Nvidia CEO Jensen Huang's pay slashed by $2.5 million after company misses sales targets

[midian182]

In brief: Nvidia boss Jensen Huang found his pay check was light by almost $2.5 million in the last financial year. But this wasn't a Tim Cook-style voluntary reduction in pay; Huang received a 10% cut because Nvidia missed financial targets. Still, it's not as if team green's CEO is going to struggle to pay the bills.

As per The Reg, Huang's compensation package for the fiscal year 2023, which ended on January 29, totaled $21.356 million. That's a lot of money, around 94 times what the average Nvidia worker earns, but still 10% lower than what Huang made the previous year.

Like the majority of tech leaders, most of Huang's compensation comes from stock awards - $19.666 million during the last fiscal year. His base salary was $996,216, and he was given $693,710 in other compensation, such as security, life insurance, and travel.

Virtually every big company has been feeling the effects of the current economic crisis. Expensive graphics cards have been hit particularly hard, with sales reaching their lowest point in December since 2005. Nvidia's revenue was flat for the 12 months of FY23, while net income fell by $5.382 billion.

"Fiscal 2023 was a challenging year, with macroeconomic headwinds, channel inventory corrections, COVID-19 and product architecture transitions affecting several of our businesses," Nvidia wrote in a proxy statement. The result was the company's Fiscal 2023 revenue and non-GAAP Operating Income performance falling short of the "CC's (compensation committee's) pre-established goals for executive compensation."

While the rest of Nvidia's executives saw their compensation go up rather than down, they still could have made more. CFO Colette Kress' target pay was $11.99 million but she was awarded $10.91 million, while VP of Worldwide Operations, Ajay Puri, received $1.37 million less than he could have earned. The median employee's compensation was $229,078.

Huang was the only executive to see his yearly compensation go down. But with a personal net worth of $26.1 billion, a $2.5 million pay cut is unlikely to be felt very much.

While Huang's pay cut was due to missed targets, Apple boss Tim Cook's 40% reduction in compensation was voluntary – and despite 64% of shareholders voting in favor of Apple execs retaining their previous year's pay packages.

Permalink to story.

https://www.techspot.com/news/98665-nvidia-ceo-jensen-huang-pay-cut-25-million.html

 

[Nanochip]

If they lowered gpu prices, sales would increase. But they won’t, they’re the real scalper now.

 

[Blqckout]

Where can I get 700k compensation please?

 

[DSirius]

Well, Karma is starting to catch up Jensen Huang. This may be just the beginning.
"But with a personal net worth of $26.1 billion, a $2.5 million pay cut is unlikely to be felt very much."
I bet that Huang felt it and care about this more than it shows, not as much financially as for his reputation.
Conducting a business driven primarily by greed is not a viable model in any market with finite resources or in recession.

 

[Achaios]

Every time I see the Leatherman I remember his statement about GPUS:

"The Era Of Cheap Graphics Cards is Over For Ever"

So long as this guy is the boss at NVIDIA we won't see cheap GPUs.

 

[Neatfeatguy]

How is he going to afford a new leather jacket? WTF were they thinking when slashing his pay? He's going to freeze and die without a plethora of leather jackets.

 

[nnguy2]

Every time I see the Leatherman I remember his statement about GPUS:

"The Era Of Cheap Graphics Cards is Over For Ever"

So long as this guy is the boss at NVIDIA we won't see cheap GPUs.

Just like every villain, he's announcing how he's going to execute his evil plans.

 

[wiyosaya]

Awwww. Poor Jensen. Let's have a pity party for him.

Maybe nVidia can make up some of their lost revenue in his paycut.

 

[waclark]

If they lowered gpu prices, sales would increase. But they won’t, they’re the real scalper now.

That's not a foregone conclusion. They might increase, but what is the impact to profits? Sales have declined for Nvidia (and others) as have profits. How many more 4090s would you sell at $1500 rather than $1600? The problem isn't price reduction, it's that prices at $1000 and up are well above what most of us want to pay. Nvidia's Net Margins, according to Google are around 16% as of Jan 2023. If you reduced the cost of the 4090 by 16% it would cost about $1350. Do you think a lot more 4090s would sell at $1300? I supposed a few more, but I doubt you would see double sales.

 

[Nanochip]

That's not a foregone conclusion. They might increase, but what is the impact to profits? Sales have declined for Nvidia (and others) as have profits. How many more 4090s would you sell at $1500 rather than $1600? The problem isn't price reduction, it's that prices at $1000 and up are well above what most of us want to pay. Nvidia's Net Margins, according to Google are around 16% as of Jan 2023. If you reduced the cost of the 4090 by 16% it would cost about $1350. Do you think a lot more 4090s would sell at $1300? I supposed a few more, but I doubt you would see double sales.

fair point. but what any corporation must keep in mind when setting prices is: there is no profit if the product is sitting in inventory. I went to micro center the other day and there is a ton of GPU product in inventory sitting in on the shelf.

Remember, there is elasticity of demand. If you chase profits above all else, and you set a price too high for the market to bear, then one likely outcome is high inventory and low sales, and ultimately low profits.

The corporation must read the room and meet consumers where they’re at. Nvidia clearly got drunk off of mining demand and scalper prices. They likely thought: “if consumers were willing to spend so much on ampere, and Lovelace is better in performance and performance per watt, then consumers will pay more for Lovelace too right?”

But they clearly haven’t read the room. Much of the demand from last generation was artificial and unsustainable in my opinion, it was in large part mining demand, and a lot people who couldn’t travel, and wanted to game during the Covid lock downs.

But Covid is over and ethereum is now proof of stake and cryptocurrency demand has cooled off in general. So GPU demand is normalizing back towards being driven primarily by price-sensitive gamers.

And gamers clearly aren’t interested in the Lovelace tax. Which is good. This means nvidia must innovate on their end to keep production costs low, to be able to still make a good profit in a market where consumers aren’t interested in exorbitant prices.

 

[poohbear]

Its still peanuts compared to his overall wealth from Nvidia stock, which made him a billionare. It was a $2/share stock 23 years ago. It reached a high (before stock split) of $1300/share during the peak of the pandemic. $21 million is pocket change for him.

 

[neeyik]

what any corporation must keep in mind when setting prices is: there is no profit if the product is sitting in inventory.

For Nvidia, that only applies to its own FE models. Companies like MSI, Gigabyte, Zotac, etc have to purchase the GPUs from Nvidia before they can use them, so static inventory hurts them far more -- initially, at least. Eventually, poor sales mean that those firms just won't be buying any more chips from the jolly green giant, until stocks have cleared, and Nvidia is clearly banking on the market picking up over the year.

They likely thought: “if consumers were willing to spend so much on ampere, and Lovelace is better in performance and performance per watt, then consumers will pay more for Lovelace too right?”

Initial sales of the 4090 were high, though, especially in light of how expensive it was, so they probably felt justified, to begin with. Don't forget that Nvidia expected a significant downturn in the market and was possibly banking on having sufficient revenue until the dip hit. But then, of course, we had 12VHPWR-gate and 4080-but-actually-a-4070-gate.

As expensive as the 40 series of cards are, I don't think that the price is solely to blame for Nvidia not hitting its targets (after all, the current RX 7000 series of cards aren't exactly bargain-priced). For me, it's the dearth of models at various price points -- up until very recently, there has just been three Ada cards to choose from. Had Nvidia broken convention and launched an Ada card for every segment at the same time, I suspect it would have fared much better financially.

 

[Nanochip]

For Nvidia, that only applies to its own FE models. Companies like MSI, Gigabyte, Zotac, etc have to purchase the GPUs from Nvidia before they can use them, so static inventory hurts them far more -- initially, at least. Eventually, poor sales mean that those firms just won't be buying any more chips from the jolly green giant, until stocks have cleared, and Nvidia is clearly banking on the market picking up over the year.


Initial sales of the 4090 were high, though, especially in light of how expensive it was, so they probably felt justified, to begin with. Don't forget that Nvidia expected a significant downturn in the market and was possibly banking on having sufficient revenue until the dip hit. But then, of course, we had 12VHPWR-gate and 4080-but-actually-a-4070-gate.

As expensive as the 40 series of cards are, I don't think that the price is solely to blame for Nvidia not hitting its targets (after all, the current RX 7000 series of cards aren't exactly bargain-priced). For me, it's the dearth of models at various price points -- up until very recently, there has just been three Ada cards to choose from. Had Nvidia broken convention and launched an Ada card for every segment at the same time, I suspect it would have fared much better financially.

All nvidia had to do was keep the ampere MSRPs. They would’ve done very well and moved a ton of product. Lovelace is a good architecture. Tsmc is a good manufacturer. Lovelace is very power efficient. The pricing is what’s bad. And inventory is piling up.

Yeah nvidia charges Msi Asus et al for the gpu itself. That’s a fair point. But it’s not a good look for the premium GPU brand to have a mountain of inventory on store shelves. Nvidia seems to be a company that has EGO. So it has to hurt to know that gamers are rejecting Lovelace. I don’t think I’ve ever seen anything like this market before, in the past 10-15 years

Regarding rdna3, prices were too high there as well and amd allowed the 7900XT to decline in price. My issue with rdna3 is that it’s not as power efficient as Lovelace, and also I don’t like the high idle power consumption with dual high refresh rate monitors.

I had my eye on the sapphire nitro 7900 xtx but its power consumption is insane and it can consume more power than the 4090 under certain circumstances, while performing worse. No thanks. So for now I will stick with rdna2 unless the 4090 falls in price.

 

[waclark]

fair point. but what any corporation must keep in mind when setting prices is: there is no profit if the product is sitting in inventory. I went to micro center the other day and there is a ton of GPU product in inventory sitting in on the shelf.

Remember, there is elasticity of demand. If you chase profits above all else, and you set a price too high for the market to bear, then one likely outcome is high inventory and low sales, and ultimately low profits.

I'm quite familiar with elasticity of demand. I used to work in product marketing some time ago. The goal for any product is sell the most product at the highest profit. The thing is, increasing price can increase profits and even though you may sell less, you can maximize profits due to the cost of producing a product (manufacturing, shipping, support etc). So, when putting this into the context of Nvidia, will they sell more 4090s if they discount by $100 and will they sell enough to make up the $100 loss of revenue and profit?

The corporation must read the room and meet consumers where they’re at. Nvidia clearly got drunk off of mining demand and scalper prices. They likely thought: “if consumers were willing to spend so much on ampere, and Lovelace is better in performance and performance per watt, then consumers will pay more for Lovelace too right?”

Possibly true, but, there is also the whole, how much do customers want to spend on a GPU. $500 seems to be a sweet spot but no one is going to deliver 4090 performance for $500. At least not today.

But they clearly haven’t read the room. Much of the demand from last generation was artificial and unsustainable in my opinion, it was in large part mining demand, and a lot people who couldn’t travel, and wanted to game during the Covid lock downs.

But Covid is over and ethereum is now proof of stake and cryptocurrency demand has cooled off in general. So GPU demand is normalizing back towards being driven primarily by price-sensitive gamers.

And gamers clearly aren’t interested in the Lovelace tax. Which is good. This means nvidia must innovate on their end to keep production costs low, to be able to still make a good profit in a market where consumers aren’t interested in exorbitant prices.

I don't think the demand was 'artificial' per se. It was unexpected and unusual but there was a demand for the hardware. Did it make sense? Well, I'm not a miner, so I can't say if it made any sort of financial sense.

Actually I think AI is going to drive the next round of demand. And I wouldn't say gamers aren't interested in Lovelace, Nvidia is selling well compared to AMD. In fact, Nvidia regained some market share while AMD lost market share. What gamers aren't interested in is spending $800 or more on a GPU. That's why AMD's new models haven't sold as well as the last generation. That may change when they bring out their lower cost mid/low end GPUs.

Nvidia is managing the market by controlling inventory. Right now, there is a lot of inventory, but that will start to move once all the 30 series models are exhausted. Pricing should come down, and we are seeing some movement, though for myself, prices haven't dropped enough to pull the trigger on any GPU regardless of manufacturer. But, I do agree that Nvidia (and AMD) need to manage production cost because I do believe that this is partly why GPUs are priced where they are.

 

[Nanochip]

I'm quite familiar with elasticity of demand. I used to work in product marketing some time ago. The goal for any product is sell the most product at the highest profit. The thing is, increasing price can increase profits and even though you may sell less, you can maximize profits due to the cost of producing a product (manufacturing, shipping, support etc). So, when putting this into the context of Nvidia, will they sell more 4090s if they discount by $100 and will they sell enough to make up the $100 loss of revenue and profit?

Possibly true, but, there is also the whole, how much do customers want to spend on a GPU. $500 seems to be a sweet spot but no one is going to deliver 4090 performance for $500. At least not today.

I don't think the demand was 'artificial' per se. It was unexpected and unusual but there was a demand for the hardware. Did it make sense? Well, I'm not a miner, so I can't say if it made any sort of financial sense.

Actually I think AI is going to drive the next round of demand. And I wouldn't say gamers aren't interested in Lovelace, Nvidia is selling well compared to AMD. In fact, Nvidia regained some market share while AMD lost market share. What gamers aren't interested in is spending $800 or more on a GPU. That's why AMD's new models haven't sold as well as the last generation. That may change when they bring out their lower cost mid/low end GPUs.

Nvidia is managing the market by controlling inventory. Right now, there is a lot of inventory, but that will start to move once all the 30 series models are exhausted. Pricing should come down, and we are seeing some movement, though for myself, prices haven't dropped enough to pull the trigger on any GPU regardless of manufacturer. But, I do agree that Nvidia (and AMD) need to manage production cost because I do believe that this is partly why GPUs are priced where they are.

Fair points. Good talking with you. It will be interesting to see how or if nvidia will manage production costs, given that they likely want to move to 3nm for Blackwell next gen.

 

[waclark]

Fair points. Good talking with you. It will be interesting to see how or if nvidia will manage production costs, given that they likely want to move to 3nm for Blackwell next gen.

Yes, I hope Nvidia, AMD and Intel all figure it out. However, I do think we are in for a long wait to see high-end $500 GPUs. I've got my fingers crossed. LOL

 

[neeyik]

All nvidia had to do was keep the ampere MSRPs. They would’ve done very well and moved a ton of product.

That would require Nvidia to absorb the increased costs in fabrication, packaging, storage, distribution, etc. For example, the AD102 and AD103 are both only 3% smaller than the GA102 and GA104, so there's no significant improvement in maximum wafer yields, and they're made on a far more expensive node than that used for Ampere.

Now if one assumes that it was willing to do this, the only way that this could be internally justified would be through having plenty of evidence that it would sell lots of inventory -- it's the only way one can offset higher revenue costs to maintain profit margins. The general expectation, though, was that this wasn't going to happen, as every sign was pointing to a large downturn in the market. Nvidia's own Gaming sector figures kept dropping each quarter through FY2023, too.

Lower prices would certainly have helped improved sales figures, but with a $200 difference between the MSRPs of the RTX 4080 12GB (as it was called at the time) and 3080, they would need to be nearly 30% higher to generate an equivalent revenue. Could that have happened? Anyone's guess, really.

As an aside, the 4080 16GB has the same MSRP as the 3080 Ti, and the 4090 is $100 more than the 3090 (and $400 less than the 3090 Ti). Perhaps if Nivida gave the 4080 16GB a Ti name, running it with higher clocks than it currently does, and left the 4070 Ti as 4080 and made it, say, $100 more than the 3080 (rather than $200), perhaps matters right now would be a tad rosier for Nvidia.

 

[Nanochip]

That would require Nvidia to absorb the increased costs in fabrication, packaging, storage, distribution, etc. For example, the AD102 and AD103 are both only 3% smaller than the GA102 and GA104, so there's no significant improvement in maximum wafer yields, and they're made on a far more expensive node than that used for Ampere.

Now if one assumes that it was willing to do this, the only way that this could be internally justified would be through having plenty of evidence that it would sell lots of inventory -- it's the only way one can offset higher revenue costs to maintain profit margins. The general expectation, though, was that this wasn't going to happen, as every sign was pointing to a large downturn in the market. Nvidia's own Gaming sector figures kept dropping each quarter through FY2023, too.

Lower prices would certainly have helped improved sales figures, but with a $200 difference between the MSRPs of the RTX 4080 12GB (as it was called at the time) and 3080, they would need to be nearly 30% higher to generate an equivalent revenue. Could that have happened? Anyone's guess, really.

As an aside, the 4080 16GB has the same MSRP as the 3080 Ti, and the 4090 is $100 more than the 3090 (and $400 less than the 3090 Ti). Perhaps if Nivida gave the 4080 16GB a Ti name, running it with higher clocks than it currently does, and left the 4070 Ti as 4080 and made it, say, $100 more than the 3080 (rather than $200), perhaps matters right now would be a tad rosier for Nvidia.

The 3080 msrp was $699! That’s what got gamers so excited back in 2021. A $699 card that was much cheaper and faster than the previous generation’s dominant card, the 2080 TI! That set off a shockwave of excitement. But We know what happened to that fake “msrp.”

Now we have 4080 at $1199 that is almost twice as expensive as the 3080. There’s no excitement in that. What justifies the almost doubling of MSRP?

We have a 4070 that’s more expensive than the 3070. The 4090 is relatively close in price to the 3090 yes, but the majority of the market doesn’t buy the 4090.

The fact that nvidia raised prices significantly, was going to call the now 4070 TI a “4080” just makes nvidia appear to be anti consumer. Many gamers were very frustrated at the high prices and low availability of last generation, and were hoping for a reprieve; for prices to normalize and availability to improve. Unfortunately what happened is availability improved, but prices skyrocketed. And well unfortunately for nvidia and it’s partners, inventory is piling up.

This is probably why evga said “I’m out.”

 

[neeyik]

Now we have 4080 at $1199 that is almost twice as expensive as the 3080. There’s no excitement in that. What justifies the almost doubling of MSRP?

The performance metrics of the 4080 are roughly double that of the 3080, and it’s better than the 3080 Ti for the same MSRP. Perhaps in the mind of Nvidia’s bean counters that’s how the price is justified but let’s face it, multi-billion dollar companies aren’t pulling in multi-billion dollar revenues because they’ve put the needs and desires of the consumer first.

This is probably why evga said “I’m out.”

Profit margins in the whole third party PC component market are low — 15% isn’t unusual. The likes of Asus, MSI, Gigabyte, etc survive by having a huge portfolio, where small margin lines are bolstered by better performing ones; smaller firms like Zotac get by through OEM contracts. EVGA had neither, unfortunately.

 

[Nanochip]

The performance metrics of the 4080 are roughly double that of the 3080, and it’s better than the 3080 Ti for the same MSRP. Perhaps in the mind of Nvidia’s bean counters that’s how the price is justified but let’s face it, multi-billion dollar companies aren’t pulling in multi-billion dollar revenues because they’ve put the needs and desires of the consumer first.


Profit margins in the whole third party PC component market are low — 15% isn’t unusual. The likes of Asus, MSI, Gigabyte, etc survive by having a huge portfolio, where small margin lines are bolstered by better performing ones; smaller firms like Zotac get by through OEM contracts. EVGA had neither, unfortunately.

The 80 series card has never been a $1199 msrp. The 980, 1080, 2080, 3080 were all much cheaper. And to double the msrp of the last generation 80 card seems unjustified. It just seems like nvidia saw the scalpers making big profits in the ampere generation, and decided to become the real scalper now.

And well it’s backfiring. And I’m happy that gamers are voting “no” with their wallets. I certainly support intel amd nvidia making a profit. No doubt about it. But an $1199 4080 vs a $699 3080 just comes off as highway robbery. It reeks. In my humble opinion of course.

 

[Nobina]

What's crazy is that in spite of sales going down due to prices being inflated to high heavens, Nvidia is still against lowering prices, they hope people will get used to it, and many times they did, but it has to stop eventually.

 

[neeyik]

The 80 series card has never been a $1199 msrp.

The 3080 Ti was $1199 but that was during pandemic-authorized-scalping days, of course But yes, the 3080 was a good $500 lower than that — this is why I was saying that if Nvidia had released the 16GB version of the 4080 as a Ti model and the the 12GB 4080 as just a 4080 (with higher clocks on both), the prices would almost certainly been a lot more palatable. Unfortunately they went with the ‘Look at me, I’m the scalper now’ meme instead.

Nvidia is still against lowering prices, they hope people will get used to it, and many times they did, but it has to stop eventually.

I suspect it will be a long time, if ever, before Nvidia lowers prices — it makes so much from compute and AI, it can afford to wait it out.

 

[waclark]

/snip

Lower prices would certainly have helped improved sales figures, but with a $200 difference between the MSRPs of the RTX 4080 12GB (as it was called at the time) and 3080, they would need to be nearly 30% higher to generate an equivalent revenue. Could that have happened? Anyone's guess, really.

As an aside, the 4080 16GB has the same MSRP as the 3080 Ti, and the 4090 is $100 more than the 3090 (and $400 less than the 3090 Ti). Perhaps if Nivida gave the 4080 16GB a Ti name, running it with higher clocks than it currently does, and left the 4070 Ti as 4080 and made it, say, $100 more than the 3080 (rather than $200), perhaps matters right now would be a tad rosier for Nvidia.

Lower prices would help "if" they were low enough. I don't think $100 off a 4080 or 4090 would matter much. Even $200 off the 4080 still puts it at $1000 and that, I believe, is much more than most people want to spend on a GPU. Now, if the 4080 were priced like the 4070Ti and the Ti were priced like the 4070, it might be a different story. What we never seem to know, is what are the gross margins on these GPUs and how much can they realistically reduce prices? Can Nvidia make money on an $800 4080 or a $1200 4090?

 

[tellmewhy]

Bill Gates once had said that the company is permanently two years away from extinction.

This is something that applies to all technology companies. Nvidia's sin is that it deliberately degrades and nerf hardware in order to split the market. Intel did something similar with artificially locked multipliers and still hasn't escaped the danger.

Demand for mining occurs? LHR on cards and emergence of other hardware more expensive for mining only. Same with the demand for AI. It's gotten to the point where AI companies are designing their own chips.

When someone sells a couch they don't get the right to decide what the furniture layout will look like in someone else's house. Nvidia needs to make a line of two or three all-purpose PC cards and adjust to a competitive price point. It is its responsibility to keep production costs low and production capacity high. Maybe they should take the hasle and build fabs.

Anyone who thinks they are a irreplaceable can stick their finger in a glass of water, once they take it out they will get the answer.

When you have the top product and have all to dislike you, then you probably you make something wrong. People don’t dislike Lamborghini even if it is expensive and they can’t buy it.

It's very simple: if there's extra demand, raise the price, don't nerf the hardware. People are disgusted by nerfed hardware even on a subconscious level.

 

[neeyik]

Can Nvidia make money on an $800 4080 or a $1200 4090?

Sure, though the follow-up question is how much profit does a GPU company need to make, per item, to cover the growth in investment needed to remain on top of the game? Let’s say there’s a net profit of $300 on a $800 4080 and Nvidia gets 75% of that profit. $7.3 billion was spent on R&D in FY2023 and $5.3 in FY22 so if that 4080 was the only product made, 8.9 million cards would need to be sold for the profit to cover that increase in R&D expenditure.

Of course, far more products are actually sold but global shipments of all dGPUs (including AMD and Intel) in the last quarter of FY2022 was estimated to be 13 million. If we assume that the average is somewhere like 80 million per annum, and Nvidia sold 85% of those, then it would suggest that there is scope for prices to be that low. However, the bulk of those shipments are for low-mid and mid-range cards, as well as laptops, so the profit margins will be lower.