Well This Can't Be Good: AMD Axes Carrell Killebrew & Other Employees

[bryanW1995]

QFT

That would be an ideal scenario. The thing is, though, Intel seems to be well on their way to developing a DX11 GPU without even needing AMD or nVidia. I'm doubtful that nVidia would take such a massive risk as well; AMD's CPU division has a terrible history at the high end of the market over the past 5 years, and I can't see nVidia being interested in developing anything that is not cutting edge.

You mean laugh-abee? By 2015 intel will have a gpu to rival the best gpus of 2010 at the pace they're moving. Is that going to be enough for them? Maybe with ati/amd graphics division out of the way it would be, but I think that it would be cheaper for intel to buy up all the ati IP than to reinvent the wheel all over again so to speak. They've already thrown $3 billion at the problem with minimal results, and even those are based upon nerfing IQ to levels not seen since the good old days.

 

[bryanW1995]

Not sure I agree. ATI has consistently been the 'bright spot' in AMD within the past 4-5 years. That span covered some great products from the 3xxx, 4xxx, 5xxx, and 6xxx series'. They released Eyefinity, kept power usage respectable, helped lower mid-range GPUs in the market, and have offered some good mobile GPU options as well. If they were not saddled with a debt-ridden parent (AMD) they might be doing better. The solid things that are making money for AMD today (Llano) are a direct result of ATI expertise. I would be interested to see how fiscally successful ATI could have been under a different company other than AMD.

How is fusion working out for them exactly? They still don't have it for sale yet, and by the time it actually comes to market the company could be bankrupt. I don't have all the data in front of my, but a rough guess is that the ATI division has made around $300 million since the acquisition. $300 million profit on a $5 billion investment over a 5 year period is around 1% annual ROI. Not exactly strong enough to keep pace with intel. There were a LOT of other things they could have done with that money to secure their future, and frankly, AMD chose poorly. I don't know if it will kill them, but they are certainly not the company that they could/should have been if they'd been more astute with their money back in 2006.

 

[PCboy]

Agreed in principle. Unfortunately, AMD really couldn't afford ATI at the price they purchased it for, and it hurt both companies because of it.

Agreed. I wish they hadn't bought them in the first place.

 

[bryanW1995]

People, its the new Boss doing this. Not BD, not HD7k. Nothing specific like lackluster BD or shrinking gpu market has anything to do with these cuts specifially. Its all about trimming the fat away you could say. This is all the CEOs doing.

I strongly disagree. Yes, RR is flexing his muscles and making changes to put his stamp on the company, but would he be doing this if hd7xxx was a home run? They already have a very good idea of what their performance level will be, and they certainly have similar info on Nvidia's upcoming offerings as well. I hope I'm wrong b/c I was planning to go AMD next round, but this is eerily similar to the Dirk lynching before the BD fiasco.

 

[Vesku]

I hope it's just more of a low power, mobile focus and not that the next gen radeon is going to be mediocre. Killebrew struck me as more of an efficient performance and features focused lead.

 

[Silverforce11]

If they are truly clearing out ATI devs which brought them so much success, i am not going to buy the next radeon product from AMD.

I've been an ATI/NV user for a long long time, i could care less about AMD products.

 

[bryanW1995]

Running off Killebrew specifically smacks of a difference in design philosophy. RR seems more like a biggest/fastest/mostest type of guy, he's probably not a very big fan of small ball/power efficiency.

 

[SirPauly]

How is fusion working out for them exactly? They still don't have it for sale yet, and by the time it actually comes to market the company could be bankrupt. I don't have all the data in front of my, but a rough guess is that the ATI division has made around $300 million since the acquisition. $300 million profit on a $5 billion investment over a 5 year period is around 1% annual ROI. Not exactly strong enough to keep pace with intel. There were a LOT of other things they could have done with that money to secure their future, and frankly, AMD chose poorly. I don't know if it will kill them, but they are certainly not the company that they could/should have been if they'd been more astute with their money back in 2006.

How do your figure they don't have APU's for sale?

 

[vshah]

QFT

That would be an ideal scenario. The thing is, though, Intel seems to be well on their way to developing a DX11 GPU without even needing AMD or nVidia. I'm doubtful that nVidia would take such a massive risk as well; AMD's CPU division has a terrible history at the high end of the market over the past 5 years, and I can't see nVidia being interested in developing anything that is not cutting edge.

I imagine if nvidia bought AMD's CPU assets it would be for the x86 license.

 

[Silverforce11]

The only thing profitable for AMD is Llano and ATI gpus. Both are "graphics" products.

Llano could have been much better if manufacturing could keep up with demand. It's a huge market, essentially all mobile space bar massive bricks is a market for Llano. Cheap desktops, HTPC, OEMs etc.

Trinity looks bright. Next gen Radeons looks bright. What upcoming AMD CPU looks promising to you? Cos Piledriver sounds like a pile of crap to me, BD with +10% IPC is still so far behind.

 

[cotak13]

Or Intel.

More likely intel than nvidia. At Nvidia it would be a case of turf war with the bright people there. At intel there is enough money and space for him to maybe get his own project and do something amazing.

 

[cotak13]

People, its the new Boss doing this. Not BD, not HD7k. Nothing specific like lackluster BD or shrinking gpu market has anything to do with these cuts specifially. Its all about trimming the fat away you could say. This is all the CEOs doing.

You are completely right on that one. When a CEO chops 10% after 2 months and a bit in office, he hasn't had the time to really study each and every aspect of the company.

What this really is is a change in direction. Nothing more. He's betting that AMD can keep it's relative market share and revenue for the time being (say about a year or 2 coasting on work already done by those chopped off). And as a result the cut will free up money to reinvest in his new direction (whatever it is). The risk to this move is that he cuts too deep and AMD loses market share and revenue wiping out his "savings". If that happens it's likely he'll be shown the door and the company will be in dire straits like a few years back. And Dirk and Rick will be ROFLTAO. Personally I think what he's doing is extremely risky. AMD's not a company with big cash reserve (hence the need for slash and burn to free money up). And we all know how competitive the graphics market can be. If the new GPU is a flop or is delayed into the market due to these cuts we could well see them file chapter 11 within 2 years.

Also if you guys actually read the stuff online (especially people posting on the CPU forum) you'll know that the original group of people who came up with all the Good CPUs for AMD also designed BD. And that they all left for greener pastures before BD even came close to production. That might be the reason why it turns out a bit of a disappointment.

 

[LOL_Wut_Axel]

Ugh. Please stop calling it ATI. We're not in 2009 (which was a great year for them, but still).

 

[Silverforce11]

As long as ATI staff is still there making gpus and apus, i'm gonna call it ATI.

 

[Idontcare]

Killbrew was one of the brightest lights at AMD. Particularly after reading Anand's backstory on the HD 5000 series. Where does someone like that go now? could be a big score for Nvidia....

Companies are sometimes forced to make decisions in dire times which they'd rather never had been forced to make.

No doubt they would have rather retained Killbrew given that they had done so since the advent of Eyefinity.

But within whatever limited view AMD management has for their future, Killbrew was not viewed as being a critical contributor towards it.

What we don't hear is the other side of the story. For all we know the guy was technically brilliant but a real pompous ass in the workplace and his shortcomings put him on the short list long ago despite his otherwise remarkable contributions.

Every one of us who have worked in a company that saw layoffs has seen this happen. The guy we all wished would get fired but could not be fired because of some otherwise notable (but history by then) contributions to the company...and then along comes opportunity (layoffs) and management takes advantage of it to clean up a bit.

Where he goes from here is a good question, probably onwards and upwards with a new employer like Intel.

 

[RussianSensation]

Let's put it this way: Where would AMD's growth be without Fusion and platforms?

OR, where would Bulldozer be if AMD spent $5.4 Billion designing 2 separate CPU architectures: (1) a truly revolutionary desktop CPU that would actually beat Intel on the desktop/server space; and (2) the most efficient per transistor/watt mobile CPU for laptops? Instead they blew > $5.4 billion (because they did an LBO which means paying interest on the loans too!!) on a GPU division that will take 20 years of present value cash flows to make the investment worthwhile. Any idea why not a single company made a competing bid for ATI at the time?

It's pretty hard to design a revolutionary CPU when you've just saddled yourself with massive debt and interest payments for the next 5-10 years. If AMD's GPU division cost more than $5B to create from scratch, why is AMD's market cap just 4.2B? Obviously the ATI part is worth NOWHERE near $5.4B today. It's probably worth $1-1.5B if sold separately today.

Not sure I agree. ATI has consistently been the 'bright spot' in AMD within the past 4-5 years. That span covered some great products from the 3xxx, 4xxx, 5xxx, and 6xxx series'.

Not sure from which angle was HD3000 series good? It was one of the worst performances from AMD in years, only "bested" by the HD2900 flop. HD3000 was not competitive with GeForce 8.

Regarding the popular HD4000 series: it's very easy to have a successful generation when you are practically giving them away at $199 and $299. From a business perspective, HD4000 generation was only to regain mind-share among gamers after the horrendous HD2900 and lacklustre HD3800 series. I bet it barely made any $ for ATI.

HD5000 series, pretty much the same thing. HD5850 was selling for $270 and HD5870 for $370.

HD6000 series, here buy a $230-240 HD6950 and unlock it into a 6970. That alone pretty much cannibalized the sales of HD6970 series.

AMD GPU division has provided some of the best deals for us gamers. But as a business, it was pretty much sacrificing Average Selling Prices / Profitability for market share. Despite all that, NV has been selling their GPUs for higher prices and making bank, while still having almost 60% desktop discrete market share.

It's very easy to have a good generation in the eyes of hardware enthusiasts when you are selling your high-end single GPU card for $250 that unlocks into a $370 card, while and your competitor is selling theirs for $300-500. But you are in the business of making $$$, not making Joe Smith drool over his unlockable HD6950. Look at HD6850/6870. Those cards are selling for $120-150. You can barely make $$ on that. AMD probably needs to sell 10 HD6870s to make as much $ as NV makes off 1-2 GTX580s.

AMD's current graphics card strategy is not at all like the "winning ATI" of the past, which often sold top-end single GPU cards for $400-650. ATI of the past had far higher cash flows, far higher profitability/margins, far higher average selling prices, and almost never resorted to undercutting its competitor on price by significant amounts (definitely not $100s of dollars). ATI of the past always wanted to have the single GPU performance crown. ATI was never regarded as the "budget" (i.e., cheaper) performance brand. ATI stood on equal footing with Nvidia and the firm's prices reflected their belief that their products were as good.

Historical launch MSRPs:

9700Pro = $399
9800XT = $499
X800XT = $499 and Platinum Edition for $549.
X1800XT = $599
X1900XT = $599, and XTX version for $649

By pricing their high-end cards at $400-650, ATI's mid-range cards sold for $250-300, not for $140-150 like HD6850 and HD6870 are today.....ATI was making more $$ in all desktop segments since it had a much more steep price curve. Today, HD6970 tops out at about $350, hardly anyone cares for HD6990 and AMD's mid-range sells for what historically was a price range for low-end cards.

I am not going to list prices for HD3000, 4000, 5000 and 6000 series. But you know they were MUCH lower. If HD7970 is better than GTX580, AMD should launch it at $499-549, not $369, and actually make $$$. If Kepler is better, they can lower the price. We'll see if Read changes the pricing for AMD's GPU division.

Problem is, now that AMD messed up the entire GPU pricing structure of the market after low-balling HD4000-6000 series, gamers have been spoiled. We now *expect* to pay $299-350 for a high-end AMD GPU and $149-199 for a mid-range GPU. If AMD never resorted to price cuts, we would still be willing to pay $399-499 for a high-end GPU. Now that the consumer is used to the new pricing structure, it's going to be very difficult to AMD to raise prices now.....

 

[LOL_Wut_Axel]

As long as ATI staff is still there making gpus and apus, i'm gonna call it ATI.

Might as well call Apple "Apple Computer" as well.

Yes, it's trivial, but get with the times. ATI DOES NOT exist anymore. It's a dead name.

 

[NIGELG]

Maybe some of the Armchair CEO'S should apply to AMD since they know how to run the company better than anyone else..

And some people wanting to pay 499 to 650 bucks for a high end card...don't worry if AMD folds you'll have your wish because you'll have no choice.

 

[Tsavo]

I just got word that they laid off Ruby.

:|

 

[RussianSensation]

Maybe some of the Armchair CEO'S should apply to AMD since they know how to run the company better than anyone else..

And some people wanting to pay 499 to 650 bucks for a high end card...don't worry if AMD folds you'll have your wish because you'll have no choice.

I am not saying I can personally run the entire company better, but ATI certainly did as a standalone firm vs. how the GPU division is run under AMD. They had provided chips for TVs, had a working smartphone chip design that they sold to Qualcomm, had a far more profitable GPU business and had a chipset motherboard business with Intel. If you carved out ATI's division and sold it today, it would be worth a fraction of its previous self. Clearly, AMD has increased ATI's GPU business value....or rather not.

You can sit here and argue all you want that $300 mid-range and $400-500 high-end GPUs are ridiculous but ATI had no problems selling those and neither does NV today, while still being loved by gamers. Maybe you and I have a different idea of what a successful business entails. To me it means making $$, not giving away your product for a fraction of what you can sell it. Market share is meaningless if you are not making $$ by having it. Should Starbucks sell its coffee at Dunkin Donuts prices? Should Apple sell its laptops at Dell prices? Should Mercedes sell its products at Toyota prices? If you have a premium product, you should price it accordingly. ATI did. After AMD purchased ATI, their graphics cards started to sell at mainstream prices while still maintaining nearly premium performance. Talking about leaving profits on the table.

Look at Apple, only 5% market share in smartphone space, but > 50% of the industry's profits. That's a well-run business. Look at AMD, around 50% GPU market share, and hardly any profits.

Like I said, once a consumer is re-conditioned to pay a lower price for a similar performance level, it's too hard to reverse his/her psychology. If ATI never lowered the price of HD4870 to $299, then you would have never complained about paying $399-499 for that card. Afterall, the GTX280 would have still be sold at $649 by NV. So ATI would still have made $$ and retained its higher profitability. Unlike NV though which can recoup its profit margins in professional / scientific markets with Quadro and Tesla, AMD doesn't have that luxury. So how exactly are they supposed to make a lot of $$ by selling $299 HD4870, $269 HD4890 or $299 HD6950s? Good luck with that.

 

[iCyborg]

Killbrew was one of the brightest lights at AMD. Particularly after reading Anand's backstory on the HD 5000 series. Where does someone like that go now? could be a big score for Nvidia....

It does make me wonder what exactly goes into a decision to let go a senior engineer like that (actually he was a director which is a management path, maybe that was his undoing...). Even if he was underperforming. I mean, wouldn't an underperforming principal staff engineer still be loads better than an overperforming junior engineer and a much bigger loss to the company? Reminds me of that Peter's Principle where everyone is promoted while he can successfully do his job until he rises to his level of incompetence. So no matter how good one is, eventually he'll get to a position that might be over his skills and a possible target for a layoff as he's no longer "pulling his weight"... (I thought I would use he/she instead of he, but it massacred the text, just try to read it, so f**k political correctness, it's a he )

Some areas like PR/marketing were hit much harder than others, but all areas and all locations were hit, there are some pretty good R&D people from ATI that I know were laid off, it may save money on the short term, but can't help to feel that on a long term it's a loss, especially in the GPU section that's been doing fairly well, competition-wise. Perhaps they expected more when they paid $5b...

 

[iCyborg]

Like I said, once a consumer is re-conditioned to pay a lower price for a similar performance level, it's too hard to reverse his/her psychology. If ATI never lowered the price of HD4870 to $299, then you would have never complained about paying $399-499 for that card. Afterall, the GTX280 would have still be sold at $649 by NV. So ATI would still have made $$ and retained its higher profitability. Unlike NV though which can recoup its profit margins in professional / scientific markets with Quadro and Tesla, AMD doesn't have that luxury. So how exactly are they supposed to make a lot of $$ by selling $299 HD4870, $269 HD4890 or $299 HD6950s? Good luck with that.

GT200 was a much larger chip than RV770. Your way would be to have higher margins but also allow nVidia to have decent ones, their way was to have lower margins but force nVidia to have much lower and possibly borderline profitable. I can't say which approach is better without knowing more...

 

[RussianSensation]

GT200 was a much larger chip than RV770. Your way would be to have higher margins but also allow nVidia to have decent ones, their way was to have lower margins but force nVidia to have much lower and possibly borderline profitable. I can't say which approach is better without knowing more...

Please explain to me why HD4870 was $299 when GTX260 was $399? HD4870 was at least as good as a GTX260. In fact, NV had to bring out a GTX260 216 to actually compete with an HD4870. Under such a situation, AMD should have priced the 4870 at $379-399 since it was actually better than the original GTX260. Talking about dropping the profitability ball.....

Wait a second, is AMD's purpose to bankrupt NV by lowering NV's margins OR to make $$$ for its own shareholders?

What's wrong with allowing NV to have decent margins? You should be worrying about making your products better than competition's, so you can price them higher and make more $$, not worry about your competitor's margins. If you make a premium product and the consumer is willing to buy it, everything will work out fine. Why undercut your premium product from day 1 if you think it's competitive? AMD isn't Walmart. It doesn't have large enough economies of scale to undercut NV for an extended period of time to make NV financially weak. Also, high-end discrete GPUs aren't like toilet paper that people buy once a month. So you can't really offset low margins by selling them by truckloads.

So why in the world would you lower your own margins to try to financially hurt NV? That would end up with you hurting first before NV even breaks a sweat since NV makes $ in the lucrative professional / scientific markets and also has Tegra. Put it this way, how is AMD selling Derpdozer for $280, but have been selling their excellent HD6950 for 8 months for just $230, that also unlocks into a $330 6970?

 

[iCyborg]

Both Intel and nVidia have better brand recognition than AMD. If they priced 4870 at $400 they might have had better margins, but lower market share because nVidia is "the way it's meant to be played"...

Look, I'm just speculating, given the chip sizes their margins may have been in line with nVidia's, and if they know nVidia can't cut the price to match them, they can get market share and have solid profit, and put pressure on nVidia. How do you know their margins at $300 weren't still better than nVidia's at $400? And IIRC, nVidia was totally dominating with 8800 GTX for a while, they had to catch up.

I won't comment on Bulldozer pricing, but I thought they "fixed" 6950 unlocking a while ago? And $100 difference, wasn't that the 1GB version that couldn't be unlocked? I don't recall ever seeing 6950 under $270, and I'm from the same city as you...

 

[Vesku]

Yeah, AMD didn't gain nearly the marketshare it deserved for price/performance of 4xxx and 5xxx series. Not sure why OEMs and many people seem to go out of their way to avoid great products.