CORE LOGIC CORE LOGIC CORE LOGIC. Why don't people see this? DELL wants to buy everything from ONE place. If AMD can ship/sell them a package deal the way Intel does, thats what they will buy. I'm sure Dell isn't keen on using 3rd party chipsets, its just one more thing to break and another company for them to deal with. AMD picks up revenue from ALL the OEMs that would end up buying AMD branded core logic chips. HP, Dell, IBM, etc would all much rather have everything coming from one place, and AMD would love the extra $$.

I think the gfx part is really gonna be left alone to their own volition to do as they've been doing, other than maybe using AMD fabs to do some chips, and maybe being able to pick AMD engineering brains for help if they need it and vice versa.

Reading this reminds me of when 3DFX bought STB, granted 3DFX was already pretty screwed at that time, the point is that this doesn't make a whole lot of sense.

I am no professional analyst but I just don't get it how so many of you miss THE point.

Buying ATI is perfectly logical for AMD, for several reasons:

1, Notebooks
Not only here, but on virtually every other tech forum people keep commenting to this news with wild guesses about Torrenza and the IGP business an they totally forget about notebooks.

Notebook sales are skyrocketing and some say will overtake desktop sales in no time. AMD has always been an underdog here, but with the Turions and X2s (i.e. low power Athlon64s with a new name) they would have really had a chance. If you look at the TDP of these new chips they are at Pentium-M levels, yet the AMD based notebooks usually run MUCH shorter on battery. The reason is not the processor, but "everything else".

As much as AMD has been able to sell its marketing crap about just how cool it is to provide "choice" to work with many partners instead of building the whole motherboard with chipset, wireless module and processor themselves like Intel, it's been nothing but soar grapes. It's like saying that you drive a Chevy Metro / Suzuki Swift instead of that BMW because it requires less fuel and you are environmentally conscious - not because you're broke.

By building / designing the major parts of the laptop they would be able to design a better overall power consumption plus manufacture it at their state-of-the-art fab(s) which are ahead of TSMC and Co.

Coincidentally, the largest AMD chipset supplier in that segment is ATI if I'm not mistaken.

2, The complete package
Many IT professionals I've talked to still don't trust a VIA, Ati or nvidia chipset. Even though they admit the K8 is as good as or better than a P4, they opt for the latter because it's "mad by Intel". Also they either buy Dell or if they go with IBM or HP they still make sure it's Intel based. In fact, I can understand them from their perspective. When you build your own PC and discover it randomly restarts because your memory doesn't like your MB you are angry, and frustrated and try to solve it somehow. Now imagine that with 20-30 or more PCs and a business standing still while you are guessing the error.

The IT people I talk to blame the motherboard and the chipsets for this, not AMD, but still, they buy Intel because it builds the MB as well and there are no compatibility issues. In fact, the only new K8 machines at our firm are the ones I bought directly.

3, Plan for the future
With plans for GPUs integrated into a CPU (as the 3rd or 4th core) or even being added to the MB like a CPU for upgradeability it's no wonder Intel is planning on entering the graphics market big time. AMD also has to be ready for the challenge or it will get caught with its pants down like now with Conroe.

As for Ati's reasons, they are anyone's guess. Being so tight with Intel lately was a bit like having a Great White as your best friend. Do they really want to sit and wait until A, Intel intros its own gaming GPU and dumps them or B, buys them out (hostile takeover), grabs the licenses and then eats the company alive?

Plus, this would be the chance to get hold of the AMD chipset market for good. Just imagine how much sooner and how much more information would ATI get about upcoming AMD stuff than nvidia? They would really have to screw up not to produce vastly superior chipsets in the future.

ATI seems to have an engineering style that favors low heat and low power consumption. Combine that with a can-do/will-do company philosophy and you may have the reason for AMD's interest.

ATI knows that Intel will dump them at the first instance so that's their impetus.

Wonder what kind of licensing rights (and tech knowlege) ATI would bring with them. Could we end up with AMD making chipsets for Intel motherboards?

Actually, I hope it doesn't happen. Could be rather bad for nVidia.

This isn't about graphics at all. This is about chipsets. AMD has never been good at making chipsets, and while ATI isn't the best that's out there, their new Radeon Xpress stuff is showing some real promise. Buying up ATI is probably a lot easier for them than starting up a massive new in-house chipset-development team geared toward constant revisions and performance enhancements.

Soon you'll see the package deal with AMD offering Athlon64s on AMD motherboards featuring the AMD Radeon Xpress chipset with onboard Radeon graphics. It'll have everything a computer vendor will want in an all-in-one package, since it will allow them the AMD equivalent of slapping an Intel logo on the box and that's it.

For AMD to beat Intel they need to offer all-AMD solutions, and that means good chipsets, including ones with decent integrated graphics. Buying ATI gives them that.

I just hate what this would do to ATI (and I say this with an AMD chip in all 3 of my systems)... If I was holding any ATI Stock I would either be protesting strenuously, or I'd sell all my stock ASAP. And my reasons for saying this stem from ATI's current relations with Intel, they are partners and this would hinder that relationship.

But I really suspect this is just more "rumor mongering" going on. Wait two more weeks and this will go away. Then wait another two and it will come back again ... just because it gives the Analysts something to say about something that means nothing at all. (Well at least in this case I hope I'm not wrong!)

I still doubt it is true--what's much more probable, if anything major between AMD and ATi is probable beyond the concept of a close working partnership, is that AMD and ATi are collaborating on some kind of joint venture which they simply haven't announced at the moment. Other than that, CNN and everybody else involved in spreading the rumor lately seem to have all picked it up from the same place--the Internet. It reads like the first rumors we saw weeks ago, only slightly more refined and polished, and that's about it. It's the typical round-robin rumor mongering the Internet is celebrated for.

Of course, AMD doesn't need to buy ATi to get ATi to design core-logic chipsets for them, as ATi is already very motivated to do that--at most all AMD would need to do is to license those designs it wants, if AMD would have some purpose for doing so. Indeed, nothing is preventing AMD from contracting with ATi to design future core-logic for upcoming cpus, etc. Nowhere at all do I see a compelling reason for AMD to *buy* ATi simply for core-logic. Last, of course, it behooves AMD to have multiple core-logic chipset designers and manufacturers (like nVidia, ATi, SiS, Via, etc.)--it would not behoove AMD to emulate Intel's single-source cpu bus-licensing philosophy--and that's why I don't believe this rumor. AMD has prided itself on opening up its cpu buses to third parties, and to buy ATi for the sake of core-logic ownership would be to reverse the incredible progress AMD has made in the last several years.

My opinion is that whatever it is that some 3rd parties have observed in the way of "meetings" and "collaboration" between AMD and ATi personnel recently, it isn't an AMD buyout of ATi. It's something else not yet announced, is what it is, and these people, not being in the loop, have simply misinterpreted whatever they've seen--if they've actually seen anything out of the ordinary. I've read several opinion pieces written in various places around the web that attempt to reconcile this patently irreconcilable idea of AMD buying out ATi, and not a one of them has provided me with a compelling rationale for AMD to try and do something like this.

OTOH, I've said for years that it won't be Intel derailing AMD in the future. If AMD is derailed it will come as a result of AMD driving itself right off of the tracks, and this kind of activity might suffice in that regard. But like all stock-market rumors that grow legs, this rumor seems dedicated to convincing some people that AMD is about to make a big mistake, so I wouldn't be a bit surprised to learn later on that the source of the original rumor piece (which was very poorly framed and reasoned) was a stock analyst into whose ear Intel was whispering sweet nothings about AMD and ATi...:D Never underestimate the power of a well-placed dollar here and there to aid in the dissemination of rumor designed to influence investor behavior.

I think the only way this merger can be a good thing is if AMD/ATI have a very, very clear idea of what they're going to do together that they couldn't have done separately through some partnership. I think this would almost have to be a new CPU/GPU/PPU product of some sort, rather than just chipsets. If this is really just some lame response to Intel's "platformization", then I don't think it's going to be worth the hassle.

Has anyone considered the possibility that perhaps this is more of a long range strategy. How? Well, we have all heard about HT3 and the fact that it will have a "socket" for accelerators and the like.

Well, the possibilities are endless, but how about a cpu style socket for a GPU.

Interesting concept to say the least!

Besides no stockholders still heard anything, then there is another few problems.

For GPU, will it end up with AMD/ATI fighting Intel(+nVidia maybe) against lower process technology aswell? Would it just as CPUs end up as 90nm vs 65nm? Intel is also now making chipsets on 90nm while most others do it on 130nm.

For me it just seems that AMD/ATI would need the better product anyday to even compete against the equal Intel block parts. Be it GPU, Chipsets, CPUs. And if Intel buys nVidia (They can do it in cash) you would have a direct 65nm vs 90nm contest with the products we know today.

But we will see, whenever the stockowners get info :P

Also as a note, GPU in a socket with HT or whatever makes no sense. Unless its for the lowend parts, and even here its very questionable. Also remeber latency means nothing for a GPU, its oneway traffic all the way, only bandwidth matters. GPUs just need the memory very tight tied to the GPU. Even DIMMs cant be used here. Also GPU in a socket would kill SLI/CF options rather quick. And coprocessors on that socket is a nice dream. But its something extremely rare thing. Just look today on how many that use these specialized coprocessors that you need special software etc for.

There are many reasons why AMD should buy ATI.

1) In the short term it's probably mostly about chipsets.

2) In the mid-term manufacturing will become more important (capacity utilization is one factor, using top-notch manufacturing process for GPUs another, but keep in mind also that if ATI in future uses only AMD fabs to manufacture GPUs it will cut its design costs by tens of MUSD annually).

3) In the longer-term it's probably a bet on convergence of CPUs and GPUs (CPUs are becoming massively parallel, GPUs are becoming more programmable, 3D acceleration will be required by the OSes, so GPUs and CPUs will not be as different as they are now and some of the functionality may migrate between the two).

But the real question is not whether AMD should buy ATI, but how much it can afford to pay. And the bad news for AMD is that numerous economic studies have shown that buyers tend to overpay and their shareholders get screwed in the end. A few months ago AMD could use its high-flying stock to buy ATI cheaply, but now the terms of the deal are unlikely to benefit AMD shareholders.

So even if it is true that the deal has been approved by the AMD board, it's far from certain that AMD shareholders will be enthusiastic about it.

Ok everyone, repeat this with me: MONEY.

AMD is about to engage Intel in a bidding war -- AMD just doesn't have the same deep pockets that Intel has. This is a fight that Intel and AMD cannot sustain for lengthy periods of time, but they will weather this storm of price cuts and lower margins.

The buyout of ATI would mean several things:
1. AMD would have another "relatively" stable source of income.
2. ATi is already the GPU brains behind the 360 and Wii, this merger gets AMD into the console GPU supplying market.
3. AMD would have ONE major piece they truly lack against Intel: a CORE LOGIC CHIPSET (Corrado says this best)

Whom bother to more for the chipset other than the owner of the chips themselves? Don't get me wrong: ATI chipsets aren't considered "the best" chipsets -- with AMD owning them, they MUST improve.

In the end, I'm betting it's #1 in my list. All the other pieces are pretty sweet, but I'd rather have seen the AMD/Nvidia rumors myself.

I've been spotted at Best Buy.

Rumors that I'm buying the chain are untrue.

If this is true - it must be motivated by a long term view of where the industry is headed. AMD must see a future that we arent imagining right now. Platforms may be it - PCs will fade away and be replaced by 'devices', perhaps acquiring ATI is the first step in assembling the talent and capabilities.

I dont see and short term gain from this - not for servers or pcs. If Dell starts selling AMD desktop chips - will they go ATI like HP ? Its only in servers and workstations that HP offers nvidia based solutions - all the desktops are ATI ... Either way it proves nothing viz this acquisition.

When you consider that Intel is the biggest microprocessor AND gpu manufacturer,

It makes sense that the 2nd biggest microprocessor manufacturer would merge/buy the 2nd biggest graphics manufacturer.

The chipset business is another area where I'm sure AMD would like to dip their fingers into, but unlike many other people here, I don't think that its the primary reason for the merger/purchase.

Plus, ATi is fabless isn't it? They could save some money by using any of AMDs unused capacity, but then again AMD is already outsourcing also. In my opinion, the chipset stuff is just icing...

Where is AMD gonna get the money? I mean, they're upgrading their existing fabs, and building a new one. How many billions is that? Don't forget that the next quarters are gonna hurt as AMD makes huge price cuts and Conroe gets all the attention. From a timing point of view, I think this is the worst time.

In terms of fab capacity, AMD can't make enough chips to satisfy demand as is. Remember ATI last year and the x800's? I'm pretty sure gfx card fabrication is pretty tight already. From a fab point of view, I think this would only make the problem worse.

Intel has recently just cut off several products, and really trimmed themselves up. ATI has a huge business, selling many other things than gfx cards. If AMD were to take over, they'd also take on many products, to the point where they might not be able to manage AMD and ATI's products effectively. From a managerial point of view, I think this is just a bad idea.

Isn't there some AMD or ATI shareholders on here?

If the Torrenza platform is a factor, it's perhaps more to make custom coprocessor chips than GPUs. Why use ATI's design teams I dunno, but perhaps they feel they're getting an especially good price on it.

Perhaps they're considering some of ATI's intellectual property?

Perhaps they'd like to diversify as they get beat around by ICM? Hell, it worked for a while with Flash.

The platformisation perspective does make some sense, because then AMD would have a motherboard division, a graphics division (one where it should hopefully offer better than the GMA series in any metric that's important), and can then offer systems based around this platform to compete with Intel's Centrino. They haven't done this in the past, but as PCs become increasingly commoditised I think this may become a significant purchasing factor for businesses.
-Mole

Still no word on why ATI would say yes to AMD. Its not like AMD has all this Fab space to spare at the moment. And AMD's always catered to nVidia in the past. If anything Intel is their buddy more so then AMD.

ATI stock is rising today on rumors that AMD will make a friendly offer of between $21 and $23 as early as next week.

That's a steal there. Heck, i've got $24 in my back pocket right now, I'm in :-D

In seriousness, I assume that's either a typo or per share...
-Mole

Heh, I really still dont believe it. but it looks very real now.

But whats the prospect?

Either ATI GFX gonna suck or be very good. What will nVidia do? And will this give Intel an excuse to make highend GFX too? (Excuse in terms of monopoly)

On the other hand I would prefer total solutions from AMD and Intel. From top to bottom :)

6th paragraph: "marriage between ATI and NVIDIA" ?
...and AMD would be more probable.

I still think that would be for other reasons than mentioned.

I can only say one thing about this...

This just can't happen! It's really bad for the industry. We have a pretty good balance now where all party's are rather happy.

I also can't see benefits for ATI. Mayby the fabs from AMD, but can they handle all the waffers ATI needs?? I don't really believe they can. They need a lot of waffers for their processors allone. I don't see them produce GPU's, chipsets, xilleons and all the other stuff (mayby when their next fab will come online, but this one must first be build & is there money when they buy ATI??... Having to face lot of competition from Intel now, don't see them make lots of profits, these pricecuts mean less income, but same costs)

A very very expensive gamble.