RD580 production hampered by M1575 southbridge supplies?

[trinibwoy]

We don't know what we don't know, 'tis true this time as in every other. What we have is on the plate in front of us from a generally reputable (tho sometimes wrong) source says what it says about motivation here. I don't question the motivation as being legitimate, just the previous statement as being rather sleazy if this is where they were headed. And if they hadn't made up their mind yet, a simple "no announcements at this time" would have sufficed nicely back in December.

Honestly, trying to evaluate the integrity of a company spokesperson's statemements is very low on my list of useful activities. They were bullshitting - so what?

 

[trinibwoy]

If Nvidia sees that a better southbridge from ATI is coming soon, from ULi's perspective they would be better clearing loads of those M1575s as this is their best chance. Its not the case and that is why most of us have the oh-so-not-good-but-doesnt-matter feeling watching these low tactics.

Given the stated and implied reasons for the aquisition of ULi, I think clearing old inventory would be one of the first things on the list. Nvidia bought people and real-estate - not products or IP. A clear contrast to the Hybrid acquisition for example.

 

[Arty]

Unless they (Nvidia) underestimated demand and thought ATI's SB would already be out. If that is the case maybe they are scarce b/c Nvidia simply did not make enough as they assumed the demand would already have ceased for them.

Could be.

Too much damage control going on though.

 

[Geo]

Honestly, trying to evaluate the integrity of a company spokesperson's statemements is very low on my list of useful activities. They were bullshitting - so what?

YMMV, of course. I get a signficant portion of my kicks from reading the tea leaves and guessing what lies ahead, so prolly get moderately more annoyed at willful misinformation (as opposed to "oops", which also isn't fun, but in to every life some oops will fall, both your own and others).

 

[Bouncing Zabaglione Bros.]

How much money do you think Nvidia makes on each ULi southbridge. And how do you think it compares to the benefit to ATi?

It's money going to Nvidia rather than ATI. You've got the choice of taking half the chipset revenue from ATIs chipset products (as ULi were doing), or handing ATI the additional revenue of a southbridge (as Nvidia have now done).

 

[Mariner]

How much money do you think Nvidia makes on each ULi southbridge. And how do you think it compares to the benefit to ATi?

The question should be, how much money does Nvidia make on each nForce chipset sold in comparison to each ULi southbridge? I'd guess it would be one heck of a lot more profitable and it has the added bonus of crocking a direct competitor to your own product because the current ATI southbridge isn't up to par. It's all just the niceties of the business world and I'm sure ATI would be acting in a similar manner if the tables were turned.

Assuming the rumours about ULI southbridge 'shortages' are true (and I'd guess they are), I'm sure that it's no accident that this is occurring - but then, I didn't ever expect anything different. I'd guess that NV will keep the chips trickling out until the newer ATI southbridge becomes available, just enough so that the motherboard manufacturers don't get too pissed off!

 

[trinibwoy]

YMMV, of course. I get a signficant portion of my kicks from reading the tea leaves and guessing what lies ahead, so prolly get moderately more annoyed at willful misinformation (as opposed to "oops", which also isn't fun, but in to every life some oops will fall, both your own and others).

If it was a public statement then I have no issue with it. I'd be surprised if that statement also doubled as an official communication to ULi's existing customers on the future availability of ULi parts. That would be mighty grimy.

What I find telling is that, according to the digitimes article, Nvidia believes that they can stall the adoption of ATi chipsets by reducing supply of ULi southbridges. If they are correct, then I see no reason why Nvidia would not want to make that play. Any delay in ATi's (inevitable) rise in the chipset space would be in Nvidia's best interest.

 

[trinibwoy]

The question should be, how much money does Nvidia make on each nForce chipset sold in comparison to each ULi southbridge? I'd guess it would be one heck of a lot more profitable and it has the added bonus of crocking a direct competitor to your own product because the current ATI southbridge isn't up to par. It's all just the niceties of the business world and I'm sure ATI would be acting in a similar manner if the tables were turned.

I didn't buy that at first - that holding up ULi shipments would result in more Nforce shipments. But apparently there is a lot of doubt in the OEM space about ATi motherboard solutions so maybe that is the "real" reason (and not just simply screwing ATi).

 

[Geeforcer]

"NVIDIA intends to supply ULi customers with current products for the foreseeable future."

Supply customers with current products != fulfill all orders or increase/maintain production in the face of increasing demand brought about by RD580 release.

 

[Mize]

It's money going to Nvidia rather than ATI. You've got the choice of taking half the chipset revenue from ATIs chipset products (as ULi were doing), or handing ATI the additional revenue of a southbridge (as Nvidia have now done).

I disagree with this.
nV can effectively delay XFire (the only useable variant) for weeks or months. That's the real value in this equation.

 

[Geo]

Supply customers with current products != fulfill all orders or increase/maintain production in the face of increasing demand brought about by RD580 release.

Maybe in a computer program. In the world of normal human-to-human communications, at least disclaiming the "maintain" part would be a contentious read. Tho it's not impossible that as ATI mainboards increase that a shortage is caused even by maintaining, for which NV would have no responsibility. Nevertheless, what is on the table is Digitimes description of what is going on here, an "acceleration" of cutting back on ULi production.

If you have a new source to the contrary you'd like to toss on the table, feel free.

 

[Sxotty]

It's money going to Nvidia rather than ATI. You've got the choice of taking half the chipset revenue from ATIs chipset products (as ULi were doing), or handing ATI the additional revenue of a southbridge (as Nvidia have now done).

BZB is definitely right on this point, and for that reason I find it slightly difficult to believe Nvidia would screw themselves out of money on purpose. Unlike many here I don't believe they represent the evil empire, I believe the seek profits first, and thus my assertion that perhaps it is more of a mistake than a willful attempt to be evil.

 

[Geo]

Having checked the thread, so far as I can tell you are the first to use "evil" or any variant. I did use "sleazy", but not to describe either the acquistion or the decision to ramp down production. So, y'know, quit projecting, please (unless I missed it of course, in which case feel free to point).

 

[Mize]

BZB is definitely right on this point, and for that reason I find it slightly difficult to believe Nvidia would screw themselves out of money on purpose. Unlike many here I don't believe they represent the evil empire, I believe the seek profits first, and thus my assertion that perhaps it is more of a mistake than a willful attempt to be evil.

I think you're way off on comparing the net margin of a southbridge with the value of a competitor's product delayed. What killed the X1800 more? The cost of chips or the hard launch date?

I would quickly sacrifice a low margin product to delay a competitor's basis for a high margin product by months.

I don't like what nV did here, but I would have done the same probably.

 

[Blacklash]

The Sapphire board won't be using the ULi SB and it still isn't out.

ASUS doesn't seem to have a problems supplying RD580 boards with the ULi chipset either.

 

[trinibwoy]

I think you're way off on comparing the net margin of a southbridge with the value of a competitor's product delayed. What killed the X1800 more? The cost of chips or the hard launch date?

I would quickly sacrifice a low margin product to delay a competitor's basis for a high margin product by months.

Ditto.

 

[Geo]

http://www.beyond3d.com/forum/showpost.php?p=729378&postcount=1

Orton just said "We don't see it as an issue at all" re ULi SB availability.

He also talked about having a 1.5 week (1M chipsets) SB "oops!" at a fab. Because of constraints on capacity they weren't able to make it up immediately, and this impacted their ability to ship as many as they wanted of both NB and SB.

Which makes me wonder if Digitimes' sources had the wrong end of the stick on what was going on with shortages. . .

 

[trinibwoy]

Which makes me wonder if Digitimes' sources had the wrong end of the stick on what was going on with shortages. . .

In what way? Orton didn't seem to refute the report of shortages, he just seemed to brush off the perceived impact (not like he could do anything else).

 

[Geo]

"we don't see it as an issue at all" isn't refuting the idea of shortages? Now, maybe he's flat-out wrong, but I do indeed see that as intended to refute the idea of ULi SB shortages of any signficance.

 

[trinibwoy]

"we don't see it as an issue at all" isn't refuting the idea of shortages? Now, maybe he's flat-out wrong, but I do indeed see that as intended to refute the idea of ULi SB shortages of any signficance.

geo, I think you're confusing two things here - the allegation that a ULi southbridge shortage exists and the assertion that such a shortage has a material impact. Which one exactly do you think Orton is refuting? I pick the latter.