Careless talk costs wives
|
|
|
How AMD messed up Christmas
THE WAR in the IT industry is better than the real thing!
You only have to see the details between the piles of information swirling all around you and have a teeny weeny drop of imagination. I tell you, it's fantastic.
And you probably know exactly what I mean. Let’s see. We have alliances, empires growing, shattering, marriages of convenience, betrayals, assassinations and of course, weapons of mass destruction. Do you know what today’s weapon of mass destruction is? Of course you do. You wouldn't be reading this if you didn't call yourself an IT-ophile - at least in secret, when nobody hears you. Well? Introducing today’s weapon of mass destruction… is the Core 2 microarchitecture from Intel!
Nay, I am no Chipzilla fan. Actually, AMD is painted on my heart... with green I might add. But a fan-boy I am not. I'm a technical guy just like all other technical guys.
I spawn zeroes and ones for a living and I got silicon running through my veins. So like any other open minded individual with good knowledge of what’s going on in the hardware world, I don’t need a fortune teller to know that things are (and will be) going magnificently for Intel while AMD seems to be (and continue) on a downward spiral.
There’s war between these two and there has been war for almost 40 years, a war with casualties, dramas and all the suspense you’ll ever need. Of course, only geeks like most of us can truly savour this confrontation (hey, no offence). But the confrontation is dangerously tilting in the favor of one of the competitors, and while all of us now reap the benefits of the punching between these two beasts, the threat of monopoly is looming ever closer to our scalps.
There’s a saying that when two fight, the third one takes the prize and while the two ‘zillas fight it out in the silicon arena, we, the buyers, get to save some extra cash while being able to afford processors we were barely dreaming about three years ago. All this could change soon…
One slip
For tens of years, AMD has been just a shadow behind Chipzilla. Just like a shadow, copying almost every move their “master” made. While the general behavior of AMD was “shadowy”, they were however shadowed by Intel in almost all areas. At the dawn of the new millennium however, the Chimp maker became increasingly naughty, dealing a few good slaps of the face of the tyrannical aristocracy also known as Intel. But the boys in blue didn’t pay too much heed to the old prophecy telling them that the green scary goblin will come and take away their prize if they don’t wake up from their arrogant numbness.
Then came the NetBusted architecture from Chipzilla and the P4 eggs fryer. And then came the K8 architecture from AMD introducing the first commercially successful 64 bit processors. Of course, all this is history… we all know it, we all heard it time and again. But I needed to remind myself how things were. How Intel slipped.
Because it did. Badly. In its arrogance and pride, it thought the little dirty green boy could be bullied off the market and wrestled out in the corner with just one arm. It was usually right. However, in this case, it got it wrong. Very wrong.
For two years, AMD spoon-fed Intel its own crap, strangling market share while practically humiliating Chipzilla in almost all performance related aspect and relegating Intel to second place in all but the notebook arena - which is actually where Intel found its salvation. And as if the nightmare wasn’t enough, Intel also had to endure a damn solid beating in the power consumption aspect. The roles reversed.
I remember reading in 2005 a funny line off AMD’s site. It had a rather unfair and calumnious PDF in which it made fun of Intel’s NetBusted architecture. It said something like “it’s amazing they (Intel) got a permit from the fire department to emit that much heat”. I considered this to be pretty slimy and under the belt on AMD’s part, but it was still funny. It is then when my image of AMD begun to change. It was becoming the bully, but, as it became apparent only later, it wasn't smart enough to take advantage of this.
So that’s the drop of history I’ll never forget. And no AMD fan will ever do so. Not even the Intel fans will forget the events at the middle of this decennium. Because it was Intel’s wake up call, it was the first time the bullied kid hit back, at least that’s how I see it. It’s too bad that the kid started laughing and basking in the spotlight, while the bully pulled out his shotgun…
Mistake or inability?
Actually, it’s pretty damn hard to imagine that AMD didn’t see it coming. I mean, DAAMIT, people come and go a lot from those fabs’n’labs.
I find it hard to believe that the Chimp didn’t know what Core 2 would be with 6 months before Intel launched it, maybe even earlier. So why did AMD let this happen? And why didn’t the counterstrike come swifter and more efficient?
It might have believed its K8 architecture was invincible, just like Intel did in the past - until AMD taught them the opposite, at least, the hard way). But while Intel can be excused, since it has been leaders for the better part of 30 years, AMD cannot be forgiven for underestimating its rivals.
Intel learned a lot in a few years and changed its ways drastically. In only three years, Intel became aware that it needed to change, to pay more attention to its clients while also showing more respect and openness to the end user and the press. It started being very open and, most importantly, stopped behaving like it had a stick up its arse, it stopped being so conservative. Intel became efficient and innovative while AMD didn’t do much in the past three years. And now that it released a significant product (the native-quad core processor), even though it is a good piece of innovation, it fell short of being the miracle that it desperately needed.
So what about the other possibility? Maybe AMD simply didn’t have the “brain” to do more. This doesn’t mean that its engineers are morons, it’s just that Chipzilla simply came with a fantastic architecture.
Now that about a year and a half has passed since the release of Nucleus 2, history proves that the two years of undeniable AMD supremacy were nothing more than a mistake, a spike, or should we say, a downfall for Intel.
A spelling mistake. Really, Intel made it look easy these past 18 months. It kept on pushing out one CPU better than the last, again and again, while AMD was doing the same thing, but to me, it was the 90s again… when all of AMDs releases were way behind the Intel counterparts. And really, while AMD tries to put on a brave face, this won’t fool me, the press, or the fans of both camps. I was and still am an AMD fan but, like any fair individual, I know damn straight that the Chimp is back to its cage at the circus… the spotlight is back on where it always was…
The Money Factor
Chimpzilla’s situation was quite nice during 2004 and 2005. It was expanding and paying its debts. More importantly, its larger competitor had serious problems and the war resulted in human casualties as Intel laid off personnel in almost all areas, in order to increase their efficiency. AMD instead, arguably decreased its efficiency by buying ATI. For the ATI part however, we should wait a few more years, who knows, maybe the move was a good thing. But right now, since AMD has indebted itself like a junkie from his dealers and started selling the furniture in the house for the highest bidder, it sure looks like things are going downhill really hard.
Another thing that we must find very upsetting about AMD is its price policy for the period summer 2004 to summer 2006. It’s obvious that when you got a technological advantage, you should exploit it and earn more money. However, AMD failed to do this the right way. First of all, it got too greedy when it came to it products. It charged ridiculous prices, attempting to win too many times more the production cost for a single top-of-the-line CPU.
The situation is proved by the price cuts we saw in the past year and a half, when the price fell by almost 70% for some CPUs. I mean, yeah, it’s ok to win some cache but c’mon, this is too much. Of course, AMD needed the cache, but this is not an excuse for exploiting its competitive edge so much as to bleed us customers dry of every penny for 100MHz more. The upsetting thing is not the strategy but the fact that it lost exactly the one thing that made it “cute” for the guys who always cheered for the team in second place.
While Intel abused its monopoly to shove down our throats the expensive line of NetBust CPUs, AMD lost its popular charisma by abusing its fragile edge to get more cash in order to… what? Build a super-K10 processor? Nope. Save the world from hunger? Nay. Destroy an alien race threatening to take over the Earth? Try again. Buy ATI Could be. Was that a smart choice? Yes, no, maybe? I guess Fusion and the future will tell us… speaking of which…
The Future
It’s quite clear that if Chimpzilla doesn’t find something spectacular soon, it’s going to hit serious trouble. Previews of Fusion, Bulldozer and Bobcat got me sizzling with anticipation, especially the low voltage processor but after the f(enom)iasco of this November/December we should try not to read too much into AMD’s future plans . All in all, this is a sad Christmas for AMD and something tells me it won’t be having such a great holiday. In the meantime, things are going swell for Intel, which successfully moved to 45nm while AMD is still having trouble with 65.
The fight for the next 2 or 3 years seems to be very well in Intel’s favor, and while we all love having great CPUs under the hoods of our PCs, it would be really sad to see AMD KO-ed. Of course, a total defeat is highly unlikely, and after all, rumor has it that an old South Korean man named Samsung is interested in having a Chimp pet. As a matter of fact, Samsung would really help out AMD’s situation and such a move would probably make Intel shiver. If Intel is a bully in the schoolyard, then Samsung is the schoolyard. µ
Tags: Amd
Share this:
Share
all this doom and gloom for amd. what is wrong with people. it took intel 3 years to leapfrog amd. i would never use intel no matter what new design or new process they acheive. its called loyalty folks. we need competition. do you want to spend 1000.00 or 1200.00 for a cpu, i sure dont. if i had to guess it costs about 50 bucks to make that, probably less as many factories as intel has, intel has for the most part great marketing. and shed loads of money , but the simple fact is people want choices, and amd provides that. i mean how do you feel about intel having record revenue, not to mention 80 percent of the marketshare and laying off at the same time. i know its hard to swallow that one. amd has had it's share of hard times, and they may continue to, but im convinced there will always be someone out there that makes sure they survive. now for my prediction, nvidia and amd will join forces. makes perfect sense, sorry hector but you guys need nvidia's ceo, you lost some good people, but its time to fight the good fight. cya
I was a devoted AMD customer until recently. 

The requirement to upgrade motherboards with each CPU upgrade has left a bad taste in my mouth. 

Any "savings" from buying AMD, was quickly negated by the need to buy the supporting motherboard. 754, 939, AM2, etc compared with Intels 775 which has been around forever. 

I think moving the memory controller onto the CPU (I originally thought was genius) has become an albatross around their neck -- not to mention them resting on thier laurels while INTEL puttered into the lead. 

I haven't jumped the AMD ship yet, but I have been reluctant to upgrade from the 939 becuase of all the memory innovations. 

Why spend top dollar on a AM2 system if it is going to be orphaned like the 754 and 939 in six to nine months.

Just my 2 cents worth.
I really liked your article. While I am skeptical about Samsung's interest in AMD, I fully agree with the chronological events outlined here.
Intel has wrestled AMD to the floor, but now they will be easing off on the stranglehold. They want AMD weak, but not dead. Intel can't take the risk of becoming a monopoly.

Remember when Microsoft invested in Apple? If is becomes necessary, I wouldn't rule out Intel investing in AMD on some pretext of technology-sharing.
Looking at the slides for Bulldozer, I think AMD is making a big mistake, the archituctre is far from balanced. 

Phenom is an FP monster, what good was that. AMD is focusing on improvements in HPC and server applications. 

But the average consumer cares more about gaming, and normal applications. If Bulldozer can not outperform Intel Nehalem in these applications, AMD has no chance even if it was the best server chip.

As for the IPC improvements that will come with the 45nm version of Barcelona, AMD needs more than just a 10% boost. 

And this boost needs to come in applications where it has a big deficit like 3dmax, zVisuel Kribi3D, and games.
So, yes, Intel's CPU were behind AMD's in the P4/K8 days and now it's the opposite. 

But: There is a huge difference.

How much market share did Intel really lose back then? A couple percent? So, even though Intel had the inferior CPU, they still held by far the largest market share and made more than enough money.

AMD, on the other hand, is not in the luxurious position of being able to do still do well even though their CPUs are behind. 

That's their dilemma.
No presents under the AMD Christmas tree this season. Maybe a few re-gifted San Diego chips...
in 2004 (first hit on google for both companies)

Intel did 34 Billion with 9 Billion profit

Samsung did 83 Billion with 8.4 Billion profit

Intel is the Duke Nukem of the play ground and thanks to the OC 45nm Quads calculating physics generating craters is no problem.

Samsung is better off letting IBM deal with it
I too am an AMD boy but they really need to get their act right quick.

A really Excellent article.

Keep up the good-work!
ROFL, you should post more like that on toms. Did you do those calculations on an abacus?
The Article was a good summary of the feud between the Foes.but I do believe that AMD lacked execution(some serious execution of Architecture improvements).

The original K-10, which was slated for 2005 did not leave the labs.

The present architecture (called as K-10)has every good arsenal to fight against Intel,eventhough it's an iterate of K-8. but they have to fix their TLB Error,Data path between L2 and L3 Cache and add more Cache memory(If Intel can add more cache,why can't AMD do the same, since the architecture can host upto 12 Mb of Cache(2Mb each at L2 and 4Mb at L3).

now over to Samsung..If Intel has Cache memory,Samsung provide those.about 60%of the cost of production of an Intel Processor is for the Cache Memory,which is from Samsung.If you have observed, the Samsung has been a trend setter in Memory technology.so Intel needs Samsung(at least for now).

I do agree with the author and do doubt the execution of AMD at the present scenario.ATi was an essential buy, but AMD has to pay much for the deal,and at the wrong time too.when AMD sat with ATi for talks, it was at least a billion Dollar cheap.if they executed that deal on that year,we could have a healthier AMD now, and the original K-10 project would not be slated as K-11.Surprised huh? AMD had concepts of Bulldozer in mind way back in 2004.Dirk Meyer( i think so) wanted to make that happen, but to make a massively parallel architecture was a challenge and that make sense when AMD made talks with the Dudes at GPU Arena.it got intensified, when the Cell Processor was born.but as a curse for any organisation,Politics,I heard played its dirty game.

The Multi-Core game saved the Grace of not bringing in a new Architecture,but it was not a fix that a new architecture can provide when released at 2005.(at least on 2006).

The Bulldozer should execute to bulldoze Nehelem.this can only bring AMD back into the Race. I am quite skeptical to believe in AMD promises because, they have not delivered what they have promised.In Fact,Barcelona is not the best AMD can do(Cmon guys, you can do better than that).

I hate to put my money at Intel.but I do not know whether i do have a choice,since Phenom is not even executing on par with Core2 at the same clock.We need the AMD at 2003.the AMD at the days of K-6,K-7 and K-8, where the Architecture said it all.
I, too, am a long date follower of the Intel-AMD war. 

And the war is full of reversals from one side to the other. There is only one battle that Intel is constantly winning : the fabs. They are constantly ahead. 

And their lead is increasing. In 1998, both were pretty much even on .25, .18 was close (6 months), .13 was (8 months), .09 was very late (12 month), .65 was even worse (15 month) and .45 looks worse.

Granted, AMD was more aggressive on technological advances (copper, SOI, APM), but I don't remind them being ahead of schedule a single time.

CPU engineers are doing their jobs. K8 wasn't such a hard design to make. I used a proven core on a top-notch platform (Direct Connect). The only coup they pulled was a very good and on schedule dual-core design, right at the time market needed it. Now they are trying to improve everything at the same time (Core, platform, process) and they struggle. 

The phenom will make them survive, but after that I thing AMD will have to reinvent itself, becaus I don't see anything against a 4GHz nehalem.

I think AMD should take the low-end road and capitalize on power efficiency and laptop. There is a huge growth potential in the 6-hours lasting laptops, ultraportables, UMPC, smartphones, portables consoles, consoles, etc.

AMD should abandon HPC and high-end servers. They should take the low cost/low power way.

The phenom core is very good to do that, "Fusion" that to a potable graphic chip on dye and you have a winner, no mather how much faster core 3 is. 

Theese chips will sell, because they will bring what is most wanted : a usable mobile device with great battery life and low cost.
Until your closing paragraph, of course. 

If Samsung,or any other company, for that mater, BOUGHT AMD, AMD's x86 LICENCE, issued by INTEL, would be TERMINATED. 

Do you all get that? NO ONE CAN BUY AMD AND SELL and/or MAKE x86 CHIPS. THE LICENCE WOULD BE REVOKED BY THE LICENCE HOLDER, INTEL! HELLOOOO!

If they could AMD would have been bought months ago,especially now when they are worth exactly what THEY PAID FOR ATI!!!

It's do or die fellas.

SPARKS
With Fab36, AMD only has the foundry capability to supply 20-30% of the market, Intel the rest 70-80%. 

AMD had sold out basically all his chips at premium price before Core 2 came out. Even then, AMD only has 50% of the retail desktop CPU market, which is probably 30% of all desktop CPU market, and 15-20% of all CPU market. 

Intel got much fat margin in Mobile CPU and Server CPU market.

With only 2 major and capable PC/Mac CPU suplliers in the world, AMD, with only 20-30% of the CPU supply, cannot win since Intel can basicly charge any price for the rest 70-80%. 

Imagine a situation in which AMD capable of producing a CPU twice the speed of CORE 2 and selling USD500 per chip for all its chips. 

Intel can still sell CORE 2 at USD450 per chip and making much more money than AMD because Intel has the production capability to do so. 

Even with all the money, it still takes AMD at least 3-4 years to build and run 1-2 new foundry. Inbetween Intel, with much more profit, can definitely catch up.

I think AMD needs at least 4 years technical lead and steady increasing margin to fund a new foundry and also an alternative CPU design team, like intel, to successfully compete with Intel. K8 gives them probably only 2-3 years.
Nice little article, I had fun reading it.

But then I'm not sure if Phenom+HD3800s will be able to save DAAMIT's a*** from getting sunk in the gamer's PC arena.

I mean, is there anyway for a Spider to survive in DDT (Dualcore+Dualcore+TriSLi - no pun intended :D)?

Don't get me wrong, I've been an AMD fan-gurl starting from their egg-fryer processors during the 90s, and I'm even looking forward to my next upgrade but it just looks like it's not going to be such a nice bang for teh buck. ;)
AMD had its lead for 3 years, from 2003 to 2006, and we saw the same bs announcements of death for Intel at the time. 

Really what has not been looked at in this case is how remarkably well AMD has held up this past year and half. 

Apart from the initial market share shellacking that it received after Core 2 was introduced, it has come back to nearly the same share that it had previously. 

The real problem was that both the graphics and processor divisions came out with immature new-gen products, but both of which are recoverable. You're seeing it already with the marked improvement between HD2900 to HD3800 graphics series. 

And likely you're going to see the same improvements come in on Barcelona over time. There won't be domination over Intel like in 2003-2006, but there will be parity.
Pentium 4/D are desktop parts, sonny. Else what the hell was Pentium M?

Intels success with Centrino was from basically one thing:
Pentium is a more recognizable name than Turion.
Taracta & BaronMatrix: nicely said, both of you. 

I can comment that I am not "doom & gloom"y, but simply a bit pessimistic about the future of AMD. Intel did move fast, never contested that ::- ). Congrats to them. BaronMatrix: I hope what you say will turn out to be prophecy. 

Nasanaeru: yes, being recognized as a worthy competitor is indeed a good thing. And the fight is pretty juicy, Intel did a good job.
Bring on the Fusion so we can all laugh at Intel's sorry excuse for what they refer to as a graphics chip.

Intel will still have to play the price game with AMD, the Spider (and AOD) is not something they can just ignore. 

With AMD offering very nicely priced platforms for what Intel would charge for just a quad-core flagship CPU, AMD won't be completely left out, but they really needed those 2.6GHz and up units.
I don't think AMD deserves all the blame for not being able to take full advantage of their superior processors in the past. 

As I recall, Intel made all sorts of questionable deals with computer suppliers to shut out AMD. Until AMD started their lawsuit against Intel, it was very difficult to find a name brand computer with an AMD CPU.

Currently Intel makes the fastest CPUs; but does anyone care besides fanboys? AMD's fastest CPUs are plenty fast enough for almost everyone, and they cost the same or less as equivalent Intel CPUs. 

If you are in the market for a $500+ CPU, you have to buy it from Intel; but if you care more about the price/performance ratio, there are choices from both AMD and Intel.
I can say only that with athlon 64 INTEL was smashed in the corner for 1-2 years..Their processors were sliGHtly slower than athlon64 im talking about p4 and p520

That continued 1-2 years and suddenly p620-630 started to appear with 99%of the instructions from AMD"S athlon64...
After that AMD made dual core x2 socket 939

Wich RULED FOR 7-8 months .Then core duo appeared and guess what boys and girls .It was absolute copy of athlonx2 architecture without 3dnow and integrated controler..(the same architecture will be much more obvious if they copied even the integrated controler)

The instructions were very similar>99%
But 1 year after amd"s x2
I say this put 4mb cashe on 6400+ and lets see the results...
AMD isnt a shadow and never will be .They are innovative firm with low budget ,thats the real truth..
We've seen surprisingly poor execution from AMD all this year and last.

I'm due for a new PC; I still can't bring myself to buy Intel, so I'll be buying a 790 mobo, 3850, and a Brisbane (or Windsor) X2 and hope they get their s$%t together next year. 

(sigh)
I agree, these are darker days for AMD, but things happen very quickly in this market. 

I've noticed, in retail stores like Best Buy, that the majority of computers don't have Intel badges anymore though, they have AMD. 

Which, while the patrons of these stores are ranging from non-techies to techs, I think that their lower prices should sustain them through this challenge. 

I doubt there is any sales person who could sell a regular joe on a QX9770 over anything from AMD based upon price. So! 

For AMD's sake, I hope regular people start feeling their computers are outdated soon. Unfortunately though, that could mean more sales of Vista which is kind of a step backwards (or forward if you embrace DRM).

Samsung has the money and ability to take over AMD. Although every Samsung product I've owned in the past was a flaming pile of garbage (DVD players and the like, I shied away from their microchips to play it safe).

Between a rock and a hard place, AMD is.
Very well said. 

Sadly, it will gets more difficult for AMD moving foward as their remaining architectural advantages (IMC, Hypertransport, and native-quad) will be more or less neutralized with Intel's Nehalem within the next 12 months. 

If a bandwidth constrained Core2 had enough IPC improvments to easily beat the K8, how ugly does it get with Nehalem. 

It will be interesting to see if Intel actually taps its full potential given that they are already holding back on Penryn.
In 2006, Intel's reported revenue was $31.5 Billion USD with a net income of $5 billion.

In 2006, AMD's revenue was $5.6 billion with a net income of negative $166 million.

The Samsung Group had a revenue of $158 billion with a net income of $12.9 billion in 2006. 

I doubt Samsung can do a hostile takeover of AMD, but shareholders willing, I think Samsung has plenty of cash to buy a majority stake in AMD - assuming that AMD is willing.

(Numbers came from Wikipedia).
With all of the crap floating around about a company 1/10th the size of Intel, it seems like everyone liked the 90s and the way Intel put choke chains on all of the OEMs - and customers.

I don't think AMD had overly high prices. Their single core chips were low-priced and the dual cores were costly, but they were 80% faster than the single cores.

At the same time Intel was sabotaging AMD launches, threatening to pull chipsets or charge the other arm and leg for them if OEMs sold more than three AMD chips.

I think this will be a good XMas for AMD as they had record revenue last quarter and are on the verge of $2B per quarter.

We should all be down on our knees thanking AMD for fighting the good fight - with a "convicted" monopolist no less.

We now have quad core that are less than dual core chips at their intro. I don't actually think that's a good thing though.

I expect Intel's plan to backfire and AMDs share to continue to grow. They clawed back some last quarter even in the face of C2D. Once the 65nm process is 100%, the cost savings will be apparent even with a larger die as 300mm wafers @ 65nm will still provide A LOT more Phenoms than Fab 30 made K8s.

And the really funny thing is how everyone sees the Asset-Light initiative playing out. It has nothing to do with "outsourcing," it has to do with providing more capacity at a lower cost.

If anything, TSMC, Chartered and IBM will make chipsets or the low end Rana, while Fab36 makes mostly Barcelona and Fab 38 makes mostly Phenom (X2, X3, X4).

That would provide them with enough capacity for 40% of the market and people can stop saying they can't supply enough chips.

There is a common misconception that Intel did not have a competitive processor during the Netburst era (P4/D).

This is just not so. They had a processor that was very competitive with AMD's K8 (Athlon64/Opteron) processors. 

It was just that this processor was not being targeted at the desktop market; it was being used in the Centrino platform (Laptops). This processor would have given AMD a run for its money clock for clock. 

The only thing lacking in these processors would have been 64-bit capability which wasn't necessary at the time.

I believe that AMD knew this and so did Intel, which is why Core 2 is a descendant of this processor. 

The thing that AMD did not know was the swiftness at which Intel would change their direction and basically sacrificed the Netburst series of processors. 

You see, a poor company, relatively, like AMD could not fathom doing such an expensive culling as what Intel did. AMD had and has plans to have remained competitive with Intel but these plans could not have bore fruit in the limited time that it took Intel to change direction. AMD's plan is sound; they just didn't have the time to execute. 

When you have money, you can work wonders. Kudos to Intel.
Good opinion piece, however one (in my view) glaring mistake. 

K8's success had NOTHING to do with 64bit capabilities. It's success occurred because it was a better 32bit architecture which just happened to also have 64 bit capability. 

While it was marketed for its 64bit capability and the press latched on to this - it's ultimate success in the market place had nothing (or I guess I should say very little) to do with 64bit capability or upgradablility - it was simply a better architecture on the vast majority of SW (largely 32bit) at the time.

For those that disagree - if the 64bit functionality was removed at the time of release would it still not have gained market share based on performance and thermals?
Loved this article. You pulled EVERYTHING together in a fluid and entertaining manner. 

DAAMIT is a match made in heaven but those two are STILL recovering from the merger. 

Phenom & HD3800 are definitely the first steps in the right direction; THE INNOVATION IS THERE. Attention to performance HAS to be next -- or else I'm afraid for them. I've been an AMD fan since K6; I don't think they're outta the game yet. Cheers to Intel for wising up and getting serious. AMD should feel a slight satisfaction being finally recognized as a worthy adversary.
This has to be one of the worst written articles I've ever read.
It's an opinion piece, so don't put much stock innit. ;)

Samsung could buy AMD, but won't because of that one little thing. Oh what is it... OH! debt! ;)

When it comes to the playground bully analogy, Samsung is a small part of the playground. Sony and Asus would have a larger portion.

Cheers,
John
Luckily for AMD buying ATI places them in the console arena. Resulting in them working more closely with MS and kicking Intel into touch this gen with no where to go. 

ATI presence in Wii and XBOX 360 must count for something. 

I don't think NVidia will want Intel to push them out either. So this alliance could prove lucrative in the long run.
Clueless article written by a clueless dude.

Hint:Samsung can't buy AMD.
One thing that's is also on my mind is the Intel-Microsoft duopoly of the past several decades: both companies made their software/hardware as compatible as possible while increasing the performance/bulkiness of hardware/software to keep both of them competitive.

Now that Microsoft is weakening relative to other technology companies (such as Google and IBM), Intel can breathe more easily but its advantage in faster processors is also less of an advantage than before. 

This is allowing competitors like AMD breathing space to make processors that provide a comparable price/performance ratio. And if they bankrupt themselves, as you said, someone better funded will step in and try to compete with Intel.

With so many new technology/chip companies and a weaker impetus for technological growth in laptop/desktop CPUs, I think Intel can no longer can maintain a monopoly, with or without AMD's "help".
Great read.