Analyst: NVIDIA Lied about Its Cryptocurrency Earnings to Avoid Stock Crash

Megalith

24-bit/48kHz
Staff member
Joined
Aug 20, 2006
Messages
13,000
In its recent earnings report, NVIDIA says it generated around $602 million in crypto-related sales. Not so, says Royal Bank of Canada capital market analyst Mitch Steves, who did some math and claims the actual number is closer to $1.95 billion. While the figure isn’t something he can definitively prove, AMD’s earnings report lends credence to the theory ($234M/25% vs. NVIDIA’s hypothetical $1.95B/75%, with total crypto revenue being $2.75B). Some presume NVIDIA is concealing revenue in a desperate attempt to smooth over their stock troubles.

By his calculation, the total crypto revenue from April 2017 to July 2018 should be around $2.75 billion, based on the hash rate of ethereum and other cryptocurrencies that require graphics processing units. Steves estimates Nvidia captured around 75% of the total crypto market during that period and AMD captured the rest. There is no way to actually confirm the numbers, according to Steves. Steves said that AMD's recent earnings release added another piece of evidence that Nvidia's exposure to crypto should be higher than it revealed.
 
Thanks for the article. .. wow, nVidia controlled 75% of the mining market? Amazing! I had no idea they were dominating that bad over AMD. But, it does make sense I guess. They are so much more powerful than AMD.
 
In its recent earnings report, NVIDIA says it generated around $602 million in crypto-related sales. Not so, says Royal Bank of Canada capital market analyst Mitch Steves, who did some math and claims the actual number is closer to $1.95 billion. While the figure isn’t something he can definitively prove, AMD’s earnings report lends credence to the theory ($234M/25% vs. NVIDIA’s hypothetical $1.95B/75%, with total crypto revenue being $2.75B). Some presume NVIDIA is concealing revenue in a desperate attempt to smooth over their stock troubles.

By his calculation, the total crypto revenue from April 2017 to July 2018 should be around $2.75 billion, based on the hash rate of ethereum and other cryptocurrencies that require graphics processing units. Steves estimates Nvidia captured around 75% of the total crypto market during that period and AMD captured the rest. There is no way to actually confirm the numbers, according to Steves. Steves said that AMD's recent earnings release added another piece of evidence that Nvidia's exposure to crypto should be higher than it revealed.


Ok, so I think I know what you're saying... But isn't the "actual" $1.95 billion more than the "reported" $602 million...?
 
Thanks for the article. .. wow, nVidia controlled 75% of the mining market? Amazing! I had no idea they were dominating that bad over AMD. But, it does make sense I guess. They are so much more powerful than AMD.

I never was big into Crypto, but I was under the impression it was the other way around - AMD was the better performer for most miners. That doesn't mean that nVidia couldn't still command 75% of the market, but I wouldn't go and just assume that it's just because "they are so much more powerful".

Also, this doesn't seem to take into account any ASIC or non-GPU mining activity, of which there was certainly some.
 
Last edited:
Ok, so I think I know what you're saying... But isn't the "actual" $1.95 billion more than the "reported" $602 million...?

I think the implication is that nVidia doesn't want to disclose the higher number - because now that it's gone and not coming back, it will affect future earnings to a much greater extent. And that's an impression they don't want to give, because it will negatively affect stock prices (as it should).

nVidia is betting the house on the fact that something else will pick up in the mean time (RTX sales, Automotive sales, AI sales) to mask the fact that their previous revenue was so dependent on Crypto. If it doesn't, you end up putting out revised earning statements and getting your stock hammered anyway... but there's a chance you may be able to avoid it (so long as you can avoid Fraud charges)
 
I never was big into Crypto, but I was under the impression it was the other way around - AMD was the better performer for most miners.

Also, this doesn't seem to take into account any ASIC or non-GPU mining activity, of which there was certainly some.
They were. Not sure what hes talking about.
 

Attachments

  • Ethereum-Mining-GPU-Benchmarks-With-RX-Vega.png
    Ethereum-Mining-GPU-Benchmarks-With-RX-Vega.png
    41.4 KB · Views: 0
So a company took advantage of the crypto market and sold a crap load of product at list price and raised the price and sold still more. In most other realities, stockholders would be heaping praise upon the leaders of such a well run company. Instead, stockholders seem pissed that the market demand didn't last forever and as a result, future sales will be returning to what they were before the crypto market bubble.
 
So a company took advantage of the crypto market and sold a crap load of product at list price and raised the price and sold still more. In most other realities, stockholders would be heaping praise upon the leaders of such a well run company. Instead, stockholders seem pissed that the market demand didn't last forever and as a result, future sales will be returning to what they were before the crypto market bubble.

Kind of.

Shareholders did love it. Look at nVidia's price for 2018.

Then it all dried up. Shareholders ~hate~ that. Wall Street rewards growth over all else. This is the part you are missing. Shareholders don't care if it can't last forever - they will reward whomever can find the next thing to keep driving growth, and punish those that stagnate and can't keep growing, or even worse, backslide and shrink. It's completely irrational and I may not agree with it, but it's how the stock market works.

nVidia doesn't want their stock price to tank.

The article is insinuating that nVidia is trying to mask how much money they made on Crypto to hide it from shareholders so the extent of the now non-expected revenue will be masked.
 
....

nVidia is betting the house on the fact that something else will pick up in the mean time (RTX sales, Automotive sales, AI sales) to mask the fact that their previous revenue was so dependent on Crypto....

Well, Tesla will stop buying around 90 000 Nvidia Drive PX 2s in Q2 2019 when they switch to their own chips for Autopilot and since Nvidia Drive PX 2 is not exactly cheap, this will further hurt Nvidia revenue.
 
They were. Not sure what hes talking about.
I still wouldn't be surprised to see that NVIDIA captured 75% of the crypto market because the cards used much less power, ran cooler, and were actually available to buy. AMD did not produce GPUs as quickly as NVIDIA could.
 
  • Like
Reactions: noko
like this
Thanks for the article. .. wow, nVidia controlled 75% of the mining market? Amazing! I had no idea they were dominating that bad over AMD. But, it does make sense I guess. They are so much more powerful than AMD.
For those of you who are not financially savvy, It's a big deal when Executives lie about their earnings and expected future earnings guidance.
The SEC will lay the hammer down with fines if the company is not being forthright in their guidance.
The problem here is Nvidia was either negligent (not likely), or misleading when they disclosed the portion of their sales that were related to mining vs Gaming.
Gaming is a more stable market for Nvidia, so it would be in their best interest to fluff the sales numbers to indicate that mining was a smaller portion of sales than reality. This is a gray area, and they may have gotten caught playing too aggressive with their accounting and analysis when they down played the significance of mining revenue.
 
So a company took advantage of the crypto market and sold a crap load of product at list price and raised the price and sold still more. In most other realities, stockholders would be heaping praise upon the leaders of such a well run company. Instead, stockholders seem pissed that the market demand didn't last forever and as a result, future sales will be returning to what they were before the crypto market bubble.
Nvidia themselves didn't raise the price on the 10xx cards. I believe the price increase of 20xx cards was mostly from their market dominance.
 
For those of you who are not financially savvy, It's a big deal when Executives lie about their earnings and expected future earnings guidance.
The SEC will lay the hammer down with fines if the company is not being forthright in their guidance.
The problem here is Nvidia was either negligent (not likely), or misleading when they disclosed the portion of their sales that were related to mining vs Gaming.
Gaming is a more stable market for Nvidia, so it would be in their best interest to fluff the sales numbers to indicate that mining was a smaller portion of sales than reality. This is a gray area, and they may have gotten caught playing too aggressive with their accounting and analysis when they down played the significance of mining revenue.

Well reading the article, it's all speculation. He says he did the math, but has no proof. Take that as you will.
 
Well reading the article, it's all speculation. He says he did the math, but has no proof. Take that as you will.
Yes, agree that it would be very difficult to prove.
Roughly speaking, all you really need to do is compare Nvidia sales just prior to the crypto boom and you will likely see the spike in revenue from cyrpto was much larger than reported by Nvidia.
 
I still wouldn't be surprised to see that NVIDIA captured 75% of the crypto market because the cards used much less power, ran cooler, and were actually available to buy. AMD did not produce GPUs as quickly as NVIDIA could.

Course that would be "believing" in various Analyst numbers NVDA numbers etc etc... all of them have bias one way or the other and Jensen has proven time and time again will do whatever possible to hide "the dirty truth" by sweeping under the rug.

Radeon have been "superior" for mining/hashing/code cracking/complex calc purposes for a very long time, up until specific designs favored certain proucts (FPGA, ASIC et al) NVDA likely did sell A LOT more from 2015 till they reported 2018 earnings, however, this absolutely is only taking them at their word when their word is quite tainted..

Anyways, I know for many coins you were a fool to bother mining with NVDA anything, however with some coins there were NVDA focused coins last I believe was ~2 years at most that actually made mining on NVDA worth it directly (especially taking into account lower average power for the mining performance given ( X units per day sort of speak)

Likely folks were not going out of their way to buy NVDA for "mining" directly (at least not for a solid 8-10 mths after GTX 700 or 1xxx/2xxx were launched) unlike many Radeons that seemed to be highly sought only for the hashing performance (tuned down vast reduction in power very similar performing to the GTX/RTX or whatever) Odd actually because Vega 56 / 64 the Fury cards beyond being quite potent for gaming are actually very very good for mining (when voltage tuned) beyond the initial purchase price.

I suppose I always considered the Radeons better for mining ($$$ purpose) where the Nv cards were generally good "folders" ( F@H, Seti@home) etc. I know when I was mining quite a while back there were only maybe 5 coin types "worthwhile" to be mining with Nv stuff vs the 100s possibly worthy for the Radeons.

I suppose it goes to show, had Nv really wanted mining profits they would have used their billions to build custom hash machines, or at least do a "headless" GPU that has its FP performance at 1/2 or 1/4 vs 1/10-1/25 as many of their things currently do (to make sure the actual PRO level cards are fully enabled FP performance).

I do not feel sorry for Nv nor Jensen either way, a day late many dollars short to try and "act" like the story matters when it comes to their consumers.
 
Last edited:
Not really, just look at their taxes, the money trail never lies.
That doesn't detail which market segment the revenue came from. Total numbers are not the concern, it's what was sold to generate their revenue.
 
That doesn't detail which market segment the revenue came from. Total numbers are not the concern, it's what was sold to generate their revenue.

I'm sure it could be figured out. Hey, maybe Nvidia A.I could help :)
 
How do either company even figure out their revenue distribution of mining vs gaming? Like, how does NVIDIA know it generated $600 mil in crypto related sales? Do they just use a 3rd party app to see how many people use the cards (kinda like steam for gaming)? and then apply some calculation to say...yeah, we generated this amount of revenue from crypto.

It just seems odd to me that either company needs to separate the two. Its a GPU, sales are sales no matter what someone decides to use it for.

I guess another way to say this is that unless they are doing their own crypto mining, they arent really generating revenue from it. they are just selling gpu's. People use both companies cards to view porn, people use their cards to mine, people use their cards for many various aspects. their revenue should just be based off how many cards they have sold.
 
Last edited:
They were. Not sure what hes talking about.
analyst-nvidia-lied-about-its-cryptocurrency-earnings-to-avoid-stock-crash.1976604


191809_Ethereum-Mining-GPU-Benchmarks-With-RX-Vega.png

Because your cherrypicked benchmark shows the one coin AMD was good at, while Nvidia was good at everything. And once EthLargementPill came along, 1080Ti's were way over the top in ETH mining as well (50MH/s - 55MH/s).

There were cost benefits to running huge farms of AMD cards just mining ETH though -- mainly that they were cheaper because there was less demand on AMD cards from gamers.
 
Last edited:
How do either company even figure out their revenue distribution of mining vs gaming? Like, how does NVIDIA know it generated $600 mil in crypto related sales? Do they just use a 3rd party app to see how many people use the cards (kinda like steam for gaming)? and then apply some calculation to say...yeah, we generated this amount of revenue from crypto.

It just seems odd to me that either company needs to separate the two. Its a GPU, sales are sales no matter what someone decides to use it for.
previous years history coupled with current year data, give yous a fairly accurate number projected volume, conclusion, the extra volume is from the variable
simplified exemple :
year1 Q1 sold 10products
year2 Q1 sold 11products
current year Q1 sold 30 products, if previous quaters doesnt show a significant increase, then you have 18-20 product more because of variable X(crypto)
 
Last edited:
If the cryptocurrency market is falling (crashing), and you're HEAVILY reliant on it for your revenue (like, 75% of your income), and you LIE on an investor call (say, call it 25%) in an effort to limit your exposure to a shrinking market (and thereby limit the drop in your stock)...you've violated several laws and securities regulations.

Hmmm....maybe we should put Mueller on it?
 
I still wouldn't be surprised to see that NVIDIA captured 75% of the crypto market because the cards used much less power, ran cooler, and were actually available to buy. AMD did not produce GPUs as quickly as NVIDIA could.
Probably has more to do with Nvidia being able to meet supply and not because they were superior to crypto mining. There's a reason why a RX 480 is about $100 on Ebay right now, because they were the go to mining card. When the price of the RX 480's and RX 470's were too damn high, they jumped over to Nvidia.
 
NVIDIA is already down 50% from it's high so I'm not sure the strategy worked.
 
NVIDIA is already down 50% from it's high so I'm not sure the strategy worked.
sometimes things are said to buy time, maybe jensen was hoping to charge the highest possible price for RTX and if they sell extremly well, he would have been able to hide those missing 500mil$, but since RTX tanked...
 
Probably has more to do with Nvidia being able to meet supply and not because they were superior to crypto mining. There's a reason why a RX 480 is about $100 on Ebay right now, because they were the go to mining card. When the price of the RX 480's and RX 470's were too damn high, they jumped over to Nvidia.

Many reasons on why Nvidia became the better mining GPU for miners.

1. ROI was better. Nvidia GPUs could use less power and have good enough hash rates that your ROI would happen sooner rather than later. Even if the card costs more than a similar AMD card. Lower power requirements.
2. Drivers. Nvidia drivers are so much better you can truly setup a mining rig and virtually forget about it. AMD's drivers would crash and when that happens, you're not making money.
3. Supply and Demand, as you said. However, if this was the case then AMD would have never been on top. Nvidia always had a higher supply than AMD.

During the beginning of crypto mining AMDs card were better at it by a large margin, but after a while better mining pools were created giving the edge to the much, much more efficient Nvidia cards. Sure some AMD cards can mine faster than Nvidia, but at what cost?
 
Keep in mind that the cards nVidia sold due to mining may have never been used for mining in the first place. AMD cards being bought out left and right for mining would have left a hole in the gaming GPU market which nVidia would have stepped into. In this case what Jensen said may actually be 100% true in that their figures showed that cards bought to be mined on were the said $ amount. However, the only reason nVidia was able to hit the high numbers was because AMD cards were non-existent with regards to gamers.

nVdia's numbers may be correct to the point that most of the cards bought were used for non-mining purposes. However, those cards may have only been sold because AMD cards were purchased for mining and leaving the gaming sector open almost completely for nVidia. nVidia's sales would have dramatically increased due to the mining craze even if only a small percentage of cards were used for mining. The statement of sales numbers would technically be correct but at the same time completely misleading especially since nVidia's sales would not be sustainable once stock of AMD cards came back to the channel instead of being mostly grabbed by miners.
 
For cryptonight, Nvidia was never in the same ballpark. For a project like Monero, nobody used Nvidia really. For GPU mining, it was Vega > *.

Equihash projects did very well with Nvidia hardware. Zcash, and a ton of smaller alt-coin projects when the market was booming did much better with their hardware.

But yah. Nvidia stuff was easier to setup and run. The AMD side needed power table tweaks in the registry, overdriventool, etc.. Bit of a pita.
 
  • Like
Reactions: N4CR
like this
I buy and sell way to many used pc parts and as far as I can tell about 70% of the used mining card floating around are nvidia. This story lends some insight into the investor class action against nvidia.
Yup lots more around than amd, even here.
 
Ok, so I think I know what you're saying... But isn't the "actual" $1.95 billion more than the "reported" $602 million...?
Remember in the stock game. It does not matter how much you make so long as you made more than last time, AND meet expectation. So nvidia lowered the bar...for next year.
 
Instead, stockholders seem pissed that the market demand didn't last forever and as a result, future sales will be returning to what they were before the crypto market bubble.

that's because most people are getting more and more evil (in this case 'greedy') as time marches on.
 
Sure that title picture isn't him describing the boob console?
 
Many reasons on why Nvidia became the better mining GPU for miners.

1. ROI was better. Nvidia GPUs could use less power and have good enough hash rates that your ROI would happen sooner rather than later. Even if the card costs more than a similar AMD card. Lower power requirements.
2. Drivers. Nvidia drivers are so much better you can truly setup a mining rig and virtually forget about it. AMD's drivers would crash and when that happens, you're not making money.
3. Supply and Demand, as you said. However, if this was the case then AMD would have never been on top. Nvidia always had a higher supply than AMD.

During the beginning of crypto mining AMDs card were better at it by a large margin, but after a while better mining pools were created giving the edge to the much, much more efficient Nvidia cards. Sure some AMD cards can mine faster than Nvidia, but at what cost?

I said Nvidia lied on their earnings a few quarters ago. People told me nooooo. Didnt they say crypto only accounted for like 7% of their growth in sales or some such nonsense? I’m too indifferent to go look it back up. When guys like me go from buying one mid range graphics card every 3 years to buying over forty 1080TI in a single year, me and several of my tech friends, that ain’t no measouly 7% increase.

To your point...

I’m a miner who started out with AMD cards in summer of 2017 and finished with 100% Nvidia cards. Why? AMD cards required much more babysitting in windows using algo switching than Nvidia cards. Nvidia cards will mine with nicehash for months on end without touching them. Afterburner works better for nvidia card to tune, undervolt and overclock. My AMD systems were a headache. The wattman settings would reset. The drivers took forever to load so you could overclock cause you had to do each card individually instead of all at once like Nvidia using afterburner. It took me weeks to setup a 12 card AMD RX580 system and it remained a huge administrative headache. A 12 card Nvidia 1060 system around the same time was stable as a rock.

Nvidia cards used less power, generated less heat, were cheaper and easier to source closer to MSRP longer, they ran more reliably, and towards the end, made better profits than AMD cards too. All my miner friends switched to Nvidia cards.

That from my experience mining with about 60 cards from mid 2017 on. I still mine, but only have one rig left and it has Nvidia cards.
 
Last edited:
Because your cherrypicked benchmark shows the one coin AMD was good at, while Nvidia was good at everything. And once EthLargementPill came along, 1080Ti's were way over the top in ETH mining as well (50MH/s - 55MH/s).

There were cost benefits to running huge farms of AMD cards just mining ETH though -- mainly that they were cheaper because there was less demand on AMD cards from gamers.

There were plenty of coins AMD hardware was good with, but which coin you chose to run was a function of the power you were willing to burn. Some were memory heavy (Eth), some were core heavy (Zec).

nV was cheaper to buy, had "good enough" returns and easier to get due to supply. Its no surprise they sold a metric crapton of cards.

Also setting memory timings on AMD hardware wasnt the easiest thing.

At one point I was running:

1x VegaFE
1x Threadripper 1950X
11x 7950/7970/R9 280X
4x R9 580 8gb
2x R9 480 4gb
6x R9 290X
1x R9 380
1x 1060 6GB
1x Titan (OG)
2x 750Ti


And some other junk. Power didnt matter at the time so everything was overclocked to the nuts..

EDIT

Will agree AMD cards were harder to set up but I didnt have to babysit them much. If you set up your algorithms correctly they switched fine for me. Afterburner worked fine for the 290X's, and Sapphire Trixx seemed best for the 7900 series cards.
 
So you're saying short NVDA?

Well this report (written by Mitch Steves of RBC) was issued on Thursday and NVDA was actually up on Friday. Unless everyone just missed it and reacts to this one article then it shouldn't affect price until there's additional info corroborating the claim.
 
Back
Top