nVidia Q3: Revenue up 53%

[witeken]

Yet,you also forget that the enthusiast chips have not increased in price by anywhere as much.

An example:
1.)Core i7 980X - 239MM2. RRP:$999
2.)Core i7 5820K - 356MM2. RRP: $389
3.)Core i7 6800K - 246MM2. RRP@ $434

So the six core chips have plummeted in price.

Now if we want to make Intel look worse,the RRP from maximum enabled chip to maximum enabled chip has gone up from $999 to $1723,ie,Core i7 980X to Core i7 6950X,over 6 and a half years.

But thats the thing - the Pascal Titan X is a second tier salvage product of a 500MM2 GPU which costs $1200. The second tier salvage product of the GTX580 generation which was of a 500MM2 GPU was the GTX570 which cost $330,

So that is a 3.64X increase in RRP. Intel looks rather like a charity in comparison.

Hence,my point still stands,Nvidia has been successful in marketing that price increase,and good for them. What I find is funny PCMR enthusiasts seems so much in denial how much they are cash-cows now for other areas,just like Intel used enthusiasts as a cash-cow for Atom chips,which were sold at cost or a loss.

I generally agree with your post. The thing here is that high-end and midrange is, from a companies' perspective, always relative to the competition. If Nvidia's -- I don't know -- say 1070 is as good as AMD's best card on the market, in their eyes the next tier (1080) will have premium performance compared to the market/competition, so they can pick and chose any price they want consumer to pay for that premium performance. Nvidia has done that for a long time to charge a lot for their best product, so no sherlock I would say. They just don't call it a Titan now, I guess.

If you add in the cost peculiarities of the 16nm node (and if you live in Europe the anti-democratic quantitative easing), you can certainly get some interesting pricing structures, irrespective of the underlying product characteristics like die size which are invisible to the consumer.

Now if someone were to make a performance per dollar comparison like the die size discussion .

 

[happy medium]

680/980/1080 were never high end

come on man, you cannot be serious.

"NVIDIA GeForce GTX 680 Review: Retaking The Performance Crown"
http://www.anandtech.com/show/5699/nvidia-geforce-gtx-680-review

"At the very high end the GTX 980 will be unrivaled
http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review

"The GeForce GTX 1080 logically succeeds the GeForce GTX 980 and is priced not too far apart from its predecessor. It will launch at $599 where the GTX 980 debuted at $549. NVIDIA, however, is positioning it as a premium part since it's supposedly faster than any currently available single-GPU card.
https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_1080/

That's my definition of high end as well as every other reviewer on earth.

 

[esquared]

I am going to lock this thread for a while. nearly everything over the last several pages has nothing to do with the 3rd quarter earnings.

I may open it later in the day. And, if/when this is opened any further off topic posting will be met with infractions and possible removal from VC&G.


esquared
Anandtech Forum Director

 

[esquared]

Opened again and moved to the NVidia subforum.

Keep post on topic.


esquared
Anandtech Forum Director

 

[SlickR12345]

Nvidia has great profits only because they've been selling overpriced cards that are supposed to be mid range, but are selling as high end. They release the $350 version as the high end, so the GTX 980 was $550, then wait for the suckers to buy them up and several months later release the 980TI which is the true high end. Of course the 980ti is also cut from the Titan X, which means 980 is literally 2 tiers bellow as a cut down card.

If the Titan X is the high end part, based on history it should cost around $500-600, the 980TI as a cutdown, but very close should cost $400-450 and the 980 is the higher mid range that should cost around $250-300

 

[Qwertilot]

Well the poor mods tried

The whole thing is a bit silly anyway of course - NV are producing a very consistent +~30% performance increase year on year. Both top performance and performance/price (although that sometimes slips on to a two year schedule.).

You buy when the value proposition is tempting on a personal level, or not if it isn't. Really hard competition seemed to produce minor bumps but no dramatic variation on the basic schedule.

 

[StrangerGuy]

Nvidia has great profits only because they've been selling overpriced cards that are supposed to be mid range, but are selling as high end. They release the $350 version as the high end, so the GTX 980 was $550, then wait for the suckers to buy them up and several months later release the 980TI which is the true high end. Of course the 980ti is also cut from the Titan X, which means 980 is literally 2 tiers bellow as a cut down card.

If the Titan X is the high end part, based on history it should cost around $500-600, the 980TI as a cutdown, but very close should cost $400-450 and the 980 is the higher mid range that should cost around $250-300

"You see, life-critical stuff like viddy cards should be priced at what I want instead of what the market will bear...nope, totally no entitlement syndrome here"

 

[krumme]

In end oct 2004 gross margin was 32%
In end oct 2016 gross margin was 59%

Improving over the years.
https://ycharts.com/companies/NVDA/gross_profit_margin

In case anyone still wonders where profit comes from. If this data doesnt solve this discussion no data og logic will.

Data for this Date Range - Gross margin for NV
Oct. 31, 2016 59.03%
July 31, 2016 57.84%
April 30, 2016 57.55%
Jan. 31, 2016 56.46%
Oct. 31, 2015 56.25%
July 31, 2015 54.99%
April 30, 2015 56.73%
Jan. 31, 2015 55.88%
Oct. 31, 2014 55.18%
July 31, 2014 56.12%
April 30, 2014 54.76%
Jan. 31, 2014 54.15%
Oct. 31, 2013 55.45%
July 31, 2013 55.82%
April 30, 2013 54.32%
Jan. 31, 2013 52.90%
Oct. 31, 2012 52.87%
July 31, 2012 51.78%
April 30, 2012 50.10%
Jan. 31, 2012 51.41%
Oct. 31, 2011 52.22%
July 31, 2011 51.67%
April 30, 2011 50.36%
Jan. 31, 2011 48.10%
Oct. 31, 2010 46.46%

July 31, 2010 16.55%
April 30, 2010 45.56%
Jan. 31, 2010 44.65%
Oct. 31, 2009 43.38%
July 31, 2009 20.18%
April 30, 2009 28.56%
Jan. 31, 2009 29.44%
Oct. 31, 2008 40.98%
July 31, 2008 16.79%
April 30, 2008 44.64%
Jan. 31, 2008 45.70%
Oct. 31, 2007 46.21%
July 31, 2007 45.33%
April 30, 2007 45.03%
Jan. 31, 2007 43.89%
Oct. 31, 2006 40.70%
July 31, 2006 42.49%
April 30, 2006 42.35%
Jan. 31, 2006 40.21%
Oct. 31, 2005 39.08%
July 31, 2005 37.82%
April 30, 2005 35.99%
Jan. 31, 2005 33.87%
Oct. 31, 2004 32.34%
July 31, 2004 30.72%

 

[Qwertilot]

You need to split it out by market sector and things - all of the compute/professional etc stuff they sell will have rather higher margins than the cards for the domestic gaming market.

 

[krumme]

You need to split it out by market sector and things - all of the compute/professional etc stuff they sell will have rather higher margins than the cards for the domestic gaming market.

No i dont have to spilt into "things" to go from 30 to 59%.
The development is so extremely high it filters that out even if - and that is IF - eg datacenters is even high margin.

Seriously why do we have to lisnt to all this nonsense?. Profit is derived from margins. Gaming consitutes 1.2 of 2b.

On this earth besides anandtech gpu nv forum a price of product is defined by supply and demand and if a manufacturer can raise prices to get more profit they are expected to do so. Its called business and everyone would do it.

But no. When rs says a 570 is priced different than 1080ti, loads of bs hits this forum. Because no. Prices of nv gaming gpu have not risen. Its because fin fet is so expensive and professional market rakes in all the margins. And besides its just a super duper new highend segment - just look at reviews.

What the f is wrong with you guys? The obvious answer to rs is just "so what? Everyone would do that".

Instead we have this endless stream of crap.

Damn fine financial result btw but they have always been that except a few years. But they dont have a machine printing money in the basement.

 

[Qwertilot]

Probably because it isn't rubbish They might have upped pricing on the gaming GPUs, I dunno.

That will have moved it from say 50 to 59 though. The move from @~30 to 50 is because they've got huge chunks of sales/revenues in other markets with much higher margin/selling price than gaming GPU stuff.
(The pricing on pro/data center stuff surely makes it quite clear that the margins are higher )

Those other markets also, perhaps most importantly, lets them spread a big chunk of those massive development costs over many more sales.
(Although the data center stuff is big enough now that they're devoting a lot of effort to it in its own right.).

 

[krumme]

Yes its 100% pure fanboy rubbish and nothing else

Datacenters is so high growth and a rising star that the normal assumption would be that it hurts margins. Development of the market cost money. Always have been. No miracles. Besides there is no natural law saying professional markets is more profitable than consumer. Some people thinks so. But a big consumer brand is generally far more valuable and able to rake in profit as its generally more scalable.
Coca Cola, Apple and Nvidia is good examples of that within their different products. Everything in ordnung.

Probably a 1080ti is a good deal more expensive to develop than a gtx 570 due to finfet and dev cost, but no not 300%

Fortunately there is a straightforward explanation to the profit -> Margins.

Edit. Btw. You just cant look at pricing to define margins. There is a cost side too but i think you mind surpressed it

 

[Qwertilot]

Until Pascal the extra cost of the data center/AI stuff was - relatively to their massive main R&D costs! - fairly small.

With Pascal its getting bigger of course: special chips, 8 bit modes etc etc. They don't actually have huge competition there yet, although Intel are starting to ramp it up.

The other thing that has happened vs 2004 is many more notebook GPU sales. Those really do have a high premium, and they've had nil competition there for quite a long time now.

 

[krumme]

Yes. There is plenty profitabiliy in mobile and have been for years. Even during amd 5000 series days nv had optimus technology to give a better solution even if 400 and 500 was big and power inefficient vs 5000 series. Just goes to show they know what technology matters for consumers.

They got their advantage because of real value and the mix of good marketing and brand development.

 

[Pariah]

In end oct 2004 gross margin was 32%
In end oct 2016 gross margin was 59%

Improving over the years.
https://ycharts.com/companies/NVDA/gross_profit_margin

In case anyone still wonders where profit comes from. If this data doesnt solve this discussion no data og logic will.

Great analysis there. Cutting and pasting 12 years worth of data without any attempt to state how this proves anything you're trying to say.

It's much easier for everyone to look at the last 5 years which is all the anti-Nvidia posters comment on anyway when bitching about Nvidia's pricing strategy since that's the period where Nvidia changed their product stack strategy and pricing.

So here is Q3 2001 through Q3 2016:

Oct. 31, 2016 59.03%
July 31, 2016 57.84%
April 30, 2016 57.55%
Jan. 31, 2016 56.46%
Oct. 31, 2015 56.25%
July 31, 2015 54.99%
April 30, 2015 56.73%
Jan. 31, 2015 55.88%
Oct. 31, 2014 55.18%
July 31, 2014 56.12%
April 30, 2014 54.76%
Jan. 31, 2014 54.15%
Oct. 31, 2013 55.45%
July 31, 2013 55.82%
April 30, 2013 54.32%
Jan. 31, 2013 52.90%
Oct. 31, 2012 52.87%
July 31, 2012 51.78%
April 30, 2012 50.10%
Jan. 31, 2012 51.41%
Oct. 31, 2011 52.22%

Not so interesting looking, is it? Over the last 5 years, there is remarkable consistency with every quarter falling in the 50's. A 6.81% increase in 5 years, or 1.36% per year is hardly earth shattering.

Now, let's take a closer look at the numbers above and see why the last 5 years is all we need to look at. The number one complaint about Nvidia's pricing scheme that is carpet bombed throughout this forum is Nvidia's decision to move mid size die prices up to previous big die prices. The above 5 year stretch begins between the releases of the $500 big die Fermi 580 in November 2010 and the $500 medium die Kepler 680 in March 2012 when the strategy actually began. Based on all the venom spewed at Nvidia about how this strategy was simply a case of ripping off the consumer, you would expect to see a dramatic increase in margins after the release of the 680. Instead, the quarter in which the 680 was released (April 30,2012), margins were actually lower than they were the previous quarter, followed by an increase of about 1.7% in the next quarter which does little to support the narrative that Nvidia is arbitrarily increasing prices as opposed to raising them in line with production costs.

We see more of the same in the following quarters including the just released Q3 2016 results. Based on the price increases across the board for the mid to halo 1060/Titan XP and the release of high margin Pascal mobile products in August and no low end releases to drag down the margin, you would once again expect to see a pretty substantial gain in margin. A huge gain in revenue makes it clear they sold a whole lot of these new high end/high margin GPUs, yet under this ideal product stack release, margins only improved by 2.57% from Q3 2015.

Combine all of the above with the rapid decrease in market share for low margin sub $100 GPU's in recent years due to Intel's much improved IGP's and AMD's APU's and there is no data to support that Nvidia is making a higher margin (ie screwing all of us) off their GPU's than they were 5 years ago despite charging more.

 

[krumme]

and there is no data to support that Nvidia is making a higher margin (ie screwing all of us) off their GPU's than they were 5 years ago despite charging more.

Well you cant make up your mind if nv is doing business to please you? - and therefore is "screwing all of us" by raising prices.
So you made up your mind that they dont raise price contrary to what all other private business on this planet is expected to do to do best. Nv just delivered a slam dunk exceptional result with record margins and they did so without raising prices?

To that purpose you can cherrypick to fit this perspective. And it is the definition of fanboy perspective if thinking a companys purpose is not to earn money and that it by raising prices is "screwing" customers. Seriously is amd or apple screwing customers if they raise prices?
No it dont screw customers. Its taking advantage but the customer can go somewhere else and even buy a camera or an icecream if they think the 1080ti is to expensive. Its their choice. The company can complain their customers is stupid or the customers can and regulary do complain about prices. From icecreams to gfx. Thats how it is but every part is expected to do what is best for themselves on the market. In that way the market is actually working best. So its not screwing its making the market work best.

This is an endless debate because there is millions reasons that eg cost have risen and as many reasons margins and portfolio dont follow perfectly. You can start here and write to the moon. If the thinking - and its a feeling and not logic - is nv is screwing customers by raising prices its clearly difficult to accept the price difference between 570 and 1080ti as not correlated to margins and profit despite die size and record margins screaming through the roof.

 

[AtenRa]

You have to take Margins when first Fermi GF100 was released on a new 40nm node , that is Q2 2010. When GTX 580 was released Q4 2010, 40nm manufacturing was already in production for a year or more thus wafer prices were lower and yields were higher making COGs go lower.

* Quarters taken when each card was already in retail, actual release of the cards could be on the previous Quarter.

Oct. 31, 2016 59.03% - new 16nm GP102 (471mm2) TITAN X Pascal = $1199
July 31, 2016 57.84% - new 16nm - GP104 Pascal release (GTX 1080) = $599-699
April 30, 2016 57.55%
Jan. 31, 2016 56.46%
Oct. 31, 2015 56.25%
July 31, 2015 54.99% - GTX 980Ti release = $649
April 30, 2015 56.73% - GM200 release TITAN X = $999
Jan. 31, 2015 55.88%
Oct. 31, 2014 55.18% - GM204 Maxwell release (GTX 980) = $500
July 31, 2014 56.12%
April 30, 2014 54.76%
Jan. 31, 2014 54.15%
Oct. 31, 2013 55.45%
July 31, 2013 55.82% - GK110 release (GTX 780) = $649
April 30, 2013 54.32% - GK110 TITAN release = $999
Jan. 31, 2013 52.90%
Oct. 31, 2012 52.87%
July 31, 2012 51.78%
April 30, 2012 50.10% - new 28nm - Kepler GK104 (294mm2) release (GTX 680) = $500
Jan. 31, 2012 51.41%
Oct. 31, 2011 52.22%
July 31, 2011 51.67%
April 30, 2011 50.36%
Jan. 31, 2011 48.10% - GF110 release (GTX 580) = $500
Oct. 31, 2010 46.46%

July 31, 2010 16.55%
April 30, 2010 45.56% - new 40nm - Fermi GF100 (529mm2) release (GTX 480) = $500
Jan. 31, 2010 44.65%

Margins from Q2 2010 (Fermi) to Q3 2016 (Pascal) has increased by almost 30%
Q2 2010 to Q3 2016 Cost of High-End single Chip Flagship Graphics Cards has increased by 2,4x from $500 to $1199

When we used to have 50% of Performance increase vs the old Flagship GPU on a new manufacturing Node at $500 (GTX 480 vs GTX 285) ,
Today we have 30% of Performance increase vs the old Flagship GPU on a new manufacturing Node at $599-699 (GTX 1080 vs GTX 980Ti ) ,
Or the new Flagship price has increased by 2.4x to $1199 (GTX 480 vs Titan X Pascal)

 

[SirCanealot]

never seen a consumer fighting for paying more, it's absurd.

Sorry to contribute to the somewhat OT nature of this part of the discussion, but I feel very strongly about this.

I've made this point several times that I find these attitudes actually frightening. But I guess this is exactly what companies want — people who will defend them as they try to extract more capital from their customers whilst retaining customer loyalty. I would advise everyone to really think about what you are arguing about and think about what you are saying.

NVidia are free to call whatever they want 'high end'. And we are free to call them out for it if we do not feel that their products are 'value for money' any more. (disclosure: I got a 1080 recently. I kind of do not feel good about it. But it's awesome, lol)

PS: I do semantical arguments on hire, only $40 an hour, so if you want to continue this in PM I can go for a few hours on this if you want ;P

 

[railven]

Really is interesting to see the numbers of just how lucrative NV's business model is. End of the day it's rather impressive. They dominate the market and thanks to their huge profits they can spend it on marketing and branding which just continues the machine going.

I just peaked at their stock prices, hachimachi. I wonder how things would have been had Hector not let his ego block the NV Merger/buyout and Huang were at the helm.

Probably would have lend to a counter of Intel+ATI! DAMN YOU Hector!!! You ruined it for all of us!!!

 

[AtenRa]

Really is interesting to see the numbers of just how lucrative NV's business model is. End of the day it's rather impressive. They dominate the market and thanks to their huge profits they can spend it on marketing and branding which just continues the machine going.

I just peaked at their stock prices, hachimachi. I wonder how things would have been had Hector not let his ego block the NV Merger/buyout and Huang were at the helm.

Probably would have lend to a counter of Intel+ATI! DAMN YOU Hector!!! You ruined it for all of us!!!

With margins near 60%, Intel may find dGPUs viable again

But if AMD went with NV back in 2006 and Intel would had bought ATI, Intel-ATI would dominate both Desktop/Laptop APUs AND dGPUs and no matter who would be on the AMD + NV wheel they would eventually loose big time. Intel could spend huge amounts of resources for GPU R&D and with the manufacturing lead it had vs TSMC/GloFo at the time they would literally killed AMD+NV by now.

Just image Intel+ATI at 32nm vs AMD+NV at 40nm TSMC and especially when Intel had 22nm finfets in 2012 AMD+NV only had 28nm bulk.

 

[nvgpu]

Keep on dreaming, Nvidia's Tesla architecture is far superior to AMD's inferior VLIW architecture. No amount of process tech can help overcome the inefficiencies of VLIW.

Also unlike all AMD CEOs which are inept and unable to read where technology is going, Jen-Hsun has steered Nvidia to greater and greater heights and the financial results shows.






Trolling and inflammatory posting is not allowed.

esquared
Anandtech Forum Director

 

[AtenRa]

I dont want to derail but AMD doesnt use VLIW since almost 5 years ago when they released GCN on January 2012.

 

[nvgpu]

You were talking about 2006, also GCN is garbage when it comes to efficiency of shader usage and running DX11 and OpenGL games as anyone can tell you from the problems GCN has with DX11 games.

Nvidia runs DX11, DX12, OpenGL and Vulkan games just fine without needing the crutch that is DX12 async.

 

[krumme]

Sorry to contribute to the somewhat OT nature of this part of the discussion, but I feel very strongly about this.

I've made this point several times that I find these attitudes actually frightening. But I guess this is exactly what companies want — people who will defend them as they try to extract more capital from their customers whilst retaining customer loyalty. I would advise everyone to really think about what you are arguing about and think about what you are saying.

NVidia are free to call whatever they want 'high end'. And we are free to call them out for it if we do not feel that their products are 'value for money' any more. (disclosure: I got a 1080 recently. I kind of do not feel good about it. But it's awesome, lol)

PS: I do semantical arguments on hire, only $40 an hour, so if you want to continue this in PM I can go for a few hours on this if you want ;P

Its not OT because brand is a vital part if nv profit. They have used years cleverly developing it.

What is special is we have educated informed users reacting like i have not prior seen. A normal response for a strong brand would be that its "worth the premium" or simply "the best". Instead we have a company raising prices and the informed customers react by saying they did not raise prices.

If consumers stop acting in their own interest and get imune to price the market will simply stop working most efficient. Its the equivalent of the companies not optimizing their profit. Now one can say most marketing and brand development goes towards monopolizing the market but there is also an information part of it. Selecting a brand as a consumer gives you a known quality and thats a quality in itself as it eg saves a lot of time and minimizing risk is a value. But clearly we are outside the optimal if informed consumers can not see the most obvious. Goes to show the strenght of the brand.

 

[Qwertilot]

With margins near 60%, Intel may find dGPUs viable again

Well they're going off the places with the biggest margins with the xxx landings The gaming GPU market is probably a bit hard to get into.