ATi Reports Third Quarter Results (Press Release)

[Geo]

ATI Q3 2005

Idiosyncratically selected, as always –no warranties for accuracy express or implied.

Orton:

Q3 “challenging and disappointingâ€￾, but the sun will come out soon.

“Much more than seasonal softnessâ€￾, due to product mix shifting towards the low-end.

“lower than expected yieldsâ€￾ for the lead-free transition [which doesn’t get enuf ink, so far as I can tell].

“broad industry supportâ€￾ for CrossFire from mobo and AIB partners [gad only knows why on the latter].

“strong growthâ€￾ in chipsets, but margins aren’t as good on chipsets.

First to publicly demonstrate hardware accelereated h.264

“In May, Microsoft launched the Xbox360 at the E3 trade show. It is powered by ATI’s revolutionary 90nm 360 graphics gpu [“360 graphicsâ€￾?]. Built from the ground up for the console world with architectural advances such as ATI’s unified shader architecture, intelligent memory, and a new and innovative modeling engine. These features are a ‘must have’ for next generation game consoles, and provide a snap-shot into the future of pc gaming technology.â€￾

[Hmm. RSX doesn’t have any of the “must haveâ€￾. Must be a coincidence. Heh.]

CFO:

Nice growth in unit volumes. “In particular, revenues for desktop discrete declined from both system integrators and retailâ€￾ and this is the main reason they missed their guidance. “Competitors bundling strategyâ€￾ also impacted revenues. [this is an NV sli ref –more on that later]

Nice growth in Radeon Xpress 200 for amd/intel progress/acceptance. 160% sequential, quadrupled revenue year over year in chipset.

Looking forward to new FCC regs lighting a fire under DTV area.

2/3rds of gross margin decline attributable to desktop discrete. Gross margins on integrated is “high single digitsâ€￾ and the acceptance there has been so good that it is also dragging down the rest of the corporate gross margins. Have a program to get gross margins up on integrated over the coming quarters.

Missed OpEx guidance because of impact by increased R&D, which was up 14%. “The increase in the quarter was largely driven by a jump in prototyping costs related to the development of future products on 90nm process technology.â€￾ [Insert ominous music soundtrack here] Expect to be relatively flat on R&D in Q4.

Orton:

Must reduce cost in chipsets.

“We have also made significant investments in our new product introduction process, as we prepare to bring a full range of 90nm gpus to the market this fall.â€￾ [What the heck does that mean?? A process for introducing new products? How do you make significant investments in that?]

New “VP of Operationsâ€￾. [Hmm. Maybe that’s part of the “significant investmentâ€￾. Sounds like his job is to make the trains run on time. . or at least more on time.]

Re R520: “Yes, we’ve taken a hit on schedule. [Gee, who’s sig will that end up in?] We are now back on track, so stay tuned as we get prepared to take the definition of ‘performance leadership’ to a whole new level. And not just at the high-end. But with the full family of next generation 90nm products that will set the standard for performance and immersive multimedia experiences.â€￾ [Bold words! Hope he doesn’t have to eat them some day. Jen didn’t find his too tasty re NV30]


Repositioning X850 over the summer due to new products (theirs and others) as they’ve been preparing to do for months. [i.e. here come the price cuts]

“We believe there will still be a strong market for AGP thru the summer and fall.â€￾ New products expected to “complementâ€￾ current line up rather than immediately replace.

Re revenues from R520: “We have seen some delays. Particularly when you deal with a new technology at the high-end it’s a combination of both architectural and technology bringup. We have those well under control now.â€￾ Describes as “conservativeâ€￾ the decision to not predict revenues from 90nm in the quarter ending in August. They are “extremely encouragedâ€￾ about 90nm and very happy with their experience with XB360 on 90nm both ramp and yields.

On 6600 as a competitor: “Platform bundlingâ€￾ hurting them. X800 & XL are very cost effective and hitting that segment. X700 pcie doing well in the market. “high/lowâ€￾ strategy is pleasing them. Think X800 CrossFire is the best value in the market today in that segment.

Re R520 product in channel: “Late summerâ€￾ and “extremely excited about the capabilities and performance of the new product.â€￾

Midrange R5xx timeframe: “We do plan to bring a family of 90nm products to the market this fall –late summer by my definition; for the back to school fall cycle.â€￾ [Sounds like August to me, and that pushing R520 back has not pushed back the mid-range chip]

Bundling by competitor: Later to market than our competitor, and got hurt. (Bundling mobo with cards for SLI/CrossFire).

CrossFire & R520 availability: CrossFire in July, “channelâ€￾ shortly thereafter. R520, “We want to get out there with the product in the channel when we launch. So I think that you can look at them as having kind of consistent timing in terms of both availability and launch timeframe for that class of product.â€￾ [That sounded to me like not quite predicting same-day availability. . but that might be the goal, with a little slippage on availability allowable.]

No impact due to customer ramps or product repositioning (ie. X850 or CrossFire) on R520 delay. “Complex product that we had originally targeted to launch in the early summer and now we’re targeting for late summer.â€￾ [So a design issue is what I get out of that, rather than as they tried to ramp to production volumes hitting issues, or a “marketing windowâ€￾ for X850 and CrossFire] The “push outâ€￾ of R520 was a significant part of adjusting Q4 guidance downwards. [Well, at least I had that part right. . . they were originally expecting significant revenue from R520 in June-Aug]

Q: You guys suck [paraphrase] A: No we don’t [paraphrase].

IGP Margins: Gotta have margins above 30% in semiconductor business across company. IGP will be slightly different and somewhat lower, but they like the synergy with other parts.

CrossFire vs SLI: Very excited, brings some capabilities the market hasn't seen before. Believes the market “needs a choice in this segment, and needs openness in this segmentâ€￾ and ATI will bring that. Recognizes it is only a segment and not something the entire customer base will adopt.

 

[Megadrive1988]

I listened to the majority of the conference call this morning, maybe 90% of it, but I kept nodding off to sleep so I might have missed something, did they mention the GPU for Nintendo's Revolution at all? I heard Orton mention Xbox 360 once or twice...

edit: it seems they did not mention Nintendo.

p.s. I hope that ATI also has a design win for the GPU for the next generation Game Boy, be that a Flipper-derivative or something new.

 

[WaltC]

(I posted this in another forum and rather than repeat it in so many words I'll just reprint verbatim here, with a minor addition or two.... )


Under the GAAP rules they have to report the cost of employee stock options granted during the quarter, and that's what caused the "loss"--remove that and ATi actually profited by about ~$8M in the quarter. I'm not exactly sure about it, but I think that many US-based companies don't have to expense out employee stock options on their P&L's (but it's been awhile since I looked at accounting...)

I think ATi missed the boat, though, this past quarter in a couple of ways:

(1) They put too many eggs in the PCIe-OEM basket while forgetting that OEMs were still shipping lots and lots of AGP alongside PCIe, and still are. This caused them to hit shortages of both retail consumer AGP and PCIe products during the quarter.

(2) It also caused them to spring their AGP-based products late and in short supply. I think a lack of AGP supply drove both retail consumer and OEM purchases back to nVidia, as a result, as nVidia correctly ascertained that jumping on the PCIe bandwagon too early too strongly was ill advised for the short term.

(3) They've invested heavily into the core-logic chipset business but have hit some yield bumps along the way that have slowed down their sales in that category for the quarter.

That said, I think this is merely a tiny bump in the road, as ATi is poised to reap rewards from xBox360 (both tangible and intangible) later in the fourth quarter when it begins shipping, its chipset business when they get production snafus corrected, and of course ATi's upcoming R5x0 products, which are presumably about a month away as well (I'd expect them to be announced in July with shipments beginning in August and ramping to volume in September.)

This wasn't a stellar quarter for the company, but it certainly wasn't "horrible" by any means. I expect them to get progressively better off as the year advances and they start reaping the fruit of their efforts to date this year.

 

[Bouncing Zabaglione Bros.]

Under the GAAP rules they have to report the cost of employee stock options granted during the quarter, and that's what caused the "loss"--remove that and ATi actually profited by about ~$8M in the quarter. I'm not exactly sure about it, but I think that many US-based companies don't have to expense out employee stock options on their P&L's (but it's been awhile since I looked at accounting...)

Some do, and some don't. IIRC US companies are all supposed to be moving towards including stock options as expenses in their reports.

I first read about this a few quarters back because it was pointed out that Nvidia's big profits would have been turned into losses if they'd correctly reported stock options as expenses.

 

[WaltC]

Some do, and some don't. IIRC US companies are all supposed to be moving towards including stock options as expenses in their reports.

I first read about this a few quarters back because it was pointed out that Nvidia's big profits would have been turned into losses if they'd correctly reported stock options as expenses.

Thanks, BZB, as that's how I remember it as well. I seem to recall the question coming up for lots of other companies, too, as you mention, like Apple and Microsoft, as well as nVidia and many others.

 

[Bouncing Zabaglione Bros.]

Thanks, BZB, as that's how I remember it as well. I seem to recall the question coming up for lots of other companies, too, as you mention, like Apple and Microsoft, as well as nVidia and many others.

It seems to be a big problem for tech companies because of the way they tend to give stock options in lieu of high wages. Apparently Microsoft has been doing business this way a long time. I guess dodging salary expenses makes a big impact when so much of what you "make" comes from peoples' heads rather than a physical product,

UK companies have been forced to report stock options as expenses for quite a while, as all our big financial scandals triggered the writing of those laws in the late 80's.

 

[Richard]

Hmm... nVidia gets xbox contract. nVidia slips with process problems on the next desktop part. NV30 roughly 5 months late against ATi's equivalent card.

ATi gets xbox 360 contract, rumours of process problems for next desktop part, R520 now expected 4 months later than GF 7800...

Is this the start of a xbox curse? :?

 

[WaltC]

Hmm... nVidia gets xbox contract. nVidia slips with process problems on the next desktop part. NV30 roughly 5 months late against ATi's equivalent card.

ATi gets xbox 360 contract, rumours of process problems for next desktop part, R520 now expected 4 months later than GF 7800...

Is this the start of a xbox curse? :?

I wouldn't think so since unlike nVidia, ATi isn't tasked with the manufacturing and supply of the 360's components, which removes a tremendous burden from their operational expenses, I should think--with the exception of accelerated and short-term R&D expenditures related to their 360 component designs.

Also, ATi is branching into general PC core-logic chipset support, never forget, and in that vein ATi has to absorb not only relative R&D expenses but also is affected by yields and connected supply issues. ATi is behind nVidia in terms of its chronological involvement in the core-logic chipset supply markets.

 

[Richard]

I wouldn't think so since unlike nVidia, ATi isn't tasked with the manufacturing and supply of the 360's components, which removes a tremendous burden from their operational expenses, I should think--with the exception of accelerated and short-term R&D expenditures related to their 360 component designs.

Also, ATi is branching into general PC core-logic chipset support, never forget, and in that vein ATi has to absorb not only relative R&D expenses but also is affected by yields and connected supply issues. ATi is behind nVidia in terms of its chronological involvement in the core-logic chipset supply markets.

I'm talking coincidence & superstition and you come back with rational arguments backed up by facts? 8)

Seriously, point taken.

 

[epicstruggle]

I'm talking coincidence & superstition and you come back with rational arguments backed up by facts? 8)

Dont you hate when people do that.

epic

 

[ondaedg]

(I posted this in another forum and rather than repeat it in so many words I'll just reprint verbatim here, with a minor addition or two.... )


Under the GAAP rules they have to report the cost of employee stock options granted during the quarter, and that's what caused the "loss"--remove that and ATi actually profited by about ~$8M in the quarter. I'm not exactly sure about it, but I think that many US-based companies don't have to expense out employee stock options on their P&L's (but it's been awhile since I looked at accounting...)

I think ATi missed the boat, though, this past quarter in a couple of ways:

(1) They put too many eggs in the PCIe-OEM basket while forgetting that OEMs were still shipping lots and lots of AGP alongside PCIe, and still are. This caused them to hit shortages of both retail consumer AGP and PCIe products during the quarter.

(2) It also caused them to spring their AGP-based products late and in short supply. I think a lack of AGP supply drove both retail consumer and OEM purchases back to nVidia, as a result, as nVidia correctly ascertained that jumping on the PCIe bandwagon too early too strongly was ill advised for the short term.

(3) They've invested heavily into the core-logic chipset business but have hit some yield bumps along the way that have slowed down their sales in that category for the quarter.

That said, I think this is merely a tiny bump in the road, as ATi is poised to reap rewards from xBox360 (both tangible and intangible) later in the fourth quarter when it begins shipping, its chipset business when they get production snafus corrected, and of course ATi's upcoming R5x0 products, which are presumably about a month away as well (I'd expect them to be announced in July with shipments beginning in August and ramping to volume in September.)

This wasn't a stellar quarter for the company, but it certainly wasn't "horrible" by any means. I expect them to get progressively better off as the year advances and they start reaping the fruit of their efforts to date this year.

Even if the reporting of stock options is what caused a loss, it's still a loss to the share holders.

I agree with the PCIe egg basket statement. The only other thing I can think of is they didn't win as many OEM bids as they thought they would. In the retail market, Nvidia took a tremendous amound of agp sales. I think the retail market was underestimated until now.

From what I've read, I don't think it is a yield issue with their chipset problems. It sounds like it's compatibility issues.

As long as ATI has a strong product coming, they can weather a short storm. Their biggest problem right now is getting all their divisions up to speed. Nvidia's core logic division is extremely strong in the AMD market and they seem poised to take on almighty Intel. They will soon be offering workstation and server boards as well.

ATI better learn the same lesson that Nvidia did during the NV30 fiasco. IF you can't release a product on time with decent supply, then the consumer will buy somebody else's.

 

[RussSchultz]

FWIW, Microsoft now gives STOCK, instead of options.

double p.s. precognition points for me for recognizing that ATI going all PCIE happened too soon for the market.

 

[JoshMST]

Just like NVIDIA did, ATI has a broader base of products to keep the company going strong through tough times on the desktop market. In the past, with companies such as 3dfx, missing a product cycle was essentially deadly. 3dfx would have been in a lot better shape if they had released the Voodoo 4/5 at the same time that NVIDIA released the original GeForce (Voodoo 5 5500 would have walked all over that card). Instead they missed their fall release and sent it into spring... right into the jaws of the GeForce 2. Bad timing, too much debt... but mainly just missing a very important product cycle.

ATI will see some short term hurt from this, especially in the Fall buying season where most of the machine specs are already finalized. I think that NVIDIA won a lot of the high end range with their 7800 GTX, and we will see that card transplanting the X850 XT/PE in the top end systems. It does make you wonder though if many of these OEM's will then use the 6800 GT or Ultra as the next step down from the ATI part (since even a simple thing as changing a driver on a line can be bothersome for these large OEM's).

I still have the impression that the R520 is going to be a very fast card, and I wouldn't doubt that it would be faster than the 7800 GTX. I think the main issue here will be availability through most of the Fall. I just hope that (against all odds) somehow ATI will let me in on their R520 release. Not holding my breath though!

 

[JoshMST]

FWIW, Microsoft now gives STOCK, instead of options.

double p.s. precognition points for me for recognizing that ATI going all PCIE happened too soon for the market.

Yeah, I think that ATI really made a big mistake in not refreshing their AGP lineup late last year really hurt them. The 6600 GT is not all that much faster than a X700 Pro, and if ATI had an AGP X700 Pro at Christmas, that chip would have been received a lot better than it has been to date. It is unfortunate, as it is a very solid performing part. Up until this spring, the only ATI midrange parts for AGP were still the 9800 Pros. It wasn't until the X800 AGP was released that the $200 to $250 range had a new part (as the X800 Pro/XT/PE AGP were all above $400 up until that time). NVIDIA saw this golden opportunity and went full bore with the 6600 GT on AGP.

 

[Bouncing Zabaglione Bros.]

I agree with the PCIe egg basket statement. The only other thing I can think of is they didn't win as many OEM bids as they thought they would. In the retail market, Nvidia took a tremendous amound of agp sales. I think the retail market was underestimated until now.

They did, it's just that those wins didn't ship as many PCIE as they thought. PCIE is not compelling enough on it's own to cause people to upgrade, so they all waited for their next normal upgrade cycle. The problem with that is that a lot of people consider their machines mostly "fast enough", and without attractive new CPUs, or even that much killer-app gaming gaming software, a lot of people simply wern't willing to pay the high price for a new PCIE gaming rig without worthwhile upgrades to the rest of the system being available.

 

[WaltC]

Even if the reporting of stock options is what caused a loss, it's still a loss to the share holders.

But the point is that shareholders can see the loss is connected to the costing out of employee stock options as opposed to a direct loss resulting from its daily business operations.

The one real negative I can see with this report is that ATi's carried inventories are too high at the moment. I think that will hurt them the most with investors short term.

I agree with the PCIe egg basket statement. The only other thing I can think of is they didn't win as many OEM bids as they thought they would. In the retail market, Nvidia took a tremendous amound of agp sales. I think the retail market was underestimated until now.

Actually, as I said, most likely nVidia picked up a lot of OEM sales for AGP as all of the OEMs to my knowledge were selling AGP alongside PCIe, and still are. ATi was essentially right in its strategy as it picked up nearly all the initial OEM contracts for PCIe (with its native PCIe parts--which nVidia didn't have to offer at the time), but they were wrong in terms of their timing (as they simply went whole-hog for native PCIe before the PCIe OEM markets had fully ramped up to exceed AGP production.)

As long as ATI has a strong product coming, they can weather a short storm. Their biggest problem right now is getting all their divisions up to speed. Nvidia's core logic division is extremely strong in the AMD market and they seem poised to take on almighty Intel. They will soon be offering workstation and server boards as well.

ATI better learn the same lesson that Nvidia did during the NV30 fiasco. IF you can't release a product on time with decent supply, then the consumer will buy somebody else's.

ATi's problem was never that it "couldn't" release an AGP product lineup when it should have--the problem was that it chose to concentrate on PCIe prematurely at the expense of a still very strong market demand for AGP. Certainly, it would have been both easier and cheaper for them to have followed nVidia's path this year with respect to PCIe product rollout.

I remember commenting a few months ago that only time would tell if ATi was premature with its initial emphasis on PCIe at the expense of AGP. Now we know... I wasn't too surprised by this, really, except for the idea that if I could see it at the time ATi certainly should have seen it, too.

 

[PatrickL]

Has Orton given any indication why Mobility X800 (and prolly X800XT) seems so bad sellers ?

 

[siconik]

(I posted this in another forum and rather than repeat it in so many words I'll just reprint verbatim here, with a minor addition or two.... )


Under the GAAP rules they have to report the cost of employee stock options granted during the quarter, and that's what caused the "loss"--remove that and ATi actually profited by about ~$8M in the quarter. I'm not exactly sure about it, but I think that many US-based companies don't have to expense out employee stock options on their P&L's (but it's been awhile since I looked at accounting...)

I think ATi missed the boat, though, this past quarter in a couple of ways:

(1) They put too many eggs in the PCIe-OEM basket while forgetting that OEMs were still shipping lots and lots of AGP alongside PCIe, and still are. This caused them to hit shortages of both retail consumer AGP and PCIe products during the quarter.

(2) It also caused them to spring their AGP-based products late and in short supply. I think a lack of AGP supply drove both retail consumer and OEM purchases back to nVidia, as a result, as nVidia correctly ascertained that jumping on the PCIe bandwagon too early too strongly was ill advised for the short term.

(3) They've invested heavily into the core-logic chipset business but have hit some yield bumps along the way that have slowed down their sales in that category for the quarter.

That said, I think this is merely a tiny bump in the road, as ATi is poised to reap rewards from xBox360 (both tangible and intangible) later in the fourth quarter when it begins shipping, its chipset business when they get production snafus corrected, and of course ATi's upcoming R5x0 products, which are presumably about a month away as well (I'd expect them to be announced in July with shipments beginning in August and ramping to volume in September.)

This wasn't a stellar quarter for the company, but it certainly wasn't "horrible" by any means. I expect them to get progressively better off as the year advances and they start reaping the fruit of their efforts to date this year.

I find the contrast between this post and Walt's post from about a year WRT Nvidia, shall I say, quite telling.

BTW, that thread is interesting in its own right, with people calling Josh Nvidia apologist for basically suggesting that they were not going to crumble on the spot and asking if he owned Nvidia stock. I bet he wishes that he bought some at that point though. (

 

[Geo]

Damn, I was going to publicly congratulate Younger Me for staying the hell off that thread, but then I noticed that apparently YM couldn't resist a single inane comment about potential antitrust issues with Intel acquiring NV ('cause, y'know, that could happen any day now).

Ah well.

 

[Geeforcer]

So looking at the market cap differentiation between ATI and NVIDIA, is ATI at these levels potentially a purchase target for NVIDIA?