NVIDIA shows signs ... [2008 - 2017]

[Speccy]

Yes marketshare is calculated by shipped but a write off means no money was paid for those chips, acutally those write off chips are probably not even sent out the door, there would be no use since no one makes money on them and nV or AMD would rather make money then just send out discarded chips.

An inventory writeoff doesn't mean that no money was paid - it means the opposite; in financial terms it means that they feel that the money that was paid won't be re-couped.

Once a write-off or write-down of inventory has occured rarely would you scrap the goods that are writen off (unless its just completely useless). All it means is that you have financially taken the cost off the goods off your books, but you still physically have the goods - in which case you can both sell them for any price and they will effectively increase your overall margin (in subsequent quarters that they are sold) because as you've already written the costs off they are pure margin when they are sold.

 

[Razor1]

An inventory writeoff doesn't mean that no money was paid - it means the opposite; in financial terms it means that they feel that the money that was paid won't be re-couped.

Once a write-off or write-down of inventory has occured rarely would you scrap the goods that are writen off (unless its just completely useless). All it means is that you have financially taken the cost off the goods off your books, but you still physically have the goods - in which case you can both sell them for any price and they will effectively increase your overall margin (in subsequent quarters that they are sold) because as you've already written the costs off they are pure margin when they are sold.

Tech industry doesn't work that way, they would rather take the write off and scrap the parts. Because they won't make money on them, they deperciate too fast, within a year a high end graphics card is worth half, 2 years its worth less then 1/5. Its better for them to actually take the tax break from the write off then actually try to sell the parts. Corporate tax state and fed will end up in 55% range, that 200 million write of nV did last quarter effectively gave then ~90 million tax break. Now last quarter the write off was for gf8 lines most likely, thats already a 2 year chip, they won't make money on those cards now.

http://www.sfgate.com/cgi-bin/artic.../archive/2001/05/08/BU54900.DTL&type=business

But instead of scrapping or selling it, Cisco is cordoning it off in "secure" areas of its facilities. If things pick up, Cisco could use this inventory at little or no cost, which might fatten its future gross profit margins.
Using the previous example, let's say the part that was written down from $100 to $10 is later used in a new product that is sold for $500. The gross profit (excluding other costs) would be $490, compared with only $400 if the part had not been written down.
Accounting rules permit this. There is no rule that says a company can't later use or sell inventory that has been written off. And a company generally doesn't have to disclose the fact that it's using parts that have been written down or off.
Tax rules are different. A company generally cannot take a current tax deduction for inventory that has been written off if it's still on hand. "To get a tax deduction for inventory that is obsolete, you have to either get rid of it or prove that you have made substantial efforts to sell it and you just can't," says Reynolds.

two colored areas you are correct but if they take that tax break they can't use those parts

 

[OpenGL guy]

oh yeah but that comes out of the empolyee's salary (taxes that are withheld from paychecks to the employee), the corporation doesn't pay taxes on the money made on that portion though.

Incorrect. As I said, companies pay payroll taxes on employee salaries. Yes, the employee also pays "payroll" taxes on the salary (social security tax). The payroll tax is split between the employee and employer.

 

[Razor1]

Incorrect. As I said, companies pay payroll taxes on employee salaries. Yes, the employee also pays "payroll" taxes on the salary (social security tax). The payroll tax is split between the employee and employer.

no income tax isn't, only 2 taxes that are payed by employer SS taxes which is ~6% (total 12% it is split between the empolyer and employee) and medicaid taxes which are around ~1% (2% split evenly).(fica)

 

[Arun]

Wow, that Charlie article certainly is one of the most laughably inaccurate ones I've seen from him in some time. Let me make this clear: I agree NV's financial position is a joke right now and I'd argue based on short-term earning estimates and my global economic forecasts even their current share price is too high, but that doesn't mean Charlie should be able to just spout random bullshit. I probably shouldn't have bothered, but here we go:

The same story applies to Tegra. It was supposed to have products using it out about three quarters ago

Tegra was never supposed to have a product out before about November, which is two quarters ago. One of those was a PND for example because those have shorter design cycles, but I suspect it might have been canceled completely. Frankly many handheld semiconductor companies are saying publicly that they are witnessing unusually many customer delays and cancellations in the industry, it's not just NV.

Who knew? Ion is not only a brand now, but it encompasses any Geforce integrated GPU. Wow.

It would incredibly clear to anyone who knows anything about NV's MCP business (hint: not Charlie, wasn't MCP division supposed to be gone by now?) that he's not encompassing any GeForce integrated GPU. He's simply describing Ion as MCP79, aka GeForce 9300/9400 IGPs for Intel. Other IGPs aren't included! Any analyst who doesn't know Atom-based IGPs aren't shipping much yet clearly isn't earning his paycheck. So I don't see where the problem is.

Remember, this box has twice the CPU power of the one in the Revo.

So now dual-core automagically halves CPU bottlenecks? How cute. Certainly the Atom 230 can be a major bottleneck, but I have never seen any indication that HW-accelerated HD video actually benefits from dual-core, at least in Cyberlink. It just needs enough CPU power, so Atom 330 numbers should be nearly identical to Atom 230 if you don't multitask.

A much better explanation is what we have been hearing from OEMs for a couple of quarters now, that Nvidia seems to be addicted to marketshare and will try to maintain it even at the cost of cash losses.

You can't dismiss the other explanations unless they are actually wrong. This explanation is probably correct as a complementary one though.

What it is not telling you is that marketshare growth was mainly due to the company dumping about a million units of 65nm DX9 inventory that it had previously written off.

Assuming Charlie meant G9x instead of DX9 (did G78 even exist?) that is still a gross oversimplification. If it's inventory dumping, how do you explain ASPs weren't really down? They maintained share in the low-end and increased in the high-end (+IGPs), not the other way around. Certainly excess legacy (i.e. 65nm) inventory had some margin impact - but how is that unusual in such a dramatic demand falloff? It's not like AMD hasn't had any inventory write-offs since September either...

Using the numbers that Nvidia did give out, the totals don't add up unless GPU margins were about half of the overall corporate margins. That is bad. Bad bad bad. Your core business should not be struggling to have double-digit margins.

His calculations are too pessimistic for GPU and too optimistic for the rest, although he's not completely wrong here at least.

You've got the IGP business with ~32% margins at 28% of total revenue. You've got PSB with ~65% margins at 16% of total revenue. Then you've got GPU at 53% of total revenue and unknown margins. And finally 3% that don't add up, presumably bad rounding+royalties+contract R&D. This gives us a gross profit of ~$60M for MCP and ~$70M for PSB - which leaves about $70M gross profits for GPU since the total is ~$203M. Then to get GPU margins: 70/(664*0.53) = ~0.199 -> ~20% gross margins.

That is certainly *not* good, and I would assume *consumer* GT200b gross margins to be even lower (wouldn't exclude zero or negative, ugh!) but I wouldn't call that "struggling to have double-digit margins".

Lose money on every one sold but make it up in volume is not a healthy long-term corporate strategy

I guess Charlie hasn't yet figured out that it's harder to recover market share with new products than it is to recover gross margins with new products. Companies don't get addicted to marketshare just for the free bonus miles.

To use his exact words: "Let's see... the ramp is going fine." Bullsh*t. [...] Nvidia was supposed to beat ATI out to market with 40nm parts. Nvidia was slated for November, ATI for December.

Of course they failed to meet their ETAs, but that doesn't tell you anything about what the *ramp* looks like *right now*, i.e. how many products NV is churning through the production line and at what yields - which is, y'know, what Jen-Hsun talked about. Oh well, can't miss the opportunity to call someone a liar even when it makes no sense I guess...

The now dead GT212 was a die-shrink of the GT200b from 55nm to 40nm

You still have no idea what NV's 40nm roadmap was, now do you Charlie?

What I suspect is going on is that Nvidia hasn't been able to get its 40nm chips up to an economically viable yield.

How cute. Wasn't it TSMC themselves that said their 40nm yields weren't meeting their goals in their quarterly conference call? TSMC's definition of yields should be based mostly on test structures, this has nothing to do with NVIDIA screwing anything up. Of course, you won't learn that 40nm yields were low (along with a high wafer cost) for the RV740 ramp from your AMD sources... I'm not saying it wasn't still more cost-effective than an equivalent 55nm part, but let's be realistic here.

The GT214 / 216 / 218s are basically GT200 derivative parts, but we are also hearing that they are shrunken G92s.

And you still have no idea what NV's 40nm roadmap is right now, now do you Charlie?

The problem here is that the 214 to 215 evolution took so much die area that it is unlikely to ever be profitable.

And you really have absolutely no idea whatsoever what the hell GT215 might actually be, now do you Charlie?

Let me repeat that, Nvidia will have no parts out for either the Windows 7 launch or Christmas, so it is all ATI this year.

There's a reasonable chance Charlie is right there, but I'd love to watch him eat a hat if he's not given how clearcut that claim was.

How they can spin this as good is beyond us, especially if you know the performance levels of both parts.

Oh, right. Like Charlie knew R600's performance... Can't help but wonder if he even changed sources since then!

If all goes well, Nvidia can hope to start competing for marketshare again next spring. Until then, things are only going to get worse for the green team.

So they'll be magically have a chance to competitive next Spring based on GT300 derivatives, but GT300 itself is not going to be competitive with RV870? Intriguing! Sigh...

Nvidia's executives can spin the situation all they want, but all the analysts I talk with understand the credibility gap the company has by now and seems bent on expanding.

For once, I kinda agree with Charlie; the same is true for AMD or Intel though, forecasts generally aren't very valuable in this industry. But of course, he is himself even less credible than they are...

 

[Jawed]

You've got the IGP business with ~32% margins at 28% of total revenue. You've got PSB with ~65% margins at 16% of total revenue.

Where did these margin numbers come from?

Jawed

 

[Arun]

Where did these margin numbers come from?

Ion/MCP79 margins were said to be "About or better than corporate average" during this CC, so that's >30.6%. I very optimistically assumed that other MCP products had similar or slightly better margins so that the GPU margins would be conservative (i.e. on the low side) - as for PSB, it was from my Analyst Day notes iirc, where they had a graph of it hovering between 65 and 70%. And I obviously doubt they went up since then...

 

[Speccy]

Tech industry doesn't work that way, they would rather take the write off and scrap the parts. Because they won't make money on them, they deperciate too fast, within a year a high end graphics card is worth half, 2 years its worth less then 1/5.

Normally you'd rather sell them. These types of writedowns are often done because a quarter is looking bad so you throw the baby out with the bathwater in that quarter in order to make subsequent quarters look better.

Now last quarter the write off was for gf8 lines most likely, thats already a 2 year chip, they won't make money on those cards now.

GF8 spans a lot of lines. If the write-off was for low end GF8 parts, which aren't as tech or performance sensitive, then its very easy to offload them dirt cheap into the market for cheap IGP upgrades and have a positive effect on marketshare.

In the CC they openly stated that both inventory witten down and inventory written off in Q4 were sold this last quarter (though admit they had not written it down enough):

http://seekingalpha.com/article/136355-nvidia-f1q10-qtr-end-4-26-09-earnings-call-transcript?page=-1

Daniel Berenbaum - Auriga

Good afternoon, guys. Can we talk a little bit more about the inventory -- can you talk -- can you give us a little more detail about were there any inventory write-downs and did you sell through any -- or you mentioned you sold through some previously reserved inventory. Can you quantify a little better what the write-offs were or how much of the reserves that there were that you sold through with essentially zero cost, and how much still had some cost associated with it?

David White

We have actually a couple of situations -- we have inventory that we completely wrote off in a prior period. We also have inventories that we wrote down to market value. Some of those inventories in both categories were sold during the quarter, some of which we had gains on, some of which we had losses on. It turns out that the losses we experienced on the sale of that inventory plus some of the other products actually offset that by a greater amount, and so on a net basis, it was actually a negative impact to the quarter.

Daniel Berenbaum - Auriga

And was that basically because you, of the amount that you had written down, you hadn’t written it down enough?

David White

That’s correct.

 

[Jawed]

Ion/MCP79 margins were said to be "About or better than corporate average" during this CC, so that's >30.6%.

http://forum.beyond3d.com/showpost.php?p=1291980&postcount=626

They're referring to an historical corporate average margin. I don't know what that is, and whether they're referring to MCP-specific historical average or NVidia's historical average. With the continued growth of MCP sales in absolute terms it is presumably enjoying higher margins now than in prior quarters as NVidia captures more share in the Intel market (i.e. beyond Apple shores). Although, to be honest, I don't know the MCP revenue picture, historically.

Jawed

 

[Arun]

They're referring to an historical corporate average margin.

I can't remember for sure from when I listened (I could recheck if you want), but the seekingalpha transcript seems to say otherwise:

Rajvindra Gill - Needham & Company
And any sense of what the gross margins or ASPs on that platform were in this quarter?

Jen-Hsun Huang
About or better than corporate average.

Rajvindra Gill - Needham & Company
The corporate average of this quarter of 30.6% or just what you have done historically?

Jen-Hsun Huang
Better than this quarter.

 

[Jawed]

"Better than this quarter" meaning "not this quarter's average, but better (historical)". It's definitely not "this quarter's average".

Obviously, in trivial terms, it's >30.6%. I just don't know what the corporate historical average is (would be fairly tedious to compute, too) which is why I guessed with 40%, so just wondering why you went with 32%...

Jawed

 

[OpenGL guy]

no income tax isn't, only 2 taxes that are payed by employer SS taxes which is ~6% (total 12% it is split between the empolyer and employee) and medicaid taxes which are around ~1% (2% split evenly).(fica)

Yes, this is called a "payroll tax" as I mentioned in my previous two posts. It is a tax, right? I refer you to Wiki. Take careful note of the first sentence. Also note that employees have a maximum amount that they are taxed for social security, but I don't believe that cap applies to employers.

 

[trinibwoy]

"Better than this quarter" meaning "not this quarter's average, but better (historical)". It's definitely not "this quarter's average".

I think my english is pretty good and it's obvious he was referring to the current quarter's average. It's crystal clear in that snippet Arun quoted.

 

[Jawed]

I can see it. Obviously wasn't crystal clear to me

Jawed

 

[DegustatoR]

There's a reasonable chance Charlie is right there, but I'd love to watch him eat a hat if he's not given how clearcut that claim was.

Why is it important to have anything for Win7 launch at all? Does Charlie now that DX11 isn't tied to Win7? There probably won't be even synthetic benchmarks for DX11 until sometime in 2010 and games will follow even later than that (mostly in the form of DX9 console ports btw).
What's important is to switch to 40nm in low margins segment by the end of the year - and i'm thinking that NV will do that pretty soon. But if they'll do that pretty soon then what'll stop them from launching G3x0 GPUs in the 2nd half of the year? Charlie's articles?

 

[XMAN26]

http://www.theinquirer.net/inquirer/news/1137286/apple-dell-hp-owners-sue-nvidia

when it rains it pours for the G86/84M screw up.

 

[Scali]

Why is it important to have anything for Win7 launch at all? Does Charlie now that DX11 isn't tied to Win7? There probably won't be even synthetic benchmarks for DX11 until sometime in 2010 and games will follow even later than that (mostly in the form of DX9 console ports btw).

Yea, and aside from that, DX11 isn't even directly tied to hardware either.
Some of the most compelling features, like multithreaded rendering and Compute Shaders work fine on nVidia's current product line. In fact, the Compute Shaders may help nVidia to make their GPGPU products even more popular.

Now unless ATi has DX11 hardware planned at (or before) the Win7 release, I see no reason why nVidia would have to have it.

 

[Jawed]

Yea, and aside from that, DX11 isn't even directly tied to hardware either.
Some of the most compelling features, like multithreaded rendering and Compute Shaders work fine on nVidia's current product line. In fact, the Compute Shaders may help nVidia to make their GPGPU products even more popular.

Did you miss the bit where they work on ATI's current products?

Jawed

 

[Razor1]

Yes, this is called a "payroll tax" as I mentioned in my previous two posts. It is a tax, right? I refer you to Wiki. Take careful note of the first sentence. Also note that employees have a maximum amount that they are taxed for social security, but I don't believe that cap applies to employers.

but its not a tax on the money the corporation makes. Payroll tax is also concidered a business expenditure.

 

[Scali]

Did you miss the bit where they work on ATI's current products?

Jawed

There we go again. Would you mind controlling your ATi-fanboyism for once, and not go ranting everytime someone 'forgets' to credit ATi in a topic that isn't about ATi in the first place? (Or the other way around for that matter... Like last time when we were in a topic about Avivo, and you started ranting how I didn't criticise Badaboom in that thread?).
Now if I said that ATi *didn't* support them, perhaps that would give you the right to start a rant. But I didn't. I just didn't think it was necessary to mention ATi, since we were discussing nVidia here. Besides, I think most people here know that any DX10/DX10.1 parts will support it. Oh wait, that includes S3 aswell (and even Intel's IGPs?). Did you miss that bit?
Please...