Hey guys. I wrote a little market analysis on AMD's future with the R600. I wrote it for shi*ts and giggles. Please don't flame... if you disagree simply state why and leave the insults for BaronMatrix to dish out.

-----------------------------------------------------------------------------------

Now more than ever it seems AMD’s future is going to depend heavily on its newly acquired graphics division. In an age where a two year plan is one too many (just ask HP’s former CEO Carly Fiorina), investors are going to want to see a return, and soon. AMD’s financials are better than ever, but we’ve yet to see a full quarter’s results with Intel’s Core 2 Duo factored in. It’s almost a certainty that AMD will lose some market share in Q1 ’07 to Intel and now that AMD’s “K8L” and native Quad Core architecture is firmly delayed until late 2007, they’re going to have to depend on the R600 to keep investors happy. But it's highly unlikely that AMD will do any growing this year, even with a successful R600. Firstly, AMD borrowed heavily to purchase ATI and while profitable, it was no cash cow. Investors like to see organic growth, not growth through acquisitions. Secondly Intel owns the market share in graphics and is reportedly going to give it a second go in discrete graphics chips. And we've all seen what Intel can do with leadership under Otellini. Thirdly, nVidia has a steady secondary revenue stream with the PS3. AMD does have the X360, but despite strong sales this past holiday season, it isn’t going to outsell the PS2 and PS3 combined anytime soon. It's also no secret ATI practically gave the R500 away to get Microsoft's business. And to shake things up even further, Wii has been an enormous success. Finally nVidia has a very strong core logic platform and good corner on professional graphics with its Quadro line of GPU’s. Right now AMD isn’t doing particularly well in either of those markets despite its well reviewed core logic platform. The only place AMD can grow right now is graphics.

In short, AMD’s R600 is going to have to be one hell of a GPU to keep AMD looking good. Its success is going to make or break the newly formed graphics division. All things considered, nVidia’s 8800 series has been a huge success. And despite what some have said, people were willing to shell out $650 clams for a video card. Early this week several sites reported a delay in the R600 to March; more bad news. The greater the distance between generations the bigger the lead nVidia is going to get. And if anyone thinks nVidia’s sitting on its hands, you might want to reevaluate the world’s most successful graphics company. nVidia isn’t on top of discrete graphics because its a slouch. Don’t be surprised if nVidia releases a slew of mid-range DX10 cards when AMD launches the R600. Unfortunately for AMD, it’s lost out on a chance to charge a hefty premium for the R600. Had the R600 launched this month, they may have gotten away with $650. But with 8800 GTX cards now hovering around $550 and GTS’s going for a bargain $400, not many are going to pay for a $650 card that’ll deliver performance barely better. It’s also extremely unlikely AMD will be able to launch mid-range cards simultaneously, so they’ll miss out on that opportunity too.

But alas, there’s still hope for the little engine that could. AMD’s experience with GDDR4 memory puts it in a good position to push out future revisions of the R600 quicker. And another opportunity for AMD is its experience with 512bit memory interfaces, which to date nVidia’s done very little experimenting with. But nVidia could just as easily use those two technologies against AMD. Not to bash the former ATI, but nVidia has almost always been the pioneer in new graphics features and technology. To name a few, they were the first with on chip hardware T&L, pixel shading, SM 3.0, and DX10. ATI’s always been one step behind. But to be fair, ATI has always come back with a nasty punch whenever it releases a new generation. I'm sure the R600 won't be an exception, but it will have to pack the biggest punch yet. Finally, there’s always the benefit of owning your own fab. While no firm plans have been made for AMD to make its own graphics chips, it’s a foregone conclusion that it’ll eventually happen. Why does this matter? The cost of manufacturing GPU’s will reduce significantly once AMD tapes out ATI chips. Meanwhile nVidia has to continue paying TSMC to make its chips. And since there aren’t many small chip companies out there to buy that own modern fabs, it’ll likely stay that way for a long, long time.

Without a doubt the R600 will be the most important release for AMD since the Opteron. Hopefully, for the consumer’s sake, it’ll rock our world.

overall I like your little blurr....

couple quick things though. 8800GTX while a great card, the says havent been astronomical by any means. and last, even if the R600 isnt the performance leader, I believer the Torenza platform is what could make/break AMD/TI. Having low power, cost efficient equipment still carries a large market, and AMD has found its place in the server and mobile area's with comfort.

still good stuff though jesse, keep it coming

I'm considering the r600 despite a predicted high price just to support AMD (because I believe in competition and I would hate to see AMD/ATI die. ). Well of course its not all to support AMD/ATI I want a kick ass card too . And the fact that I have never owned an ATI card intrigues me .

BTW I read somewhere that pairing an 8800 with a 680i board would result in better performance. I believe this was the nVidia site though so I'm not trusting it so much yet.

No cash cow? Their margins were in the 50+% level, that's a cash cow, with 2+ billion in cash on hand at the time of the acquisition.

Depends on the market, when you're in a medium growth or declining market growth through growth markets isn't a bad idea (as seen in the mobile acquisition in the phone market), unfortunately the graphics market had it's first quarter of contraction this past quarter.

Pfft, steady stream? Nah, the PS3 is a dud, and what money do they get for the PS2? Original Xbox maybe, but those sales are drying up, specifically because M$ wasn't happy with them.

Check your figures, the X360 surpassed the PS2 in August, and has sold 5+ million units in the US sofar (2+ million from Nov to NewYear's), and 10+ worldwide according to NPD research. And the X360 outsold the combined sales of the Wii and PS3 during the holiday period worldwide. But remember that AMD/ATi also makes the Wii processor so AMD/ATi vs nV the sales are about 7 to 1, and total sales sofar from those new deals would be about 20+ to 1.

Less succesful than the X360 though despite being a new commodity, either way, more sales for AMD/ATi.

And has just lost exclusivity of intel buddy-buddy with SIS now being given the nod from intel as well.

Which is a costly portion of the business with very low margins and often a loss-leader.

Actually the AMD chipsets are fine, and better than most, the issue will be switching focus from intel to AMD, which doesn't happen quickly, but once it does, then end-to-end solutions will move the product regardless of others' performance since OEMs will look for bundles. This of course hurts nV who gets alot of current sales from AMD platforms.

Well it will be required in order to make AMD look like they haven't got a dog on their hands, but the futureof AMD isn't so much wrapped up in the R600 as it's potential offspring. What will the profitable mid-range look like, what will it bring to the CPU+VPU core that AMD needs for it's future? Those are the important factors.

It's not newly formed, it's newly acquired. It was formed long ago.

Really? Sales were much lower than the GF7800 series, so how successful is it? And with the graphics pullback last quarter how succesful can anyone really consider themselves? The card is a great card, but I doubt it's anywhere near as profitable from nV as the GF7600 is or the X1900 seires is for ATi.

But not many of them (less than 4% of the market shelled out more than $250). And it's not the most profitable segment either. The big money maker and the big sales come in the $150-250 segment.

That statement is pretty much an ad for nV. Who said nV would sit on their hands? Also, isn't intel the most successful graphics company, and until AMD bought them weren't ATi outselling nV?

Until you know what it is and how it compares to the competition how can you even make that claim? Once again too much cheerleading not enough objectivity.

Why is that extremely unlikely? They launched the X1800XL and X1300 at the same time, and the X1800XT and X1600Pro within a month of each other shippinh shortly afterwards, so it's not extremely unlikely especially if the production issues affect the R600 and not it's offspring. nV had a staggered launch last time and this time. I suspect that AMD/ATi will launch the whole lineup at the same time, but if anything they may do a staggered delivery like they did last time.

And what production experience does ATi have with 512bit memory interface? Other that the ringbus which isn't even true 512bit so much as a segmented 256+256 composition. Only 3Dlabs had a shipping 512bit product and even that was different.

How, by promoting slower is better, like the low-k is dangerous strategy?

Seriously that's pretty Fanboi-like talk there. The nV was a generation behind with the FX series, and ATi made the first single chip accelerator, first to output to TV, first with PS1.4/DX8.1, first with PS2.0/DX9, first with DDR2, first GDDR4, first with 110nm, first with low-K, first to 90nm, first to 80nm.... both companies have storied pasts, and both have led and followed.

Actually until AMD has enough capacity TSMC and UMC will conitnue to make the ATI parts, the one area that they may exploit is using AMD's R&D resources to reduce development costs which is the most expensive part of the process. I doubt they'll move anything to AMD fabs for a while (likely the 45nm IBM/AMD process will be the first to go AMD instead of TSMC/UMC).

Maybe, but it's the offshoot/spin-off products/benifits that matter far more than the marquee/vanity/niche product the R600 is itself. Single produts likely more important to AMD's future would be their mobile refresh in the spring/summer, and their move to CPU+VPU system on a chip for their products for the low end PC, home entertainment, micro-servers, mobile, and industrial solutions.

I agree with that. Main thing is to have people pushing the market forward, but the thing you totally ignore is the fact that the most important things to nV and AMD's future is their success in the low-mid range market with the new potential entrants (intel, VIA/S3, SIS) to this profitable segment where they make the money to pay for these marquee e-penis products.

The true future of both companies rests in their ability to survive that compeittion, especially from intel, and nV likely has the most to lose.

Some things to consider.

With all due respect I find your reasoning/statements false or short sided.The main reasons behind it for one can be found by the grape apes counters.What Ive found on thg is alot of leaning one way or the other for whomever a person prefers for their hardware.But the worse Ive read is the doom and gloom of this. When I read a story rant or opinion regaurding AMD,its usually ludicrous sounding like omgitsthelatestgreatestcomehitherallyouWONTbelievewhateventuallywillhappen crap,OR the short sided but well meaning worry(for competitions sake)or triumphant fanboyism of doom and gloom.Everyone here has to understand that joe blow out there DOESNT know what we know, BUT joe blow has the dough to make the difference.He's the one thatll buy the hype,WHOEVER is feeding it to him/her.When making statements as broad as THEYRE GOIN DOWN YA'LL just keep joe blow in mind.Like was said earlier,its NOT the gtx that sells,nor the xtx.Its the midrange that sells with joeblows money,I could go on but my point is made.I understand what youre saying and in a 'smarter world' youd be right, but really,do you think joe blow knows anything like we do?

I'm not sure I follow what you're saying. All my opinion was trying to convey is that AMD is up against some stiff competition despite owning a graphics company. A lot of people are saying AMD/ATI is a stronger company, but that's yet to be realized. Like I said early on, a two year plan is one year too many for investors. Will AMD survive the next year? Absolutely! What I'm saying is that people better be prepared for some bad financial news from AMD this year. The won't go in the red of course, but 2007 won't be any 2006 for AMD.

I thought it was a very nicely written article that raised some interesting points, and held my attention right to the end! Ta very much, roll a couple more out when you feel like it!

I don't have a ton of time right now. I promise I'll write more later, but I want to discuss one thing.

I suggest you take a look at ATI's 2Q 2006 report.
http://media.corporate-ir.net/medi [...] I_Q206.pdf

ATI's gross margin in 2Q was 28%. Good, but no 50+%. Net income (the number everyone cares about) was only 5% or $34,000,000. As for having $2 billion in cash on hand... not sure where you're seeing that. ATI has close to $2 billion in total assets, but only $210,000,000 in cash.

IIRC, nVidia has been the first to release new-gen architecture over the last few generations of videocards, and as the story always goes, there are few games that need anywhere near the horsepower offered by the new beasts. If you play games at normal res. like 1280x1024, name me five games that won't run smoothly with high graphics settings on anything but the 8800. ATi will release the R600 when they can; they know how long these things take, and if there's a delay, they deal with it. It's happened before.

Not accounting for the potential of the midrange offspring of the new technology (from either vendor), I did catch wind of the benchmarks via INQ, and the R600 blew away nVidia in every test game, and had an edge in 3DMark. And, we have yet to see the real DX10 performance of the new cards, which is quite relevant; conveniently, enthusiasts will have R600 available to them around the same time that Vista is becoming available. I respect the effort that you put into the analysis, but a lot of it seems to be usual doom and gloom that accompanies a delayed launch. Like Grape said, you need more objectivity to legitimate your predictions.

Regarding the Wii: It has a IBM processor not an AMD/ATI processor but the graphic is made by ATI though the core clock was around 240 mhz

Regarding the PS3: PS3 isn't a dud if it sells out. The "process" of creating them was a disaster but the overall sales (100%) is pretty good.


Anyway AMD did finish 7th last year in semiconductor revenues (though that is gross and not net) while Intel finished 1st.

AMD Revenues news

I suggest you take a look at ATI's 2Q 2006 report.
http://media.corporate-ir.net/medi [...] I_Q206.pdf

ATI's gross margin in 2Q was 28%. Good, but no 50+%. Net income (the number everyone cares about) was only 5% or $34,000,000. As for having $2 billion in cash on hand... not sure where you're seeing that. ATI has close to $2 billion in total assets, but only $210,000,000 in cash.

By the way that you worded your statement I though you were talking about AMD, since you were talking about AMD borrowing...

"But it's highly unlikely that AMD will do any growing this year, even with a successful R600. Firstly, AMD borrowed heavily to purchase ATI and while profitable, it was no cash cow."

So it's not really clear whether you talk about AMD or ATi being the Cash copw, but actually both had a ton of Cash on hand, in fact both were talked about as having so much cash as to be looking for purchases so both are cash cows technically. And a 28% versus nV's mid 30s (don't have time to check at work but the numbers from mid-year) isn't bad, and that ATi was profitable for all 3 quarters leading up to the acquisition means that it's a revenue generator not a money loser unlike when phone companies bought cell phone companies at the turn of the millenium, so it would still be a cash cow, only it'd be a calf not a huge Holstein.

To be more accurate the margins went up from 28 to 30.1% for Q3 (which accounts for the period prior to the merger, ending May, remember their Qs are 1 ahead of calendar) and the projection for Q4 was still increasing margins;

http://media.corporate-ir.net/medi [...] I_Q306.pdf

Also, you mention that income was 'only' 5%, yet that's after all other expenses including options, that's still pretty good for a company, and better than most fortune 500.

As for ATi's CASH position you can read yourself, it's 644 million in Q2 and in 518mil in Q3 (remember short term investments while not 'cash' are considered cash equivalents since they are short term equity, not tied-up assets), so really it still qualifies as a Cash cow, but I still don't see the evidence to back up your assumption that it's a dog. And whether or not it meets some universal standard of cash cow, I don't see how that even relates to your statement about growth, borrowing to increase capacity or to add profitable segments means positive growth, so your statement and your supporting comment are completely unrelated. The reason AMD might not grow so much is because they may have reduced demand for their products in light of competition or production problems, but it has little to do with their borrowing money if it's manageable debt. Having cash on hand and having no debt usually means you are not leveraging your assets to their full potential, and that's pretty much what AMD was doing before, now we'll just have to see if the previous unmet demand now results in a balance between supply and demand rather than shortages and undercapacity concerns. They may still grow, but potentially the need for that growth may be lessened compared to a year ago, however Dell and HP's bysprobably will keep them selling all he chips they can make for some time, and maybe if they do finally reach some overcapacity then they can consider shifting their production of VPUs to own fabs.

In anycase; I think your focus on 1 product, which represents the tiniest portion of the market, as being the lynch pin of AMD's success is rather myopic, and I know alot of people including staunch nV supporters who'd say that they expect AMD to in fact abandon the high end VPU market (even abandoning the R700), so saying that the R600 is so important to AMD's success will need a little more substance than what you've written sofar. If anything I agree more with the nVidiots despite themselves, because what they don't realize is that the true winners will get the mid range, low end and IGP market, like intel exploits now, and the marquee products lihe the G90/R700 might truely be exposed for what they really are which is simply expensive R&D and marketing, something that likely neither ATi/AMD nor nV will be able to afford at the same levels as previously displayed if the competition takes away all their profitable segments on which that house of cards is built.

IMO, the success of the R600/G80 is more important to the future of high end graphics, than the health of either ATi/AMD or nV. The R700 and G90 might be the last of the hyper refresh high end cards, heck the G80 and R600 may already be the last of the 18month cycles. You might see new lines in 18mth cycles, but likely not with the same large performance boosts, simply feature additions.

Perhaps you should think about that, because truely that affects us all more than if AMD, or any company for that matter, has a few quarters of negative results.

Well I was talking about the Hollywood VPU not the CPU. As the discussion is ATi/AMD v nV, then it would focus on the VPU not the CPU/core.

Sure it is if the development cost is high, sales are below projections used to price the unit, and especially a money losing unit dependant on media sales to finance it, but that's a Sony perspective, this was more about the contributions to nV. So. considering that nV makes money of the number of sales of the unit, selling out isn't as important as the number of units sold. Selling out 10 units doesn't represent alot of money, selling out 2 million units is significant contibution to the financial success of a company.

Well the question will be what happens after initial pre-orders and fans are satisfied. The Gamecube came out like gang busters and clobbered the Xbox selling like 4:1, but by this time last year the Xbox worldwide had outsold the Gamecube by about 20%.

However the situation of he PS3 versus X360 is like the PS2 versus Xbox, it's unlikely Sony will ever catch the X30 at this point, and even the Wii will likely never catch them, but IMO if anyone could it's the Wii.

But it's still significant from a size and economies of scale perspective. AMD grew their overall by 90%, and their share increased greatly too, whiel Intel contracted their revenue but pretty much maintained their share (likely meaning lower prices for same share).
But really that's like a mouse gaining weight to become a gerbil, but an elephant losing like 10 pounds, the numbers by themselves seem big, the the relative impact in miniscule.

PS, also if you were to combine Qimonda and Infineon back to their situation last year, they'd have finished above AMD in 5th and pushed AMD down to 8th overall.

Uhh, the PS2 does not have nVidia graphics, it has a special chip co-designed by Toshiba and Sony.
The Xbox 1 did have nVidia (like the PS3), but it's discontinued.
Instead, Xbox360, Wii and GameCube have Ati graphics.

Didn't say it was a bad idea, but going forward investors are certainly going to want to see ogranic growth out of AMD/ATI.

EDIT: Sorry... got PS2 confused w/ xbox.

I doubt that'll seriously affect market share. But who knows...

Have you seen the cost of a Quadro lately? How is charging $2500 for what is essentially a 7900GTX low margin? Quadro's are nothing but rebranded GeForces with (sometimes) more memory and different drivers. Please show me an article or story that states nVidia's professional segment is a loss leader with low margins.

I said the chipsets are well reviewed. I make no claim that ATI chipsets are poor performers. What I'm saying is ATI chipset sales and market share are nowhere near nVidia levels. But neither are nVidia's compared to Intel so... I guess it's not a totally fair statement.

Agreed

Wrong word used. Meant newly acquired.

I said all things considered. Those things being:
1. No DX10 title
2. No vista
3. High price tag

Shipments are lower than what nVidia expected, but not by a landslide. And that $650 price tag will probably balance out the lost revenue compared w/ 7800 or 7900 sales. Either way it's largely speculation since we haven't seen a financial result w/ the 8800 GTX factored in. (Results should be out in Feb)

Agreed.

Intel is the most successful in terms of market share. But if you think their integrated GPU's makes them successful... well... there's no hope for you. ATI has outsold nVidia in terms of revenue once or twice, but make no mistake nVidia is king of the hill in terms of revenue and GP. ATi has also only beat nVidia in market share a couple times, but those were quarterly results.

It's an opinion. I just don't think AMD is going to be able to get away with a $650 price tag like nVidia did. I'm not saying they won't, i'm just say i think they won't.

Yes, that's true. However, the reason I don't think ATI will be able to launch mid-range are the delays. They can't start making cards until the core is finally taped out, which it isn't. 1 month probably isn't enough time for TSMC, IBM, or whoever ATI is using right now to ramp up an entire line of new cards in volume. But again, I could be wrong. But even if they do manage to launch high end and mid-range simultaneously, there will be supply problems.

Was thinking of ringbus, which yeah, isn't true 512bit. My bad.