Friday, February 15th 2019

NVIDIA Reports Q4 and 2018 Financial Results - Down 45%

NVIDIA today reported revenue for the fourth quarter ended Jan. 27, 2019, of $2.21 billion, down 24 percent from $2.91 billion a year earlier, and down 31 percent from $3.18 billion in the previous quarter. GAAP earnings per diluted share for the quarter were $0.92, down 48 percent from $1.78 a year ago and down 53 percent from $1.97 in the previous quarter. Non-GAAP earnings per diluted share were $0.80, down 53 percent from $1.72 a year earlier and down 57 percent from $1.84 in the previous quarter. For fiscal 2019, revenue was $11.72 billion, up 21 percent from $9.71 billion a year earlier. GAAP earnings per diluted share were $6.63, up 38 percent from $4.82 a year earlier. Non-GAAP earnings per diluted share were $6.64, up 35 percent from $4.92 a year earlier.
"This was a turbulent close to what had been a great year," said Jensen Huang, founder and CEO of NVIDIA. "The combination of post-crypto excess channel inventory and recent deteriorating end-market conditions drove a disappointing quarter.

"Despite this setback, NVIDIA's fundamental position and the markets we serve are strong. The accelerated computing platform we pioneered is central to some of world's most important and fastest growing industries - from artificial intelligence to autonomous vehicles to robotics. We fully expect to return to sustained growth," he said.

Capital Return
In fiscal 2019, NVIDIA returned $1.95 billion to shareholders through a combination of $1.58 billion in share repurchases and $371 million in quarterly cash dividends.

Of the $3.00 billion NVIDIA intends to return to shareholders by the end of fiscal 2020, $700 million in share repurchases were completed in the fourth quarter of fiscal 2019. The company intends to return the remaining $2.30 billion by the end of fiscal 2020, through a combination of share repurchases and cash dividends.

NVIDIA will pay its next quarterly cash dividend of $0.16 per share on March 22, 2019, to all shareholders of record on March 1, 2019.

NVIDIA's outlook for the first quarter of fiscal 2020 is as follows:
  • Revenue is expected to be $2.20 billion, plus or minus 2 percent.
  • GAAP and non-GAAP gross margins are expected to be 58.8 percent and 59.0 percent, respectively, plus or minus 50 basis points.
  • GAAP and non-GAAP operating expenses are expected to be approximately $930 million and $755 million, respectively.
  • GAAP and non-GAAP other income and expense are both expected to be income of approximately $20 million.
  • GAAP and non-GAAP tax rates are both expected to be 10 percent, plus or minus 1 percent, excluding any discrete items. GAAP discrete items include excess tax benefits or deficiencies related to stock-based compensation, which are expected to generate variability on a quarter by quarter basis.
For fiscal 2020, revenue is expected to be flat to down slightly.

Latest Highlights
Since the end of the third quarter, NVIDIA has achieved progress in these areas:

Datacenter
  • Announced that the NVIDIA T4 GPU launched in public beta on Google Cloud Platform, with availability in the U.S., Europe, Brazil, India, Japan, and Singapore.
  • Set six records in AI performance with the release of MLPerf, the industry's first objective set of AI benchmarks.
  • Announced with Google the integration of NVIDIA's RAPIDS GPU-accelerated data science libraries with Kubeflow Pipelines, a Kubernetes-based platform for deploying and managing machine learning workloads in hyperscale datacenters.
Gaming
  • Launched the GeForce RTX 2060 GPU, putting exceptional performance and graphics enhanced by ray tracing and AI within reach of tens of millions of gamers.
  • Unveiled a record 40+ new gaming laptops in over 100 configurations powered by NVIDIA GeForce RTX GPUs.
  • Expanded its G-SYNC ecosystem with G-SYNC-compatible monitors, tested and driver-optimized by NVIDIA.
  • Announced that Battlefield V, the first real-time ray tracing game, will add DLSS AI super-sampling technology; that the highly anticipated game Anthem will integrate DLSS; and that Justice, one of China's most popular MMO games, will add ray tracing and DLSS.
Professional Visualization
  • Introduced Quadro RTX 4000, bringing real-time ray tracing to millions of midrange workstation users.
  • Announced the NVIDIA CUDA -accelerated REDCODE RAW decode SDK, in collaboration with RED Digital Cinema, enabling developers and studios to edit 8K video in real time without the need for additional video processors.
Automotive
  • Introduced NVIDIA DRIVE AutoPilot, the world's first commercially available Level 2+ automated driving system, with Tier 1 suppliers Continental and ZF announcing the availability of Level 2+ solutions based on NVIDIA DRIVE in 2020.
  • Announced with Mercedes-Benz that it will create a centralized computing architecture for the automaker's next-generation vehicles, enabling them to be software-defined AI cars.
Edge Computing
  • Launched the NVIDIA Jetson AGX Xavier , a palm-sized module delivering 32 TOPS to help build the next generation of autonomous machines.
  • Opened its AI and Robotics Research Lab in Seattle.
Add your own comment

56 Comments on NVIDIA Reports Q4 and 2018 Financial Results - Down 45%

#2
the54thvoid
DeathtoGnomes
huge losses
I wish I made such huge losses. Their profits are down but they're still taking in billions.
Posted on Reply
#3
Xzibit
Nvidia
Gaming down 46% Q\Q
Ouch. Huge drop in people buying GPUs or Now that they are getting sued they decided to report their loses. Last Q was in-line.
Posted on Reply
#4
londiste
Operating income went from $1B to $0.3B.
Margins are down a good 5% - 61% to 56%.
Posted on Reply
#5
Noztra
londiste
Operating income went from $1B to $0.3B.
Margins are down a good 5% - 61% to 56%.
Their margins have kept going down and their GPU prices are also falling.

Problem is the die is so big an expensive to make.
Posted on Reply
#6
B-Real
Definitely happens with companies losing their mind and selling products at insanely increased prices: NV, Apple. They may have got huge profits this time too, but it should be taken as a huge warning sign regarding the results of the gaming market's results. If they try to milk customers in the gaming market to this extent, that segment will end up like this. Techpowerup's poll of "Are you buying an RTX?" seems precise.
Posted on Reply
#7
Vayra86
Predicted it ever since the 10 Gigarays Keynote, RTX is going to be a major problem for Nvidia and this road they've chosen is a dead end. They won't even have several generations to make it stick at this rate. Gaming down by this percentage, at THIS time, which is after 2 years of mining and 2,5 years since the last generation launched, is utterly shocking. Some decline was to be expected because the GPU stack is quite capable for any mainstream res and even up to 1440p (so why upgrade, for many) but the other half of this is definitely the completely lacking performance gain and stagnant perf/dollar of GPUs. RTX does not change this, and for a new gen at this time, it really should have changed it. We are now 4 months post launch and product is readily available, out of stock only lasted as long as Nvidia announced early production/stock hiccups. Many people have been sitting on their GPU for several years now, and want high FPS or high res gaming. But, they keep sitting on it regardless.

I think what is very telling, is that Nvidia is firing on all cylinders to turn this around. We have an upcoming 1660ti ('Dafuq-product no.1 of 2019') that sheds RT cores to become a profitable midrange contender. We are already getting bundled games with the RTX line (rarely seen so quickly after a new gen launch). You don't do any of this if your new GPU stack is a sales cannon. And its obviously not because the thing lacking is content. Both in quality and quantity.

This RT-deal is not working out all too well and rightly so. Its not really real time ray tracing. Its a heavily cut down version of it that can't be implemented easily at all and needs to be adjusted for every asset in the scene. It doesn't save developers valuable time and it doesn't create a fantastically natural workflow 'because it just works'. The whole point is that it doesn't just work, as long as we use rasterization it requires additional work, for meagre benefits that are only seen by, thus far, a tiny percentage of the market.

RTX stacks all the problems of recent times on top of one another and implements workarounds to somehow still make it fit. DLSS is such a workaround, but the quality is abysmal. RTX detail levels are another workaround, but at the lower ones you might as well turn it off anyway. Large, expensive dies are a problem, and these financial results are the early warning sign of that. The price of RTX GPUs is a problem, and its unlikely to come down with the only competitor being a Radeon VII at $699,-, still as high or higher than the price of a 1080ti of yesteryear. And then there is the adoption hell, where a gamer needs to tick all the boxes to even be able to use the technology: DX12 ánd RTX GPU ánd have to have one of a handful of games that uses it. Never mind the fact that in one game its not useful since it is competitive multiplayer, another one is discontinued, and in the third one it creates unplayably dark scenes and only simulates a single global light source.

Early adoption never looked so bad as it did with RTX. The fact that it lacked the initial hype is writing on the wall for its future. If you would be an investor, this is one helluva high risk investment right about now.

londiste
I am not convinced this is purely Nvidia going wrong way at doing things.

Unfortunately, AMD has got Radeon inside the Computing and Graphics segment but they also had revenue decrease. With the uptick of Rysen and EPYC, that is a bit strange. Plus, their outlook said:

There are no more companies in that segment but it does look both are going to get hit.
Absolutely, most tech companies are taking similar hits to AMD. ~10-20%. But not 45%. NVDA had more of a correction coming because its stock soared in 2H18, but that alone is telling because the reason it soared was not just good products, but mining profits and lots of great promises. Many of those promises are now turning out not to be as promising, and its not just gaming. If you look at it, the value is right back to where it stalled at the end of the Pascal/Volta generation - the last GPU gen where Nvidia played the game like people expect them to: with sizeable generational performance bumps and improving perf/dollar.
Posted on Reply
#8
londiste
I am not convinced this is purely Nvidia going wrong way at doing things.

Unfortunately, AMD has got Radeon inside the Computing and Graphics segment but they also had revenue decrease. With the uptick of Rysen and EPYC, that is a bit strange. Plus, their outlook said:
For the first quarter of 2019, AMD expects revenue to be approximately $1.25 billion, plus or minus $50 million, a decrease of approximately 12 percent sequentially and 24 percent year-over-year. The sequential decrease is expected to be primarily driven by continued softness in the graphics channel and seasonality across the business.
There are no more companies in that segment but it does look both are going to get hit.

Vayra86
I think what is very telling, is that Nvidia is firing on all cylinders to turn this around. We have an upcoming 1660ti ('Dafuq-product no.1 of 2019') that sheds RT cores to become a profitable midrange contender. We are already getting bundled games with the RTX line (rarely seen so quickly after a new gen launch). The only thing lacking is content. Both in quality and quantity.
This has to do with the lack of perf/$ progress but not much with RTX line. 1660 will clearly be positioned below RTX 2060.
Posted on Reply
#9
Vayra86
londiste
This has to do with the lack of perf/$ progress but not much with RTX line. 1660 will clearly be positioned below RTX 2060.

I know you do not like RTX but its time to get over it. Early adopter has looked a lot worse than RTX, repeatedly.
Thing is, AMD can still fight the 2060 with their ancient Vega die without any problem. Nvidia needed a 1660ti sans RT cores to keep margins intact. The timing and naming of it strikes me as a Coffee-Lake quality launch. It was on some plan-B roadmap in case everything turned to shit.

I'm not sure how you feel early adoption looked worse repeatedly. Gaming is mainstream now and there is so much more movement than there was years ago when it was still moving up. Consoles define what games land on what platform for this mainstream and the time of PC being 'leading' is long gone. There is no killer app such as Crysis at this time, to showcase the technology and 'wow' us with it. I think this adoption hell with APIs and standards (because DXR alone won't be enough to cover console land) will kill it eventually. DX12 content is scarce and not gaining that much traction at all. The best we might see of RT is as an improved source for rasterized content, well I would call that a certain win. But brute forcing it in real time, most certainly is not. Its horribly inefficient and won't stand up against much cheaper methods (for the end user). And as every node from now on gets progressively harder to work with, you can bet the cost aspect is going to be a problem. It already is.

Compare it to VR: high cost of adoption, standardization issues, lack of content and killer apps, have to make concessions to use it. Its remarkably similar to RTX. I think VR had better chances than this, and look where that is at right now. The only one-day fly that really had meaningful sales numbers was the PSVR, which will also turn out to be nothing but Eye Toy 2.0. Gimmick.
Posted on Reply
#10
londiste
Vayra86
Thing is, AMD can still fight the 2060 with their ancient Vega die without any problem. Nvidia needed a 1660ti sans RT cores to keep margins intact. The timing and naming of it strikes me as a Coffee-Lake quality launch. It was on some plan-B roadmap in case everything turned to shit.
Vega is only a year and a half old. Even though a little slower, RTX2060 is slightly smaller and a lot more efficient even with all the cruft. Plan-B does not make sense - 1660 is below everything RTX. This is pure plan-A for midrange.
Vayra86
Gaming is mainstream now and there is so much more movement than there was years ago when it was still moving up. Consoles define what games land on what platform for this mainstream and the time of PC being 'leading' is long gone. There is no killer app such as Crysis at this time, to showcase the technology and 'wow' us with it. I think this adoption hell with APIs and standards (because DXR alone won't be enough to cover console land) will kill it eventually.
PC is leading platform more than ever. There has never been a time before where console GPUs are midrange at launch. There has never been time where major consoles are straight up PCs with a more restricted OS.
Wow factor is hard to come by because games look "good enough" already. And today, 2080Ti can mostly do 60 FPS with maxed settings. With "good enough" settings, 2080Ti is capable of a lot more and the 3-4 cards lower on the ladder will do 4K just fine. Rasterization is nearing an end because there is nowhere to go. 8K will not give the visual impact you'd expect for a number of reasons, even besides the fact that 8K screens are years away. Hell, even good 4K gaming monitors are hard to come by.
Vayra86
DX12 content is scarce and not gaining that much traction at all. The best we might see of RT is as an improved source for rasterized content, well I would call that a certain win. But brute forcing it in real time, most certainly is not. Its horribly inefficient and won't stand up against much cheaper methods (for the end user). And as every node from now on gets progressively harder to work with, you can bet the cost aspect is going to be a problem. It already is.
Effects only get more and more complex, require more and more compute power and more render passes. Some recent methods for things like AO (for example VXAO) are already halfway there to raytracing. If RTRT performance can be brought up enough to do some of it real-time that is a perfect place to start. From Metro, it looks like we are almost there.
Posted on Reply
#11
Vayra86
londiste
Vega is only a year and a half old. Even though a little slower, RTX2060 is slightly smaller and a lot more efficient even with all the cruft. Plan-B does not make sense - 1660 is below everything RTX. This is pure plan-A for midrange.
Can see that being true as well, however, it makes you wonder why not just use Pascal dies for it. Turing has nothing to offer otherwise and they had all that stock, right...

As for 'being under' the 2060... If the 1660ti will compete with GTX 1070 then it will be competing with Vega 56 as well. The midrange is getting pretty crowded.
Posted on Reply
#12
XiGMAKiD
Another round of cryptocoin might help them
Posted on Reply
#13
londiste
Vayra86
Can see that being true as well, however, it makes you wonder why not just use Pascal dies for it. Turing has nothing to offer otherwise and they had all that stock, right...
Turing has feature parity with Vega plus some additional things here and there. If you look at game benchmarks where Vega used to get a lead over Pascal, it never really gets that over Turing.
Posted on Reply
#14
Xaled
Xzibit
Ouch. Huge drop in people buying GPUs or Now that they are getting sued they decided to report their loses. Last Q was in-line.
Yeah this is a clear proof that they lied in previous results. if i dont remember wrong once theyve reported 100 millions of mining and 1 billion onf gaming!
Posted on Reply
#15
matar
Nvidia with those ridiculous RTX prices Expect even more loses ( Lower the prices like the GTX 1080Ti lunch day prices ) example RTX 2080TI $649 and so on with the rest of the RTX series and you will see more people buying GPUs and your profits goes back up.
Posted on Reply
#16
Noztra
matar
Nvidia with those ridiculous RTX prices Expect even more loses ( Lower the prices like the GTX 1080Ti lunch day prices ) example RTX 2080TI $649 and so on with the rest of the RTX series and you will see more people buying GPUs and your profits goes back up.
Probably not.

They will just launch 11xx cards instead. :)
Posted on Reply
#17
Robcostyle
matar
Nvidia with those ridiculous RTX prices Expect even more loses ( Lower the prices like the GTX 1080Ti lunch day prices ) example RTX 2080TI $649 and so on with the rest of the RTX series and you will see more people buying GPUs and your profits goes back up.
I think youre not getting the full picture of the market... you see, these “losses” untied huangs hands for further price rise. Revenue loss == higher prices.
And there is 2 awfull things that will make him to do so 100%: its shareholders and his own greed.
Posted on Reply
#18
Aerpoweron
matar
Nvidia with those ridiculous RTX prices Expect even more loses ( Lower the prices like the GTX 1080Ti lunch day prices ) example RTX 2080TI $649 and so on with the rest of the RTX series and you will see more people buying GPUs and your profits goes back up.
I don't think so either. The GPUs with RTX are very complex chips. So the price is reasonable if you consider manufacturing. And for now, the RTX and Tensor cores are more or less dead weight you have to manufacture and which can get broken. And Nvidia won't sell the cards at a loss. (Does anyone have manufacturing prices?)

And when you look at the teardowns, the coolers and the whole assembly must be quite expensive as well. At least on the Founders Edition cards, which seem to have the best chips too. And then there is the problem of maintenace for the end user, which is at least a big no go for me.
Posted on Reply
#19
unikin
Robcostyle
I think youre not getting the full picture of the market... you see, these “losses” untied huangs hands for further price rise. Revenue loss == higher prices.
And there is 2 awfull things that will make him to do so 100%: its shareholders and his own greed.
Let them do it. I'm one rise away from buying PS5 when it comes out and be done with PC gaming. That's coming from someone who's been PC gamer for 25 years and according to forums talk I'm not alone. High end PC gaming has become ridiculously expensive.
Posted on Reply
#20
Robcostyle
Aerpoweron
I don't think so either. The GPUs with RTX are very complex chips. So the price is reasonable if you consider manufacturing. And for now, the RTX and Tensor cores are more or less dead weight you have to manufacture and which can get broken. And Nvidia won't sell the cards at a loss. (Does anyone have manufacturing prices?)

And when you look at the teardowns, the coolers and the whole assembly must be quite expensive as well. At least on the Founders Edition cards, which seem to have the best chips too. And then there is the problem of maintenace for the end user, which is at least a big no go for me.
Yes, roughly 100$ per tu102 chip.

It is MASS PRODUCT, people! Margin they have from every single card is enormous! From 200-300% in worst case, probably up to 1000%! I don’t understand why you people keep sympathizing ngreedia and their chip “complexity “

unikin
Let them do it. I'm one rise away from buying PS5 when it comes out and be done with PC gaming. That's coming from someone who's been PC gamer for 25 years and according to forums talk I'm not alone. High end PC gaming has become ridiculously expensive.
You aren’t - I’m in too!) I’ve already took ps4 last Xmas - just when I realized that TV screens nowaday give you much more for the same price than pc monitor
Posted on Reply
#21
Aerpoweron
[quote=Robcostyle]
Yes]Ok, lets assume the 100$ is correct for the moment. (i would like your sources for that and some comparison to other chips of you have, just a link would do it for me :) )
You have to add the memory chips to that, as well as the PCB and all the power supply parts on the card. Plus installing the cooling assembly and manufacturing for it. And i think we can only guess, since it would be very hard to find accurate prices for all the steps and parts.

And to make it clear, i don't like Nvidia very much. They messed too often with their customers. Once a year in my opinion. And this is why i don't by or like Nvidia any longer.

But the Touring chip should be more expensive in manufacturing compared to chips with same gaming power but without the RTX and AI cores.

And we should see reasonable prices on the 1660ti when the assumption on the complexity of the TU102 chips holds up.
Posted on Reply
#23
Aerpoweron
Opps. Looks like we were talking about reality. The stock market is a whole different world. These analysts are very strange people. Tesla did't manufacture as much as analysts expected and the stock prices got down. Tesla did miss the mark by 1500cars, and made a record in income :D

So technical world is totally different from the stock market world. And analysts are mighty powerful creatures there.
Posted on Reply
#24
FordGT90Concept
"I go fast!1!11!1!"
Tesla has only ever had one profitable quarter and now they're bleeding money like crazy:


Yet, their stock valuation has been going up for five years:
https://finance.yahoo.com/quote/TSLA/

Yup, doesn't make any sense. Musk's twitter comment about buying the company out got him in hot water, as it rightfully should.
Posted on Reply
#25
Robcostyle
Aerpoweron
Ok
Keep in mind, that they get it all at gross price, level of optimization of operating expenses in this kind of companies is astonishing, it's always all together - poor or mediocre quality components, choice of cheapest supply routes - by sea, by air if it is cheaper than delivery of jam from your grandma, and etc.,etc. - otherwise you would never get all this tech - rtx2080ti, iPhone, your brand new TV or car or fridge - at anything close to "affordable", at all.
Posted on Reply
Add your own comment