The rise and fall of AMD

[Idontcare]

Fabless doesnt work on the long term , where are the once
fabless CPU designer a la DEC ?..

DEC was not fabless, they produced their chips in their own fabs.

 

[Abwx]

So you are saying both AMD and ARM are doomed, along with Qualcom etc.

ARM and Qualcom are in a better situation since they own
the whole IP of their CPUs while AMD is stuck in cross
lincensing agreements that allow Intel to have some
control over it a way or another.

On the long term i see no other solution than either
a merger with GF or a buy back of Dresden plant if ever
they have the chance to make a few bucks in the five
coming years.

 

[Exophase]

ARM and Qualcom are in a better situation since they own
the whole IP of their CPUs while AMD is stuck in cross
lincensing agreements that allow Intel to have some
control over it a way or another.

Qualcomm doesn't own the whole IPs of their CPUs, the architecture license is beholden to ARM, and not in a mutual cross-agreement like Intel and AMD's either. Mind you, this is a purchased license and ARM's entire business model revolves around making it available. But I don't see AMD being at a big risk when Intel is stuck with the agreement just as much as they are.

And I definitely don't see how that situation would change based on whether or not AMD has fabs.

On the long term i see no other solution than either
a merger with GF or a buy back of Dresden plant if ever
they have the chance to make a few bucks in the five
coming years.

AMD's products that have the least connection with their old fabs - Bobcat, Jaguar, and the discrete GPUs manufactured by TSMC - have been some of their more successful and probably cheaper to keep making. Meanwhile, their connections with the old fabs are causing them to be sucked dry.

Is it worth having your own fabs if you can't keep up with TSMC? And if GF is struggling to do so after being injected with a bunch of new capital and subsidized by serving a much wider market then how would AMD's manufacturing capability have been any better off had they not spun the fabs?

 

[Idontcare]

Is it worth having your own fabs if you can't keep up with TSMC?

Its really not. Not unless you have some specialized niche product and market (think defense, aerospace, etc) in which case your unique non-generic process node will continue to be an asset to your bottom line.

 

[Ayah]

Its really not. Not unless you have some specialized niche product and market (think defense, aerospace, etc) in which case your unique non-generic process node will continue to be an asset to your bottom line.

Or if you're an oppressive country highly suspicious of other countries.

 

[Fox5]

From reading it, it doesn't look like Raza was upset about the amount of Fabs. It looks like he was just upset about using borrowed money to build them as opposed to waiting until AMD had the cash. Without actually having access to AMD's balance sheets and interest rates at the time I don't think you could really determine who was right.

Or you could say they were both wrong and AMD should have concentrated on making flash memory with the fabs it had and found a way to make its CPUs at a 3rd party fab to avoid the situation they're currently in.

He may have disagreed with the fab, but they came online while AMD was doing pretty well and continued to do well iirc.

I think most people will agree that the big mistake was acquiring ATI for way too much.
Also, AMD should have opened their fabs to other companies (or joined them with a larger company) well before they spun off Global Foundries.

Or if you're an oppressive country highly suspicious of other countries.

I'm still waiting for China to buy up either AMD or GlobalFoundries. Their tech is still pretty cutting edge compared to China's homegrown stuff.

 

[Idontcare]

Or if you're an oppressive country highly suspicious of other countries.

That falls under "defense", as in "national defense"

 

[kalrith]

Part 2 is up: http://arstechnica.com/business/201...e-top-of-the-mountain-to-the-deepest-valleys/

 

[Charles Kozierok]

AMD has been killed many times, still it is alive

Yes, sort of like a guy who jumps off a building and as he passes the second floor goes: "So far, so good!"

 

[Olikan]

Part 2 is up: http://arstechnica.com/business/201...e-top-of-the-mountain-to-the-deepest-valleys/

... Part 2 is almost a copy-paste of hector's book...
and how he "saved" amd :|

 

[guskline]

Fascinating read.

 

[Texashiker]

Part 2 is not as good as part 1.

It seems part 2 of the article focused on poor financial decisions, rather than poor engineering.

Why has AMD suffered over the past few years? Because they are way behind the curve, rather then in the curve or ahead of the curve.

With AMDs current price range and performance, there is no incentive to buy AMD. Back in the 1990s and early 2000s, it was that performancerice ratio that drove AMD.

 

[thilanliyan]

This bit is really disappointing (written by Ruiz):
"Toshiba had accepted a hefty payment from Intel in 2001 on the promise that it wouldn’t use AMD processors. The “market development funds” totaled between $25 million and $30 million per quarter—a sum Toshiba executives likened to “cocaine” because it was a deal they just couldn’t quit.

Intel had bought Hitachi’s exclusivity as well. Whereas AMD had been shipping 50,000 Athlon chips to Hitachi in the first and second quarters of 2002, by the third quarter AMD’s shipments suddenly fell to zero.

NEC’s stance was especially disappointing. By the third quarter of 2002, AMD had won 84 percent of NEC’s Japanese consumer desktop business—a substantial achievement given our historical position as number two in the global semiconductor market. Looking at notebooks and desktops together, we supplied 40 percent of the company’s microprocessor needs. That would end shortly after Intel agreed to pay NEC more than ¥3 billion per quarter, as long as NEC would give 90 percent of its business to Intel and strictly limit its dealings with AMD. By 2003, AMD’s share of NEC’s consumer desktop business had slid to nearly zero too.

NEC went so far as to tell us firsthand about its agreement with Intel, which dictated that AMD’s share of NEC’s Japanese market had to be held to single digits. Globally, AMD’s share of NEC business would fall from 40 percent to 15 percent."

These kinds of practices by Intel hurt AMD more than the $1.25billion or whatever it was that they had to pay in the long term IMO. Lots of "ifs" come to mind, but it certainly wouldn't have helped AMD's situation for that to be happening. I don't believe that was AMD's only problem however.

 

[Ancalagon44]

What I find interesting is the proposed deal between Nvidia and AMD that was rejected because the Nvidia CEO wanted to become CEO of the new company.

As history would have it, maybe it would have been better if they had done that deal instead of the ATI deal? Not that ATI is better than Nvidia, but maybe the Nvidia CEO is more competent than any of the AMD leadership has been?

EDIT: Given how well Nvidia has executed in the past few years, perhaps it was just ego that prevented the deal from happening, even if it would have been better for AMD as a company?

 

[Phynaz]

I actually like the second installment. The story of how integrating ATI was failure - just as what happened when HP tried to swallow Compaq - is interesting. Also the research into the price of the fabs spiraling out of control.

 

[TY-1]

If AMD ever does go down I want it to go down like the end of Terminator 2: Judgement Day, with some damn heartfelt dignity.

Also, if it ever does finally "die", I wonder if we will see a resurgent ATI rise up from the ashes as it seems the GPU department are one of the few areas at AMD that have their shit together.

 

[kalrith]

What I find interesting is the proposed deal between Nvidia and AMD that was rejected because the Nvidia CEO wanted to become CEO of the new company.

As history would have it, maybe it would have been better if they had done that deal instead of the ATI deal? Not that ATI is better than Nvidia, but maybe the Nvidia CEO is more competent than any of the AMD leadership has been?

EDIT: Given how well Nvidia has executed in the past few years, perhaps it was just ego that prevented the deal from happening, even if it would have been better for AMD as a company?

Interesting thoughts. Another weak point for AMD is marketing, which is one of nvidia's strongest suits. It's interesting to ponder such possibilities.

 

[galego]

Yes, sort of like a guy who jumps off a building and as he passes the second floor goes: "So far, so good!"

That guy already open the parachute, it is other, the fat guy, which does not have one and just jumped.

 

[Idontcare]

What I find interesting is the proposed deal between Nvidia and AMD that was rejected because the Nvidia CEO wanted to become CEO of the new company.

As history would have it, maybe it would have been better if they had done that deal instead of the ATI deal? Not that ATI is better than Nvidia, but maybe the Nvidia CEO is more competent than any of the AMD leadership has been?

EDIT: Given how well Nvidia has executed in the past few years, perhaps it was just ego that prevented the deal from happening, even if it would have been better for AMD as a company?

Exactly, AMD opted to go with second-best, on both counts.

They went with choice #2 for their graphics acquisition, and in doing so they kept the status quo of choice #2 for CEO w/Ruiz (versus had they put Jensen into the driver seat at AMD headquarters).

Is it really any wonder when bad management makes bad decisions which in turn further cements the positions of bad management in place to continue making more bad decisions?

Its a cycle that can only be broken by the BoD.

 

[Exophase]

What I found bizarre was Ruiz's comment about the ATi acquisition:

"It was worth it," wrote Ruiz, "even if AMD shares sunk as Wall Street griped that we had paid too much and as our investors worried about the new debt. We had taken on $2.5 billion in financing to pay for the purchase, but we had shed so much debt in the past couple of years that I believed AMD could handle it. With leading-edge graphics technology in our portfolio, we would be able to offer integrated graphics solutions to OEMs just as Intel had been doing. And ATI’s technology was simply better than what Intel had to offer."

So the rationale was that AMD would have been able to recover similar amounts of debt as they had over the last few years. The problem is that it should have been clear to Ruiz that there's no way they could bring in a similar level of profits now that they no longer had the only 64-bit x86 server processor, no longer had anywhere close to the performance crown for desktops, and were way behind in power consumption for laptops. I could sort of understand having this mentality in say, 2003, but 2006??

I can't help but wonder if ATi was treated like a merger because AMD valued it at a price similar to AMD's own value.

Also kind of sad to see that Ruiz's best accomplishment, selling off the fabs, was done so by manipulating the only interested party.

 

[Idontcare]

What I found bizarre was Ruiz's comment about the ATi acquisition:

So the rationale was that AMD would have been able to recover similar amounts of debt as they had over the last few years. The problem is that it should have been clear to Ruiz that there's no way they could bring in a similar level of profits now that they no longer had the only 64-bit x86 server processor, no longer had anywhere close to the performance crown for desktops, and were way behind in power consumption for laptops. I could sort of understand having this mentality in say, 2003, but 2006??

I can't help but wonder if ATi was treated like a merger because AMD valued it at a price similar to AMD's own value.

Also kind of sad to see that Ruiz's best accomplishment, selling off the fabs, was done so by manipulating the only interested party.

It was a rather telling comment of his that he was willing to be so cavalier with the shareholder's equity while the value of their holdings plummeted.

And as you point out, it was equally telling of the character of the man that he felt his value-add in the process was to basically scam the interested party into becoming a bag-holder.

Just the sort of guy you want running your multi-billion dollar company :|

Is it any wonder he setup the vampire-like WSA (take or pay) and wafer-exclusivity to bleed AMD through 2024?

The guy is a class act, a shameless one too who likes to gloat and rub his victim's noses in it. You can bet the Mubadala guys are not laughing after reading of Ruiz's negotiating "prowess". And I thought it was Intel who was supposed to have the questionable ethics?

 

[Olikan]

Also kind of sad to see that Ruiz's best accomplishment, selling off the fabs, was done so by manipulating the only interested party.

yes... himself

 

[MisterMac]

What I found bizarre was Ruiz's comment about the ATi acquisition:

So the rationale was that AMD would have been able to recover similar amounts of debt as they had over the last few years. The problem is that it should have been clear to Ruiz that there's no way they could bring in a similar level of profits now that they no longer had the only 64-bit x86 server processor, no longer had anywhere close to the performance crown for desktops, and were way behind in power consumption for laptops. I could sort of understand having this mentality in say, 2003, but 2006??

I can't help but wonder if ATi was treated like a merger because AMD valued it at a price similar to AMD's own value.

Also kind of sad to see that Ruiz's best accomplishment, selling off the fabs, was done so by manipulating the only interested party.

Yea - i found that hilarious.

"I had to give myself a golden parachute!!! - i had no choice! The other guys almost found out i was fucking with them - so i had to fuck with AMD!".

I never understood why he didn't get more flack from that - or Mubadala didn't get more angry about that comparing to the investment scandal stuff.

 

[galego]

If AMD ever does go down I want it to go down like the end of Terminator 2: Judgement Day, with some damn heartfelt dignity.

I.e., you want them to go down just after defeating the more advanced competition.

Also, if it ever does finally "die", I wonder if we will see a resurgent ATI rise up from the ashes as it seems the GPU department are one of the few areas at AMD that have their shit together.

Terminator 3?

 

[Vesku]

If the AMD board of directors had even the barest sense of direction they would have booted out Ruiz and welcomed in Jen-Hsun Huang as leader of a merged AMD+Nvidia company. It was the most obvious move for ensuring long term share holder value.