AMD Reports $611 Million Loss

mpfife writes "Toms Hardware reports that declining microprocessor sales have pushed AMD deeply into the red. 'The company reported a net loss of $611 million on revenues of $1.233 billion, which is more than 20% below the guidance the company expected at the end of Q4 2006. The loss includes charges related to the ATI acquisition in the amount of $113 million, but is mainly a result of the increasing competition with Intel in the microprocessor market.'"

Comments

Let's hope they recover

[thanatos_x]

Regardless of your feelings on the Intel/AMD processors, I don't think any one of us wants to envision a world with only Intel making x86 processors. Don't get me wrong, they're doing an excellent job, but just how much of this recent surge was a result of the increased competition from AMD?

Re:Let's hope they recover

[whoever57]

If Intel is making good processors at an affordable price, why would I not want that? Monopolies aren't inherently bad.

Because if Intel does establish a monopoly, prices will increase and innovation will decrease. There is no monopoly today, so Intel's behavior today is not relevant to how Intel might behave if AMD went out of business.

Re:Let's hope they recover

[Threni]

> If they rose prices too high, a competitor could undercut them.

Er..yes, but then they wouldn't have a monopoly.

> If they did anything illegal, the government would go after them

Yeah, the government will save us, right after they save the people who lost their money with Enron.

Re:Let's hope they recover

[Embedded2004]

The barrier to entry into the x86 processor market is *extremely* high.

If AMD folds, Intel could pretty much do anything they want to with prices.

It would be bad for everyone other than Intel employees and Intel share holders.

Re:Let's hope they recover

[whoever57]

You don't know that's true. Intel would raise prices as far as it wouldn't decrease sales. If they rose prices too high, a competitor could undercut them.

The cost of entry for developing processors is huge. Undercutting them is not so simple. A logical approach would be for Intel to raise prices, but be prepared to drop them if it looked like a competitor were about to enter the market (the lead time, investment and people required are such that any competitor in x86 could not keep a market entry secret for very long). Rational behavior from Intel would be to raise prices and spend less on R&D. Exactly how much that would impact processor prices is hard to gauge (but see my last comment). Price rises would be inevitable, the only question is how much.

If they did anything illegal, the government would go after them.

Based on what we have seen from the current administration, a slap on the wrist would be applied -- but probably insufficient to wipe out excessive monopoly profits, thus a rational company would attempt to create and leverage a monopoly. In any case, given reasonable increase in prices (say 50%), it would be difficult to prove an illegal monopoly at work.

It's not my job to worry what someone might do in some hypothetical situation just because AMD is floundering around right now.

Your job: no. Might you suffer (by paying more) if AMD stops making processors: probably.

Re:Will this lead to Inte monopoly again?

[Tablizer]

I wonder if AMD will loose the competition to Intel all together.
Do we risk going back to having only one big CPU producer?

Not if investers are smart. Duopolies are the next best thing to having a monopoly, meaning it has fat profit margins. However, if it is truely a business that requires economies of scale, then if AMD shrinks down past a certain size, it could risk being left out in the cold. I think this is just a temporary blurp. No need to worry yet. Tech is cyclical, including chips.

Re:Big AMD Fan here

[supabeast!]

But I have to ask, while AMD were on top with the Athlon for several years - were they just sitting on their laurels? Maybe. I think that it's a little more likely that intel finally realized there weren't enough marketing gimmicks in the world to beat a better product at a better price, and shifted some dollars back over to killer engineering. Then along came Core Duo...

Only made 2/3rds their costs

[jhfry]

net loss of $611 million on revenues of $1.233 billion So essentially, they needed 50% more revenue to break even. Numbers like that are expected in an early startup... but for an established business that's just horrible!

I realize that numbers like this can be misleading, especially when they are doing some significant spending (new fabs, acquiring new companies, filing patents, etc.), however an investor never likes to see a company fail to break even by such a huge margin.

I really hope that whatever AMD has in the works pans out... if they stay behind the technology curve for long after spending like they have recently, they may be stuck waiting for Intel to stop innovating so actively again.

Re:Only made 2/3rds their costs

[fiendy]

You just made the most incredibly generalized financial analysis I've ever read. 2/3 their costs? Do you have any idea about fixed vs. variable costs?

I can't believe you got modded up, since its clear that you didn't even bother to take a look at their financials. Impressive though, if you can come to any kind of understanding of a business with two lines of information - profit and revenue. I just hope you're only investing your money.

Re:Big AMD Fan here

[TheRaven64]

AMD haven't been able to compete with Intel for a long time. Building a new-process fab is an incredibly expensive undertaking. Intel can afford to do it because they own the majority of the desktop and laptop markets. AMD can't. Fortunately for AMD, neither can IBM. The reason I say fortunately is because, while AMD or IBM could not easily make the capital investment required, AMD and IBM can between them, and have for some years. IBM don't make x86 chips, and don't really compete in the same space as AMD (except a small part of the server market that could go Opteron or POWER), so they make a good partner.

IBM have some very high volumes parts (some mobile chips, the CPU in every new console, etc), but they can't compete with Intel in terms of investment in the semiconductor market. If anything happened to AMD, then IBM would have some serious problems. The only way out would be to dramatically increase the sales of PowerPC chips. They might be able to do this using open source - sell appliance-type systems where the user doesn't need to know what OS or CPU is running - but it's a gamble.

AMD/Intel is about controlling the hardware future

[dtjohnson]

The AMD/Intel dogfight is about way more than x86 market share...it's about the future of the hardware platform. Intel has always been restrained by competitors who will offer us a user-friendly alternative to whatever Intel and Microsoft are dreaming up. When Intel and Microsoft were pushing the CPUID, AMD refused to go along and Microsoft had to make do with a hardware profile they whip up from the onboard devices and serial numbers. If it was not for AMD, every web site you visit today would be able to read your cpu serial number and log your machine in as a unique visitor. Instead of the RIAA grabbing IP addresses and attempting to identify the user with some cumbersome legal process, they would just log your cpuid and subpoena the corresponding machine. Microsoft is still working to that end with whatever tools they can and they know that they need amd and intel completely and irrevocably in bed with them which they know cannot happen when amd and intel are still bitter competitors. So Microsoft has never done anything to help AMD and hopes that AMD is finally sinking for good.

Re:Big AMD Fan here

[gr8_phk]

But I have to ask, while AMD were on top with the Athlon for several years - were they just sitting on their laurels?

Yes and No. That's around the time Hector the Sector Director took over. That's what our FreeScale rep told me he was called back at Motorola :-) He was not missed. From what I saw, he seemed to focus on marketing followed by fab building at AMD. The Hammer architecture was already a work in progress when he took over, so he can't be given credit for that. Modern AMD chips are still small refinements to that same design. Even the new K8L or whatever they call it is going to be mostly the same - they are going to put in a full SSE unit instead of 1/2 of one that they've been getting away with for so long. IMHO they waited too long for that change probably because someone was trying to optimize dollars and didn't notice Intel was about to make a large performance gain. Oops. Anyway, they rested on their design laurels for 6 years now and are still doing so. Maybe they just plain don't see how to ramp core performance (per dollar?) in a significant way beyond K8L. Performance per watt (or per dollar) is a good measure, but they must not forget that the numerator is still performance.

Re:AMD: Try listening to your customers

[gosand]

AMD was founded by Jerry Sanders, a high-flying salesman originally from Intel who never quite fitted in. In Andy's Grove's Bio of Intel, he describes Sanders as fast and loose and the AMD corporate culture akin to a Las Vegas Casino: Very extravagant and over the top. Nevertheless, AMD did produce some killer products which at the time made life hard for Intel.

My best friend works at Intel as an engineering manager, and he's been there for 9 years. He has told me many stories about how people at Intel took AMD *VERY* seriously when they burst onto the scene and started taking away their business. He also said that he talked to some people who said they weren't worried about AMD, and gave a list of reasons. One of them was that they were brash and stupid, and blew TONS of money on ridiculous stuff. Another more important reason was simply the fabs. Intel kills AMD in fabs, and that is why Intel is on top again. AMD gave them something to think about, and Intel did just that - they replied with a very measured response. According to my friend, the higher-ups at Intel are thankful to AMD for giving them a run.... because it made them come up with a better answer and it made Intel a better company.

Re:Big AMD Fan here

[mollymoo]

But I have to ask, while AMD were on top with the Athlon for several years - were they just sitting on their laurels?

Do you mean the 32-bit Athlon? Around that time, AMD were developing x86-64 while Intel were developing Itanium/Itanic. AMD were first to market with a 64-bit CPU normal people actually wanted; Intel's 64-bit offering was a hideous beast and they sold exactly twenty-nine of them. The P4s of the time were hot and slow, the Athlon-64s and Opterons were much nicer. But Intel came back strongly, improving the P4, adopting x86-64 and getting ahead in the multi-core race. AMD just couldn't keep up.

Even when the Athlon was on top in terms of performance, they didn't sell nearly as many as Intel sold P3s and P4s.

Re:Slashdot is a little biased reporting this

[Anonymous Coward]

Although it does not appear to be updated anymore, you might want to be aware of this section:

http://amd.vendors.slashdot.org/

If AMD *EVER* wants to get ahead again...

[Khyber]

They better take a SERIOUS fucking look at their L2 cache sizes.

I see Core 2 Duos with 2 megs per fucking core.

I see Turion 64 X2's with a paltry 256K.

That's just the LAPTOP end.

Hey, AMD, wonder why you're not going any fucking where, even though you've had a superior bus?

Remember the Pentium D (Basically a hyped up pentium 3 with 2 megs of L2 cache) that smoked many higher-end Pentium 4s in gaming?

Pay attention! My 640K AMD64 3000+ could be smoking many other machines if it just had a DECENT CACHE ON-DIE!

Re:If AMD *EVER* wants to get ahead again...

[evilviper]

I see Core 2 Duos with 2 megs per fucking core.

I see Turion 64 X2's with a paltry 256K.

And I see Core 2 Duos with 1MB L2 cache, compared with Turions with 512K per core... You're just taking the worst-case example, and complaining about it as if it's typical.

Not to mention that Turion X2s have 128K L1 cache, while Core 2 Duos have a paltry 64K of L1. L1 is much more significant than L2.

What's more, L2 cache isn't magic, anyhow. According to benchmarks, the difference between 2MB L2 cache, and 4MB L2 cache, makes AT VERY BEST less than 10% of a performance improvement. http://www.bit-tech.net/hardware/2006/07/14/intel_ core_2_duo_processors/5.html

That's just the LAPTOP end.

Actually, it isn't. Core 2 CPUs are Intel's desktop CPUs as well. AMD, OTOH, has a different line of CPUs for their desktops, with, among other things, typically 1MB of cache (in your words) "per fucking core."

Remember the Pentium D (Basically a hyped up pentium 3 with 2 megs of L2 cache) that smoked many higher-end Pentium 4s in gaming?

No, I don't remember that at all. the Pentium D is the euphemism for a Pentium 4, that they've used just in the past few months now.

Pay attention! My 640K AMD64 3000+ could be smoking many other machines if it just had a DECENT CACHE ON-DIE!

People are supposed to accept your theory, because you've shown how you know absolutely nothing about processors? I'll pass. AMD can figure out how to make fast CPUs without your "help." They've just been caught napping, and need time to catch up.

Re:Big AMD Fan here

[DohnJoe]

please, it's not FUD (FUD is a marketing strategy) and he's just pointing out that the IA-64 architecture didn't sell so well, of course you shouldn't take the 29 seriously.... thick as a brick...

Re:Big AMD Fan here

[cheezedawg]

A lot of this is directly AMD's fault. Remember the big AMD/UMC deal back in 2002? AMD was so excited because now they wouldn't have to build more expensive Fabs, so as a result, they didn't. Then the deal fell through, and AMD was left scrambling to make up for their years of anemic manufacturing investment as a result of this deal. You cannot blame that kind of mismanagement on the competition.