AMD Q414 results

[sandorski]

That's highly debatable... hence the discussion almost 10 years later.

Not really. At this point it's just some peoples whipping horse.

 

[Arachnotronic]

Actually intel purchasing amd is an intriguing thought, but govt regulators would never let it happen. But it would be quite interesting to see AMD dgpus and apus manufactured on intels process nodes.

I'd love to see Intel buy either AMD or NVIDIA. Amazing dGPUs on Intel process + great iGPUs inside of Intel's CPUs? Sounds great.

 

[theeedude]

Why would any new buyer of AMD continue to make x86 chips?

How have the previous non-Intel makers of x86 chips done?

Think IBM, VIA, AMD.

What's more, are they going to spend 5 years to produce an x86 chip they hope will compete with Intel?

Try being rational, not emotional.

OK, rationally speaking, if you are Qualcomm and you buy AMD, you have an x86 chip right away, not in 5 years. Maybe it would take a year for Qualcomm engineers to spin it with their flows to fit into a phone envelope and integrate it with a modem, but even there, you aren't starting with a clean sheet 5 year design. In tablet and PC, you are ready to go.
Now you have a choice. You can build an ARM chip that only works on Android, or x86 chip that works on Android and also allows you to sell into the Windows ecosystem. Rationally, why would you pay for a license to make an ARM chip which would limit your TAM to very competitive Android SOC segment, if you could build an x86 chip that works for that segment and also with Windows, where you'd have a duopoly with a player accustomed to 50% monopoly gross margins?

 

[Techhog]

I really think you guys are jumping the gun saying that they'll be bankrupt within the next 1-2 years, especially since their death will result in two monopolies unless Qualcomm is considered a competitor in the desktop/laptop space somehow. Granted, I know some of you want AMD to die...

If they do die though, I'll have to quit PC gaming. $2000-3000 single GPU cards from Nvidia would not be fun.

 

[CHADBOGA]

OK, rationally speaking, if you are Qualcomm and you buy AMD, you have an x86 chip right away, not in 5 years.

You have a non-competitive x86 chip right away.

Maybe it would take a year for Qualcomm engineers to spin it with their flows to fit into a phone envelope and integrate it with a modem, but even there, you aren't starting with a clean sheet 5 year design.

LOL It is going to take a lot more than a year to get any AMD chip in a state that they would be desirable for a phone.

In tablet and PC, you are ready to go.

With the current AMD chips, you would be going nowhere.

Now you have a choice. You can build an ARM chip that only works on Android, or x86 chip that works on Android and also allows you to sell into the Windows ecosystem. Rationally, why would you pay for a license to make an ARM chip which would limit your TAM to very competitive Android SOC segment, if you could build an x86 chip that works for that segment and also with Windows, where you'd have a duopoly with a player accustomed to 50% monopoly gross margins?

Intel is the only one getting 50%+ gross margins amongst semiconductor companies in the Windows ecosystem.

Why would Qualcomm think they would do any better than IBM, VIA or AMD did in the Windows ecosystem?

 

[III-V]

Not really. At this point it's just some peoples whipping horse.

Again, the subject is highly debatable -- THAT, is what is not up to debate.

I see it the exact opposite way. They overpaid tremendously. And they only started making consistent profit recently (at least if we look from 2010 onwards, not an expert on their earlier history, but I do know they did not do well early on), after the 7970 launched and they decided to throw away their previous reputation as the more cost-friendly brand. They had to spin off their fabs, which at least at one point in time, were a key strength for the company.

Why would Qualcomm think they would do any better than IBM, VIA or AMD did in the Windows ecosystem?

Well, they've got the money. IBM did as well, of course, but Qualcomm could certainly be a strong opponent.

But yeah, them buying AMD would not suddenly polish all of the turds in AMD's pipeline.

 

[AnandThenMan]

They overpaid tremendously.

If you compare the ATI purchase to other acquisitions you'd see this is absolutely not true. Intel picked up McAfee for almost $8 billion. D: That was a 62% premium.

 

[blastingcap]

I'd love to see Intel buy either AMD or NVIDIA. Amazing dGPUs on Intel process + great iGPUs inside of Intel's CPUs? Sounds great.

Will you love the prices, too?

 

[Arachnotronic]

If you compare the ATI purchase to other acquisitions you'd see this is absolutely not true. Intel picked up McAfee for almost $8 billion. D: That was a 62% premium.

McAfee was a waste of money, but Intel could afford it. AMD could not.

 

[Arachnotronic]

Will you love the prices, too?

We live in a world where AMD is basically not a factor in CPUs, and yet you could buy the 4790K during Black Friday for $299. 5820K is now available for $389.

Intel needs to convince people to upgrade, and by making CPUs/GPUs prohibitively expensive, volumes drop, factories don't get utilized, and people shift more of their money to smartphones/tablets.

tl;dr -- prices wouldn't change much.

 

[AnandThenMan]

McAfee was a waste of money, but Intel could afford it. AMD could not.

AMD could not afford to be left behind, having only CPU tech would have doomed them. Imagine today all AMD had in their portfolio was products directly competing with Intel. I can't believe anyone thinks this is a good idea. :thumbsdown:

In fact AMD would not even be able to do that because most all Intel processors have a GPU.

 

[Arachnotronic]

AMD could not afford to be left behind, having only CPU tech would have doomed them. Imagine today all AMD had in their portfolio was products directly competing with Intel. I can't believe anyone thinks this is a good idea. :thumbsdown:

In fact AMD would not even be able to do that because most all Intel processors have a GPU.

AMD could have bought Imagination Technologies for a mere fraction of what it paid for ATI at the time. This would have not given it the dGPU revenue stream, but it would have given it the IP to integrate into the CPUs strategically.

Or, as mrmt suggested, AMD could have merged with ATI.

Heck, ARM Holdings now has some of the top GPU IP for mobile and it got into this market by acquiring a small IP vendor known as Falanx Microsystems.

No, if AMD needed GPUs for SoCs, it could have found a much cheaper way to get it.

 

[frozentundra123456]

If you compare the ATI purchase to other acquisitions you'd see this is absolutely not true. Intel picked up McAfee for almost $8 billion. D: That was a 62% premium.

Just because intel overpaid for something does not mean AMD did not as well. In any case you have to look at the cost in relation to the resources of the company. I dont know how much benefit intel has gotten from that purchase, but they certainly are much more able to absorb the cost than AMD was. AMD basically went "all in" on the ATI purchase.

It may surprise some here that I actually strongly defended the ATI purchase by AMD when it happened. And actually I was a big fan of AMD at that time. I could not wait to get a competent gaming computer off the shelf without a discrete gpu. Unfortunately, after all these years, the best APU is a very marginal solution. And the other possible compelling use case for an apu, HSA, has not really taken off either. It is really impossible to say whether AMD would have been better off without the ATI purchase, but if they had been able to keep their fabs, or at least negotiate a better agreement, i think it is a distinct possibility.

 

[cbn]

Intel needs to convince people to upgrade, and by making CPUs/GPUs prohibitively expensive, volumes drop, factories don't get utilized, and people shift more of their money to smartphones/tablets.

Yes, Intel can do this easily by not adding so much iGPU to their desktop quad core and even Core i3s. Then get the people using all the older chips to move up the more affordable cpus.

With that mentioned, I would hope the AMD would do the same thing. (and in the case of chips like AM3+ go one step further and reduce or eliminate the die size bloating L3 cache)

P.S. I find it amazing that over-clockable Core i5 desktop prices have done nothing but increase since the i5-750 was released. Of course, the cpus have gotten better, but most of the extra die size was used for adding iGPU....not cpu logic.

 

[theeedude]

You have a non-competitive x86 chip right away.
LOL It is going to take a lot more than a year to get any AMD chip in a state that they would be desirable for a phone.
With the current AMD chips, you would be going nowhere.
Intel is the only one getting 50%+ gross margins amongst semiconductor companies in the Windows ecosystem.
Why would Qualcomm think they would do any better than IBM, VIA or AMD did in the Windows ecosystem?

Because they could sell x86 into their current Android market as well. They could transition Krait to x86 and merge the chip lines to get economies of scale. Windows market share would be gravy. Different situation than AMD which has to pay the whole bill effectively.

 

[Shehriazad]

Is this the right place to place offers for AMD?


I offer 52 cents in pocket change and 1 wheel of luxurious cheese that was personally appraised by a cheese critic.

Can't really imagine anyone would offer more than me.

 

[AnandThenMan]

Is this the right place to place offers for AMD?


I offer 52 cents in pocket change and 1 wheel of luxurious cheese that was personally appraised by a cheese critic.

Can't really imagine anyone would offer more than me.

Seriously?

 

[gammaray]

Again, Debt didn't increase this quarter. Revenue came in at its projected levels. It just seem like another thread with a bunch of ***holes that are rooting for a total collapse that will never happen -- but love to talk about it month after month. AMD has shrunk it reliance on the PC industry. Will they make CPU's 5 years from now? I doubt it. They will likely continue to shrink the business and probably shutter divisions and layoff more people.... The company will look a lot different a decade from now. People have been predicting the death of Matrox, VIA and AMD for the past 20 years.... and none of them are probably going anywhere.

A lot of people have a short memory -- like ignoring 2007, When AMD was losing around $500 million every quarter. 2014 pales by comparison. "This is the Voice of Doom speaking! Special Bulletin! Flash! The Sky is Falling."

while what you say makes sense, you probably don't own any shares of AMD. From a shareholder pov, the Q4 is disastrous, period.

 

[CHADBOGA]

Because they could sell x86 into their current Android market as well. They could transition Krait to x86 and merge the chip lines to get economies of scale. Windows market share would be gravy. Different situation than AMD which has to pay the whole bill effectively.

Intel is struggling to gain a foothold in the Android market with x86, what makes you think a company that has never built an x86 processor and with a manufacturing deficit, would do better than Intel?

 

[erunion]

[Without the ATI purchase] AMD would be dead and buried.

That position is absurd.

Apple has turned licensed graphics IP into Billions and Billions. AMD's route cost them billions and billions that they couldn't afford.

 

[AnandThenMan]

That position is absurd.

Apple has turned licensed graphics IP into Billions and Billions. AMD's route cost them billions and billions that they couldn't afford.

Apple makes their money because they have the entire infrastructure, top to bottom. They have the software, the marketing, the product, the mindshare. Not to mention Apple has pioneered several technologies, consider the original iPhone it changed absolutely everything in the smartphone market. No way AMD given their cashflow could even begin to follow said model, you're being completely ridiculous.

 

[theeedude]

Intel is struggling to gain a foothold in the Android market with x86, what makes you think a company that has never built an x86 processor and with a manufacturing deficit, would do better than Intel?

If Qualcomm were to buy AMD, they would not be a company that never built an x86 processor anymore, since they would own one that has.
They don't need to do better than Intel on x86, since like I said, any x86 business would just be gravy to their already strong Android business that could be served by x86.
After TSMC and Samsung move to Fin-FET, Intel's advantage will not be as significant as it is now, and in many segments, Qualcomm could do quite well with a contract foundry. Being fabless has advantages too. They don't need to maintain volumes just to keep their fabs filled and amortize process R&D.
This allows them to opportunistically pursue x86 segments of their choosing with low financial risk.

Plus it's not like they would be betting the company on it. AMD has $3B enterprise value. Qualcomm has market cap of $120B, and $18B in cash.
They could buy AMD several times over without having to borrow or issue shares.

 

[CHADBOGA]

If Qualcomm were to buy AMD, they would not be a company that never built an x86 processor anymore, since they would own one that has.

But they would be buying a company whose x86 processors aren't remotely close to cracking the phone market, due to excess power consumption and lack of integration.

So if Qualcomm aren't going to bring in new talent to dramatically improve future chips, how will it be done?

They don't need to do better than Intel on x86, since like I said, any x86 business would just be gravy to their already strong Android business that could be served by x86.

Unless they can produce leading edge performance, performance/watt, there would be no reason for any phone maker to go with them, instead of a myriad of ARM choices or go with Intel.

There is no viable x86 market for Android right now, so what riches will there be for AMD-Qual to pick up?

After TSMC and Samsung move to Fin-FET, Intel's advantage will not be as significant as it is now, and in many segments

That remains to be seen.

Perhaps that will only be the case for the first 6 or so months after TSMC & Samsung move to Fin-FET.

Plus it's not like they would be betting the company on it. AMD has $3B enterprise value. Qualcomm has market cap of $120B, and $18B in cash.
They could buy AMD several times over without having to borrow or issue shares.

You keep forgetting the license issue.

It is all well and good to believe it might get resolved in favour of a new purchaser of AMD, but how long would it take before this was known?

Would Qualcomm or any other company try and buy AMD and do the massive investments that need to be done straight away, or would they take a far more cautious and long winded approach.

An approach that sees AMD fall even further behind and lose valuable engineering staff who don't like the uncertainty about whether it is even viable for someone to purchase AMD.

 

[theeedude]

Licensing is a non-issue.
Do you think if Qualcomm buys AMD, Intel will sue to get AMD chips off store shelves because Qualcomm owned AMD is no longer licensed?
And you don't think they are going to get raked over anti-trust coals for doing it?
Microsoft had to pay $500M to EU just for including Media Player in Windows. Qualcomm itself is about to pay a 10 figure fine in China for not licensing LTE on favorable enough terms for Chinese equipment and chip makers. Intel trying to get the main x86 competitor banned would dwarf these.
You really think anyone at Intel management is stupid enough to try to force this issue?

 

[CHADBOGA]

Licensing is a non-issue.
Do you think if Qualcomm buys AMD, Intel will sue to get AMD chips off store shelves because Qualcomm owned AMD is no longer licensed?
And you don't think they are going to get raked over anti-trust coals for doing it?
Microsoft had to pay $500M to EU just for including Media Player in Windows. Qualcomm itself is about to pay a 10 figure fine in China for not licensing LTE on favorable enough terms for Chinese equipment and chip makers. Intel trying to get the main x86 competitor banned would dwarf these.
You really think anyone at Intel management is stupid enough to try to force this issue?

Well leaving aside that I don't think Qualcomm is going to buy AMD, it is somewhat fanciful to be declaring that an existing contract will be rendered null & void, simply because you would like it to be.