that might increase its product ramp *if* the company can take the higher production costs.

This would seem to satisfy both the reason Meyer might oppose such an initiative (it's economically unsound for a shoestring budget) and the reason why the board might push for it. Perhaps more specifically, it provides an option that the board might push for that doesn't first require the combined BoD to possess the intellect of a lobotomized sea turtle.

  • | Post Points: 20
Top 10 Contributor
Posts 8,555
Points 103,030
Joined: Apr 2009
Location: Shenandoah Valley, Virginia
MembershipAdministrator
Moderator
realneil replied on Sun, Jan 16 2011 7:04 PM

Joel H:
possess the combined intellect of a lobotomized sea turtle

Ahahahahahah!

Such a way with words!

Dogs are great judges of character, and if your dog doesn't like somebody being around, you shouldn't trust them.

  • | Post Points: 20
Top 100 Contributor
Posts 1,063
Points 11,520
Joined: Jul 2009
Joel H replied on Sun, Jan 16 2011 9:07 PM

Aye...Sea Turtles...

/Captain Jack Sparrow

  • | Post Points: 5
Page 1 of 1 (14 items) | RSS