Notes from the NVIDIA conference call Q&A session

I'm glad we can ressurect the "when will NV30 ship" line of discussion again.

Reminds me of a bumper sticker on a rusted out 70's Bonneville on with bull horns I saw in the dorm parking lot in college: Something this evil never dies.
 
Yes, and we're all glad that you can resurrect yet another useless comment.

You know, if nVidia would present any factual information that would give anyone some degree of confidence over the shipping day / launch of nV30, then we wouldn't have these speculative threads, now would we?
 
Joe DeFuria said:
Let me paraphrase Jen-Hsun's original answer: "I'm not going to tell you whether it has taped out or not, because you simply don't need to know and wouldn't know what to do with that information anyway. The only thing you guys need to know, as market analysts, is that the product will be ready for Christmas."

Well, perhaps the "analyst" is not interested in PR bullsh*t "predictions", but real FACTS so that the analyst can make his own predictions.

A couple calls ago, Jen-Hsun also said that the NV30 would be "launched" in August and be the "fall" product.

So perhaps instead of being fed another "this is what we believe will happen" line, the analyst DOES want to know when / if the chip actually taped out, so that the analyst can guess for HIMSELF when the product might be ready to ship...

I have two problems with this argument. First, the analyst had ample opportunity to press his question and force Jen-Hsun to answer directly. That fact that he backed down instead tells me that the analyst didn't really need or care about the answer to the technical "tapeout" date question, and he was satisfied with an answer that addressed the date of product availability.

Second, the analyst simply wouldn't have enough information to "guess" when the product might be ready to ship; it would be a poorly educated guess at best, and Nvidia doesn't need analysts making poorly educated guesses about their product launch dates. How familiar is the analyst with Nvidia's processes? What if Nvidia has made some new innovations or technical breakthroughs or taken some other unusual steps in NV30 development to accelerate the time from tapeout to product availability? Would the analyst be able to factor these developments into his "own predictions"? It would be unwise to give an analyst a tiny piece of the overall product development picture (such as the tapeout date) and let him run with it and draw his own conclusions. I would have done the same thing as Jen-Hsun.
 
and Nvidia doesn't need analysts making poorly educated guesses about their product launch dates.

No, nVidia can do that all on their own. ;)

All kidding aside, as I said, I see nothing wrong with answering the question directly, "No, it did not tape out yet, but what's important is product avilability, and we are confident...."

EDIT: I want to be clear that I am not assuming that it hasn't taped out yet. I don't know, because he didn't answer the question. Though one would think that he would have said "yes, it has taped out" if indeed it has...

I'm not even going to try and guess whether the analyst was satisfied with the answers given or not (though if I recall, he did ask a SECOND time about volume shipping?)...some people don't ask twice because they are embarrassed, etc.

All I know is, someone asked a question, and a direct answer was not given. Logical conclusion: answering the question directly (with no other explanation) WOULD PROBABLY lead the analyst to reach a different conclusion on availability than nVidia is saying.

How familiar ....What if....Would the analyst be able to...

Or what if the CEO is making an overly optimistic, though possible, guess based on the actual development of the chip? Much simpler explanation. As long as there is a possibility for it to ship for x-mas, the CEO will take the position that it will.

We hear all the time about products being "later than we thought", "expenses being higher", "revenues being lower", etc. It's very rare to hear "I know we said last quarter that it would be our sprinf '03 product, but it's ready NOW!"

I would have done the same thing as Jen-Hsun.

I would not.

I would have answered his question first, and THEN talked about expected shipping / availability.
 
What if Nvidia has made some new innovations or technical breakthroughs or taken some other unusual steps in NV30 development to accelerate the time from tapeout to product availability?
...
Second, the analyst simply wouldn't have enough information to "guess" when the product might be ready to ship;

I see major 2 factors influencing time to market from tape-out. One, how much time is needed to debug your design. This is mostly about brain power, there is no magical tech or enchant or whatnot that will make your engineers more intelligent. Second, once the design is debugged, it's off to nVidia's partners (TMSC, retailers and OEMs). There's nothing nVidia can do here. They have had the same partners for a while now and there hasn't been any "breaktrough" in speeding up production from these partners.

I think the past few years is plenty to make a good prediction as to when NV30 will reach retail once it has taped-out.
 
SteveG said:
Here is a more complete transcript of the NV30 tapeout question:

Q: Has the NV30 taped out yet, and if so, when did it?

Jen-Hsun: I think the most important thing for you to focus on is when is NV30 going to be on the market, and we are very confident that it will be in time for the Christmas season...[lots of irrelevant marketing speak here]... we fully expect it to be here for the Christmas season... [more irrelevant marketing speak]... we have built a couple of test chips [on the 0.13u process] and we are delighted with the functionality of the process...

Sorry, but you totally bent the facts with the way you "condensed" the call.

After this:

Jen-Hsun: I think the most important thing for you to focus on is when is NV30 going to be on the market, and we are very confident that it will be in time for the Christmas season...[lots of irrelevant marketing speak here]... we fully expect it to be here for the Christmas season... [more irrelevant marketing speak]

There was another question:

Caller: And how bout the .13 micron technology at tsmc? Would you say the yields look pretty stable at this stage of the game?

Jen-Hsun: The yield at this point for any new generation process is not expected to be stable because you want to improve it. And so ah, the thing that, that uh, is important to focus on is that we've built a couple of test chips and uhm, we're delighted with the functionality of the process....

The way you edited that call makes it look like he was still talking about the nv30 taping out when he said they produced "test chips". They were past that and talking about the actual .13 micron process yields. The way you posted that, it's easy to assume that they had built test chips of the nv30. By including all of the info, and the actual question he was answering, now people can see he was not necessarily talking about the nv30 when he talked about "test chips". Those "test chips" may have been nothing more than tests of the .13 micron process itself rather than the nv30. Heck, they could have made a test nv28 on .13 for all we know.

If you are going to provide a "more complete" transcript of the call, do it properly without leaving out very important information, which can give a completely different interpretation to what was actually said.

SteveG said:
Jeesh, how do ridiculous rumors like this get started?

You should know. They get started by doing things like you did in your previous post. Leaving out important details here and there. Stretching little pieces of info into whatever way you want and coming up with off base conclusions. Someone sees what you say, repeats it as fact, someone else stretches it more and repeats it.......
 
SteveG said:
A fair response, in my opinion.


How about putting it this way : they don't want to say that the chip didn't taped out because it hasn't . Why ? Because analysts are able to make optimistic predictions this way without being guilty for misleading the public into buying more stock to drive the prices up on stock market .

That's why they are so evasive , because they want to leave the impression that everything is working out fine while it isn't .
 
jjayb said:
Sorry, but you totally bent the facts with the way you "condensed" the call.

After this:

Jen-Hsun: I think the most important thing for you to focus on is when is NV30 going to be on the market, and we are very confident that it will be in time for the Christmas season...[lots of irrelevant marketing speak here]... we fully expect it to be here for the Christmas season... [more irrelevant marketing speak]

There was another question:

Caller: And how bout the .13 micron technology at tsmc? Would you say the yields look pretty stable at this stage of the game?

Jen-Hsun: The yield at this point for any new generation process is not expected to be stable because you want to improve it. And so ah, the thing that, that uh, is important to focus on is that we've built a couple of test chips and uhm, we're delighted with the functionality of the process....

The way you edited that call makes it look like he was still talking about the nv30 taping out when he said they produced "test chips". They were past that and talking about the actual .13 micron process yields.

You make a very good point here, that was my mistake. It is a little harsh to say I "bent" the facts though, as that implies it that I intended to mislead. I honestly didn't realize that the "test chips" statement could be taken to mean "NV30 test chips", but in rereading I see that could be implied by the way I ran the quotes together. I tacked on the "test chips" bit to show that the .13u process seems to be going well, and should have separated it from the other quote. I have now edited the original post. Hopefully in time to head-off any new rumours. :rolleyes:

By the way, nice addition of the "ah"s, "uh"s, and "uhm"s to your quote. You know, non-word sounds usually are not added to meeting transcriptions. Talk about putting a spin on someones statements. You try to imply that he sounded hesitant and uncomfortable in discussing the .13u process, when that was not at all the case. Nice try though.
 
Heh, no spin involved. I didn't include any "uh's" or "uhm's" he didn't say and didn't do it to make him look evasive. Just trying to be exact. Would be kinda dumb for me to get on you for not including everything, then doing it myself. I've never transcribed anything before. Took me forever cause I kept replaying the call and seeing that I'd missed an "uh" or "uhm" here and there. ;) Now I Know I didn't need to waste my time. (it does really annoy me when people "aaaahhhh" and "uhhmmmm" alot though. Especially reporters or people speaking publicly) Mistakes happen. Although accidental it did ultimately bend the truth. Sorry for being harsh. At the time I didn't know if it was on purpose or accidental. Thanks for changing it.
 
Just what are we trying to conclude here? Whether or not the NV30 will ship this year? Whether or not the NV30 will be ready in November? December? January? Whether or not nVidia is bluffing with this uncertainty?

I think that since the term tape-out has caused to much confusing on the web (not here at beyond3d mind you, but at Anand and other places!) it seems a bit fishy to make the uncertainty in the call into a do-or-die for the NV30 this year.

This is all we have from the CEO:

We’re very confident that it will be in time for the Christmas season.

and:

We fully expect to be there for the Christmas season

On the other hand we don't know 1) whether the NV30 has taped out, 2) whether they have taped out parts of the chip-design already and received them, 3) whether they have delayed the chip to include last-minute changes to fix any DX9 HLSL-access to PS 2.0+ and VS 2.0+, 4)whether they have first silicon back and are trying to confuse ATI, or 5) whether they have f#*ked up and are trying desparately to get a grip on things.

Yes, I understand that all this speculation is fun, and I like to follow this argument also, but I would put my money on a Christmas shipping (albeit probably not at large volume) but it may still meet demand.

Is a late fall/Christmas launch of the NV30 a good thing? Yes, because it is good for the sake for competition that ATI with R300 finally have the upper leg some months.

Is a shock delay of the NV30 into early 2003 a good thing? No. You do the math...
 
I'm amazed at the nitpicking and lengths people are going to over what is obviously a delta of approximately one month and a half uncertainty (Is it end of november, december or early January???) when obviously the only people in the position to know that are NV engineers.

As for me, I'll err on the side with the official sources as the best guess as to a release date.

I don't particularily care what you call it (whether its a fall product part, a fall and a quarter part, or a flipflop part). Obviously its the more interesting product in there lineup for the readers of this website.
 
LeStoffer said:
4)whether they have first silicon back and are trying to confuse ATI

I'm sorry, but this keeps getting suggested and just doesn't make any sense... Why would Nvidia want to confuse ATi? What possible benefit could this be to them? The R9700 is already shipping, they can't lull them into complacency anymore!
 
Nagorak said:
LeStoffer said:
4)whether they have first silicon back and are trying to confuse ATI

I'm sorry, but this keeps getting suggested and just doesn't make any sense... Why would Nvidia want to confuse ATi? What possible benefit could this be to them? The R9700 is already shipping, they can't lull them into complacency anymore!

I was just trying to say that we really don't have a clue. ;)

Maybe #4 doesn't make much sense, but my point is that all we know is that nVidia says that the NV30 will be out for Christmas. Nobody - not even nVidia - knows more before they have first silicon back. Period.
 
I think ATI knows alot more then people here give them credit for, they are also looking @ .13 micron, Nvidia is not the only company wanting to make that jump
eek5.gif
 
Not to completely interrupt the bickering :rolleyes: but thought some of you might be interested in this littel quote re: Doom3's release

From gamespot (http://gamespot.com/gamespot/stories/news/0,10870,2877682,00.html)

"Id Software has said that work on the Xbox version won't start until the PC game is complete. At QuakeCon 2002, id expressed hope that the game would be done before next year's E3 in May. For more information, check out our previous coverage of the PC game."

Ok, so I was off by about 6 weeks. ;)
 
jjayb said:
Heh, no spin involved. I didn't include any "uh's" or "uhm's" he didn't say and didn't do it to make him look evasive. Just trying to be exact. Would be kinda dumb for me to get on you for not including everything, then doing it myself. I've never transcribed anything before. Took me forever cause I kept replaying the call and seeing that I'd missed an "uh" or "uhm" here and there. ;) Now I Know I didn't need to waste my time. (it does really annoy me when people "aaaahhhh" and "uhhmmmm" alot though. Especially reporters or people speaking publicly) Mistakes happen. Although accidental it did ultimately bend the truth. Sorry for being harsh. At the time I didn't know if it was on purpose or accidental. Thanks for changing it.



LOL I did the exact same thing ove at the [H]ardOCP fourms. Those "
uh's & ah's are a muther.... "(in Isacc Hayes like baritone) shut yo mouth....I'm just talkin' about nvida y'all:)
 
Back
Top