• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

AMD Reports 2015 Third Quarter Results

I always ask myself this rhetorical question: What if AMD, instead of buying out ATI, invested that money in improving their fabrication process, would AMD be in a much stronger position today?
What would have happened if AMD had just licensed ATI's IP (kept their $2.3bn in cash and not taken out $2bn in loans), and extended their original relationship with UMC ? UMC had Xilinx as their main customer at the time and were in desperate need of a large prestige customer since IBM and SMIC were gearing up to snatch away UMC's market share, and UMC did have 10 fabs at the time- including two new 12" wafer fabs (one a JV with Infineon/UMCi). Instead AMD overextended itself with the Dresden money pit that sucked their cash reserves dry.

What might have been...

Dreaded auto analogy: AMD as Formula 1 team - IP (acquired and developed) is the engine, Engineers are the drivetrain, Stevie Wonder (AMD BoD) is the driver.
 
Last edited:
How much do you think AMD is worth? Their market cap is the consensus view of the market. However, it is incredibly easy to prick that bubble, as the market is being stupidly pessimistic. I'll illustrate how. Let's say, since we're speculating, that AMD's Radeon graphics division is worth just HALF of what nVidia is worth. After all, they're both world-leading GPU design companies, and neither one of them is dramatically ahead of the other in terms of the key technologies, so you'd have to be really silly to argue that the Radeon division is worth LESS than half of nVidia. Suffice to say, I'm being incredibly conservative using just 50%. So 50% of nVidia's current market cap is $6 billion. Now, let's look at how much AMD's CPU division might be worth. I'll arbitrarily assign a value of $4 billion, as the x86, x86-64 technologies have got to be potentially very valuable IP to a company like Apple or Microsoft or Qualcomm, who might like to be able to design their own x86 processors. So we end up with a figure of ~$10 billion value for AMD. Subtracting the ~$2.2 billion of current debt, we're left with a figure of ~$7.8 billion. At this market cap, this would mean a share price of:

Its not that simple for multiple reasons:

1. GPU. Nobody wants a business which does not produces money. Why would you get into it? If they manage to spin off the GPU division and prove that the GPU division produces money, then yes its worth something maybe even more than 50% of the competitor, otherwise its worth nothing, absolute 0.

2. x86 license. There were long discussions about it but I think its not transferable. Even if it would be ... who is crazy enough to invest tens of billions to fight with Intel? And why would you do that, the lion share is in the ARM technology nowadays, so if Apple/Qualcomm will buy something from AMD related to CPU is patents or maybe some know how, but the x86 license itself is not worth much.

3. Who will want to do business and strategic partnership with a company which tomorrow might go bankrupt? Typical enterprise server clients want support 5-10 years at least. Typical OEM also want support for the products with AMD tech which they sell to their customers and they want to know, that products will be supplied and contracts fulfilled. If an OEM will take the risk it will want to be worth it, which means AMD has to offer very agresive prices which will lead to less profit for AMD.

4. General perception of the AMD products is that they are not very good, neither the CPU, nor the GPU and it will be very hard fighting against that. Market share also tanked, lower than 25 % for both, which means 75% of the people choose and will recommend something else. Basically now people buy AMD only to support the underdog, or because the price is very low for what it offers, which means they have to sell cheap, and again not make profit. Also because of this shops will bring in what they can sell, they don't want to get stuck with stocks they cannot sell. As a result also manufacturers are producing goods people want as they also don't want stocks they cannot sell. Its a tough loop to get out from. And to prove my point, just look in your favorite online shop. I just checked mine, there are about 200 laptop models in their direct stock, only 6 of them with AMD cpu (3%). On the GPU side things are a bit better 40 have AMD GPU (20%)

I think, if someone BIG will backup AMD no matter what and they will be able to prove that their business is strong, if they will be extremely customer oriented and will jump immediately to resolve even the smallest complain they will be in a much better position.
 
The issue is and always will be Intel "cheats" still to this day walk into a dell store and tell me what logo sits under dells. There are very few companies that use AMD for more reasons than
It was a HORRIBLE move! Not because they did it. Buying ATI on it's face was good, it was the hugely overpriced purchase that was a bad idea. Combined with generally bad ideas on running a business, they still have that debt around their neck like a millstone while they attmpt to tread water.

I don't disagree they paid to much. The acquisition of ATi is all that is keeping AMD afloat(ish) right now. If the APU's didn't exist AMD wouldn't be in consoles and as we speak consoles are making more money for AMD than the PC market...
 
If you like throwing money away I will take some. Thanks!

What the heck are you talking about? Fury X is within or surpassing 980 TI by 5% in every game out there sans Project Cars.
 
Well that only took 2 pages to go from analytical discussion to green vs red.
 
Dreaded auto analogy: AMD as Formula 1 team - IP (acquired and developed) is the engine, Engineers are the drivetrain, Stevie Wonder (AMD BoD) is the driver.

Yeah but the music would have been awesome.

Baow baow baow baow--- baowm baowm baowm baowm, ...Baow baow baow baow--- baowm baowm baowm baowm , do do superstition -- do-do do do-do dooooooo

What the heck are you talking about? Fury X is within or surpassing 980 TI by 5% in every game out there sans Project Cars.

No, sorry, my card runs 20% faster than that (1500Mhz, it's a 980ti thing).
 
Except for the fact that it doesn't. It can't even match up with a vanilla ti

http://www.techpowerup.com/reviews/AMD/R9_Fury_X/19.html


GTA V? Now I know what side you are on. How about you use a proper comparison graph?!

perfrel_2560.gif
 
GTA V? Now I know what side you are on. How about you use a proper comparison graph?!

perfrel_2560.gif

Still the vanilla ti is ahead by over 5% if it was up vs lets say the G1 ti we would be looking at 15% easily put another 10-15% because of the OC a card like the G1 can do.
 
1) Crossfire scales better than SLI when it works

2) All 4 cards will be running volt mods and unlike Nvidia AMD cards scale with voltage even on ambient cooling. 24/7 clocks will be somewhere between 1100 and 1200mhz but for benching I can probably push 1300mhz and then I can crush the 4 way firestrike leader boards on HWbot.

3) Freesync is cheaper than Gsync

4) I hate Nvidia drivers

5) 4 way pretty much requires water cooling to run at a reasonable temp and the cheapest water cooled 980 Ti is more expensive than a Fury X when you buy 4 cards this small price difference stacks.
 
Pretty sure it was Michael Schumacher
Schumacher%2004%20Monza%20Hans.Ext%20chin%20FG%20lft.jpg

No fair. Full respect to F1's greatest driver (outside perhaps Ayrton Senna) - you cannot compare these giants in their field with AMD's frankly disastrous BoD.

This is always AMD's issue - great (fantastic even) software and hardware engineers but with the wrong people driving the wagon. And don't confuse shrewd business practices (Nvidia or Intel) with decency but fact is, AMD's CEO's for years have been lacking in 'prowess'.
 
No fair. Full respect to F1's greatest driver (outside perhaps Ayrton Senna) - you cannot compare these giants in their field with AMD's frankly disastrous BoD.

This is always AMD's issue - great (fantastic even) software and hardware engineers but with the wrong people driving the wagon. And don't confuse shrewd business practices (Nvidia or Intel) with decency but fact is, AMD's CEO's for years have been lacking in 'prowess'.

No fair..?

I wasn't try'n to be fair. I was just pointing to the fact that AMD sponsored Schumi. I wasn't adding to his analogy. His reference to F1 and AMD reminded me of Schumi. You guys seem to be filling in the rest.

As far as fair goes.

Didn't you just finish comparing your water blocked KINGPIN to a reference debate between the double D's.

No, sorry, my card runs 20% faster than that (1500Mhz, it's a 980ti thing).

:lovetpu:
 
Last edited:
Pretty sure it was Michael Schumacher
Schumacher%2004%20Monza%20Hans.Ext%20chin%20FG%20lft.jpg
That picture is perfect. Red, Green, AMD stuck in the middle getting lung cancer off of Marlboro cigarettes. XD
 
Remember when people right here on this very forum stated that AMD buying ATi was a bad decision? Pepperidge Farms remembers.

Classic Rhino.
 
AMD did indeed overpay for ATI. I think that was when AMD mostly set into motion their own demise. At the time of the purchase (July 24th 2006) AMD was trading for $19.10 a share with a market cap of $14.9 billion dollars. A year later they were down to $13.87 a share and a market cap of $10.8 billion dollars. A year after that they plummeted to $4.36 a share and a market cap of $3.4 billion dollars. They have had some ups and downs through the years but the general trend has been downward. Today AMD is trading for $1.91 a share with a market cap of about 1.49 billion dollars.

They paid $5.4 billion dollars to buy ATI and today the entire company is worth a little over 1/4 of what they paid for ATI.
Its not that simple for multiple reasons:

1. GPU. Nobody wants a business which does not produces money. Why would you get into it? If they manage to spin off the GPU division and prove that the GPU division produces money, then yes its worth something maybe even more than 50% of the competitor, otherwise its worth nothing, absolute 0.

Utter balderdash. Of course the Radeon division produces money. It's not even debatable. The issue is not over Radeon making money, it's always about AMD's CPUs, which really are not that bad, they were just designed for Windows 10/DX12 4 years too early.

2. x86 license. There were long discussions about it but I think its not transferable. Even if it would be ... who is crazy enough to invest tens of billions to fight with Intel? And why would you do that, the lion share is in the ARM technology nowadays, so if Apple/Qualcomm will buy something from AMD related to CPU is patents or maybe some know how, but the x86 license itself is not worth much.

Again, utter nonsense. The x86 license IS transferable, and that's the final word on the matter from AMD's own CFO, Devinder Kumar: http://www.kitguru.net/components/g...intel-cannot-block-our-merger-or-acquisition/

Essentially, what's going on is that AMD's x86-64 is built into every 64 bit Intel CPU. If you don't believe me, and you're running a 64 bit version of any Windows OS, go to your "C:\Windows\WinSxS" folder and have a look at the folder names in there. Whether you're running an AMD or Intel CPU, those folders are still labelled: "amd64_*" Why? Because the 64bit extensions INSIDE your Intel CPU were written by... Advanced Micro Devices. Really, the truthful Intel marketing slogan would be "AMD Inside!" (every one of our CPUs!)

If Intel were to try to refuse an x86 license in the event of a merger/buyout, AMD can refuse to renew the x86-64 cross-license with Intel, too. AMD could get a court injunction against the sale of any Intel CPUs with their micro-code, which means essentially all of them. Do you STILL think Intel would simply refuse to negotiate the x86 license with AMD or an acquirer of AMD's intellectual property?

3. Who will want to do business and strategic partnership with a company which tomorrow might go bankrupt? Typical enterprise server clients want support 5-10 years at least. Typical OEM also want support for the products with AMD tech which they sell to their customers and they want to know, that products will be supplied and contracts fulfilled. If an OEM will take the risk it will want to be worth it, which means AMD has to offer very agresive prices which will lead to less profit for AMD.

Again, this is a spurious argument. If AMD, with its universally acknowledged limited resources, has been in business for 45 years, and been making x86-compatible CPUs that are competitive with Intel since 1982, do you REALLY think a company with the financial resources like Microsoft ($92 billion in cash on hand), Apple (~$200 billion in cash on hand) or even Qualcomm (~$35 billion in cash on hand) will be daunted in an effort to acquire AMD and inject enormous investments into their x86 business? "Hey, there's only ONE competitor in this extremely lucrative computing market! I'll take it!" would be the more likely response.

4. General perception of the AMD products is that they are not very good, neither the CPU, nor the GPU and it will be very hard fighting against that. Market share also tanked, lower than 25 % for both, which means 75% of the people choose and will recommend something else. Basically now people buy AMD only to support the underdog, or because the price is very low for what it offers, which means they have to sell cheap, and again not make profit. Also because of this shops will bring in what they can sell, they don't want to get stuck with stocks they cannot sell. As a result also manufacturers are producing goods people want as they also don't want stocks they cannot sell. Its a tough loop to get out from. And to prove my point, just look in your favorite online shop. I just checked mine, there are about 200 laptop models in their direct stock, only 6 of them with AMD cpu (3%). On the GPU side things are a bit better 40 have AMD GPU (20%)

This is the silliest statement you've made so far. So who's perceptions are you talking about, exactly? You and a few friends? Because Sony and Microsoft are currently all-AMD inside for their high volume game consoles. There's no Intel or nVidia in there, period. And Hewlett Packard's, Dell's, and IBM's (now Lenovo) server units were all building AMD-powered servers for their customers, and selling them quite handily back when AMD had the Athlon 64-based Opterons. Seems to me that these companies and their customers didn't think AMD products were 'not very good'! And all that has to happen for history to repeat itself is for the Zen processor to start coming out of the Fab, and be good. And it will be. Intelligent people will look at benchmarks, and if AMD's chips are faster than Intel, or even pretty darn close, AMD will be able to charge the same price or more than Intel charges for their competing chip. This was proven back in 2002-2005, when Opterons were selling for $1,000+/chip, and AMD had over 10% of the server market:

http://www.theregister.co.uk/2005/07/25/amd_ten_percent/

so please don't bother to argue the point.

I think, if someone BIG will backup AMD no matter what and they will be able to prove that their business is strong, if they will be extremely customer oriented and will jump immediately to resolve even the smallest complain they will be in a much better position.

Finally, you said something I can agree with. However, I happen to think AMD could pull this off, even without 'someone BIG' buying them out. As an AMD shareholder right now, however, such a buyout, at the price per share I projected, would be icing on the cake.
 
Last edited:
No fair..?

I wasn't try'n to be fair. I was just pointing to the fact that AMD sponsored Schumi. I wasn't adding to his analogy. His reference to F1 and AMD reminded me of Schumi. You guys seem to be filling in the rest.

As far as fair goes.

Didn't you just finish comparing your water blocked KINGPIN to a reference debate between the double D's.



:lovetpu:

Ah, I see. I recall the AMD sponsorship.

And your fairness retort to me? I was comparing most 980ti's to be honest. It's only fair to compare the overclock on Maxwell to the 'overclockers dream'.
Which is relevant to AMD's current situation where the company promises too much and delivers under expectation.

As for referencing my Kingpin, you should know they don't overclock any better than most 980ti cards unless using LN2. But it is so much prettier.

I love TPU too.
 
Ah, I see. I recall the AMD sponsorship.

And your fairness retort to me? I was comparing most 980ti's to be honest. It's only fair to compare the overclock on Maxwell to the 'overclockers dream'.
Which is relevant to AMD's current situation where the company promises too much and delivers under expectation.

As for referencing my Kingpin, you should know they don't overclock any better than most 980ti cards unless using LN2. But it is so much prettier.

I love TPU too.

I found it funny from a hypocritical standpoint. Two guys having a spat over reference numbers and a third inserts himself on a AIB OC while accusing me of not being fair for just linking a picture of Schumis helmet and a minute later thanks RT for pointing out the thread is turning into green vs red.

Now comparing overclock ability is relevant.

All because of Schumis helmet. :laugh:

Schumacher-2005-helmet-F1.4.jpg
 
Last edited:
Of course the Radeon division produces money.

Its nice idea, probably true, but I'd like to see that. As of now they need HBM and double the bus width to compete with nvidia, and in general significantly higher number of transistors (see 960 vs 380, 3B vs 5B)

The x86 license IS transferable

What you say is correct, probably it will be renewed due to a stalemate. However I still question the viability of investing in x86 as a technology for Qualcomm or Apple, given that ARM can probably provide the same level of performance, if one will make a full blown desktop CPU. And at least for Apple backward compatibility is not an issue, linux/unix systems are running happily on ARM since ages. Obviously users will need to download a new build from the App Store, but it shouldn't be that hard. Microsoft on the other hand is quite dependent for now and they might invest in it just to keep their windows business up. Who knows ... we will see. My concern is that if someone wants to get into this business, they could just go to Via and buy the license, and then invest like 1B in AMD to get the extension and that's it they don't need to pay 5B on a dying AMD.

AMD, with its universally acknowledged limited resources, has been in business for 45 years,

Things have never been so bad for AMD. In the past, yes they were the underdog, had limited resources, but there was ton of good will, their products were competitive from price/performance perspective, whole market was growing like crazy. Not to mention that they were not so limited. They had a FAB, they were a real CPU manufacturer. And the example about opteron that you gave confirms what I said. At the time they had good product, strong perspective, profitability, obviously companies wanted to do business with them.

So who's perceptions are you talking about, exactly?

General perception. See reviews, see forums, not necessarily mine.
In fact I'm a big AMD fan and the Haswell that i now have is the first Intel CPU I have in my desktop since 98 when i bought my first x86 computer (had Z80 before).
Even this one I wouldn't have as despite the bad press, I initially bought the FX 8300, in very good price, overclocked it to 8350 levels and returned it the next day, due to low performance.
Single thread was abysmal, lower than my old phenom 2, FPU abysmal, lower than my 4 years old phenom 2. Multi-thread performance was great, but I couldn't accept that some tasks will work poorly on the new CPU. Intel gave me a good 40% flat performance increase both in single and multi-thread.

On the GPU side, I still went AMD, again against popular opinion, which was placing competing nvidia product with better time-frames, better power consumption and so on.
I did it partly to support them, partly because I couldn't understand how nv can work so well with the small bus, I was sure there is a trade off somewhere which nobody saw so far. And guess what, I'm a bit disappointed with this as well. I did not return it this time, as I got my 30% increase in all applications but ... I have a problem with the VRMs which most of the time go to 90 - 100 degrees C in gaming, 125+throtling in furmark. I contacted Asus and they told me its perfectly fine ... don't worry, just don't use furmark for long periods. This is not necessarily AMD fault, but a poor design on Asus side. VRMs should have touched a heat pipe from the big cooler. From another perspective this is a side effect of the fact that AMD GPUs consume a lot of power.
 
How much do you think AMD is worth? Their market cap is the consensus view of the market. However, it is incredibly easy to prick that bubble, as the market is being stupidly pessimistic. I'll illustrate how. Let's say, since we're speculating, that AMD's Radeon graphics division is worth just HALF of what nVidia is worth. After all, they're both world-leading GPU design companies, and neither one of them is dramatically ahead of the other in terms of the key technologies, so you'd have to be really silly to argue that the Radeon division is worth LESS than half of nVidia. Suffice to say, I'm being incredibly conservative using just 50%. So 50% of nVidia's current market cap is $6 billion. Now, let's look at how much AMD's CPU division might be worth. I'll arbitrarily assign a value of $4 billion, as the x86, x86-64 technologies have got to be potentially very valuable IP to a company like Apple or Microsoft or Qualcomm, who might like to be able to design their own x86 processors. So we end up with a figure of ~$10 billion value for AMD. Subtracting the ~$2.2 billion of current debt, we're left with a figure of ~$7.8 billion. At this market cap, this would mean a share price of:

$7,800,000,000 / 779,180,000 shares outstanding = $10/share. That means that AMD, in a buyout scenario, would likely be worth about $10/share. Currently it's trading ~$2.00/share. The reason for this is that most investors are waiting to see if the Zen processor will be good. Pretty soon, engineering samples will be benchmarked and those benchmarks will leak. Once we see these benchmarks, we will either see AMD suddenly appreciate in value (if they're good), or, we'll see AMD's share price remain in a holding pattern approximately where it is now. At some point, any other company that is interested in buying them will have to consider doing it soon, as the x86 license/x86-64 cross-license agreement with Intel will be extinguished by bankruptcy. This means that anyone interested in buying it had better do it sooner rather than later.

I can't understand how you did this calculation but let us redo every thing again.
1- You can't assume that ATi/Radeon worth 50% of Nvidia because Radeon's share is 20% and Nvidia's is 80% that mean 1 to 4. Second Nvidia has 75% of severs GPGPU and 24% for intel. Third Nvidia controlling most of workstation graphics. Fourth Nvidia has about 70% of computer graphics IP. Fifth Nvidia has real estates like their new campus which cost them about 340 M$. Sixth Nvidia has small projects like ARM, Tablets, Settop boxes, Power Foundation, Cars information systems. Seventh Nvidia's big R&D team. Eighth History of stable profit in the last five years.
2- You can't put value for company more than what its assets worth because that is what you get when liquidize the company and I couldn't find AMD's brand value, The x86 IP doesn't worth that much because: First the licensing agreement with Intel is to tight, Second the PC market shrink, Third the big variety of CPU IP's other than x86 like ARM, MIPS, Power PC and the ease to get licence for any of these IP's.
I can tell you what is the most valuable asset AMD has. It is their loyal fans who kept defending the brand for the last five years and kept buying their products, but AMD forgot about that and kept wasting their money on PR monkeys and over paid management, they loose key R&D people every day. with out AMD loyal fans AMD worth -2.2 B$.
 
I found it funny from a hypocritical standpoint. Two guys having a spat over reference numbers and a third inserts himself on a AIB OC while accusing me of not being fair for just linking a picture of Schumis helmet and a minute later thanks RT for pointing out the thread is turning into green vs red.

Now comparing overclock ability is relevant.

All because of Schumis helmet. :laugh:

Schumacher-2005-helmet-F1.4.jpg

Says the guy whose post hasn't mentioned AMD's financial statement.

FTR, I merely indicated how far ahead the 980ti actually is (yes with overclock). Which is really relevant to the thread because had Lisa Su been correct and the Fury X was an overclockers dream, I'd own one right now.

I waited for the Fury X release, baited on false promises. As good as it is, it was not what they said it would be.

All relevant to the thread topic of falling earnings due to underperforming (compared to stated marketing or conferences) products.

Even going off topic, if they had released the Fury X with good air cooler and then Nano, it would have made a more desirable stack. And they gambled too much on component availability.

So yeah, comparing the merits of the abilities (over clocking) of the Fury X's rival is actually pretty relevant in a thread about diminishing revenue.

Why do you feel AMD is suffering? What factors do you think play a role?

As far as Schumacher's helmet goes, my initial response was understandable from a logical point of view. @HumanSmoke compares BoD to blind guy driving but you reply, more like Schumacher driving. I miss your AMD meaning and think instead you feel BoD are aces like Schumacher. QED.

Edit: Obviously the gfx wing is only part of the problem...
 
Last edited:
Back
Top