News Nvidia 4Q22 Earnings

[Hitman928]

First, Nvidia's official financial calendar makes this their 4Q23 earnings, but I am calling it 4Q22 to match the calendar time frame to avoid confusion.

Highlights:

4Q Revenue of $6.05B (-20.8% Y/Y)

Fiscal-year revenue of $27.0 billion, flat from a year ago

NVIDIA’s outlook for the first quarter of fiscal 2024 is as follows:

• Revenue is expected to be $6.50 billion, plus or minus 2% vs. $6.32B consensus
• GAAP and non-GAAP gross margins are expected to be 64.1% and 66.5%, respectively, plus or minus 50 basis points.

Nvidia Non-GAAP EPS of $0.88 beats by $0.08, revenue of $6.05B beats by $30M (NASDAQ:NVDA)

Nvidia press release (NVDA): Q4 Non-GAAP EPS of $0.88 beats by $0.08.Revenue of $6.05B (-20.8% Y/Y) beats by $30M.Fourth-quarter revenue was $3.62 billion, up 11% from a year ago...

seekingalpha.com

More to follow.

 

[Hitman928]

Nvidia is now a data center company.

Data Center

• Fourth-quarter revenue was $3.62 billion, up 11% from a year ago and down 6% from the previous quarter. Fiscal-year revenue rose 41% to a record $15.01 billion.

Gaming

• Fourth-quarter revenue was $1.83 billion, down 46% from a year ago and up 16% from the previous quarter. Fiscal-year revenue was down 27% to $9.07 billion.

Professional Visualization

• Fourth-quarter revenue was $226 million, down 65% from a year ago and up 13% from the previous quarter. Fiscal-year revenue was down 27% to $1.54 billion.

Automotive and Embedded

• Fourth-quarter revenue was a record $294 million, up 135% from a year ago and up 17% from the previous quarter. Fiscal-year revenue rose 60% to a record $903 million.

 

[Stuka87]

So nVidia's gamble that people would happily pay huge prices for GPUs didn't pay off. And well, not even having a mid range card certainly doesn't help things. Most of the 3K series chips are left overs from what we know. The market for a 1200-1600 dollar cards is rather small compared to the market for mid range cards.

 

[NTMBK]

Gaming down 46% YoY?! Good grief. Price cuts can't come soon enough.

 

[Hitman928]

So nVidia's gamble that people would happily pay huge prices for GPUs didn't pay off. And well, not even having a mid range card certainly doesn't help things. Most of the 3K series chips are left overs from what we know. The market for a 1200-1600 dollar cards is rather small compared to the market for mid range cards.

Not sure how much is PR bluff versus an honest belief, but:

"Gaming is recovering from the post-pandemic downturn, with gamers enthusiastically embracing the new Ada architecture GPUs with AI neural rendering," CEO Jensen Huang said in a statement accompanying the results.

 

[Ranulf]

Yes, yes... that is why there are new gaming sales every dang week almost. Even Modern Warefare 2 just had a sale dropping the base game to $45.

 

[Mopetar]

Their gaming figures are basically worthless when everyone knows they included huge numbers of cards sold directly to miners under the umbrella of gaming.

Trying to draw any conclusions from the data is pointless.

 

[Aapje]

Assets in inventory is double compared to last year.

 

[coercitiv]

Not sure how much is PR bluff versus an honest belief, but:

"Gaming is recovering from the post-pandemic downturn, with gamers enthusiastically embracing the new Ada architecture GPUs with AI neural rendering," CEO Jensen Huang said in a statement accompanying the results.

It's time to remind everybody about Jensen's core belief relative to the first wave of Ada products (found via SkaterBencher):

4080 targets a very large segment of the GPU market (enthusiasts & core gamers) which tend to not be affected by market conditions here or there. So I expect Ada's launch to be very successful

Anybody who wants to listen to his entire answer: https://video.ibm.com/recorded/132127846 at 35m30s

Like @Aapje mentioned, assets in invetory are growing but gamers enthusiastically embrace... probably with thoughts and prayers. /s

 

[maddie]

Assets in inventory is double compared to last year.

I think everyone is betting on a 2nd half sales rebound. That's the new mantra. We''ll see how it works.

 

[Qwertilot]

Gaming down 46% YoY?! Good grief. Price cuts can't come soon enough.

Wasn't it massively up last year or two though? Looks more like a reversion towards the long term mean than anything actually very worrying.
(Some combination of Crypto & Covid you have to presume.).

 

[Aapje]

I think everyone is betting on a 2nd half sales rebound. That's the new mantra. We''ll see how it works.

I wonder how much impact AI has. I worry a bit that it may be mining-lite, creating a large demand for GPU's until specialized processors take over, which I think will happen.

But I don't expect a recovery in sales to consumers at these prices. Why would it?

Wasn't it massively up last year or two though? Looks more like a reversion towards the long term mean than anything actually very worrying.

Sales have actually been fairly modest for the last 2 years, presumably due to manufacturing problems. So the prices were high, but not so much the number of cards sold. They were much higher during the 2017/2018 boom, where production could apparently increase much more.

And the long term trend IS worrying, since sales have been declining for a long time now: https://www.tomshardware.com/news/sales-of-desktop-graphics-cards-hit-20-year-low

It is also quite possible that we'll see a much faster decline now that prices are going up so fast. It's quite an assumption that when prices suddenly go up by 70%, we'll return to the sales numbers from when prices per tier increased by about inflation.

 

[Stuka87]

I wonder how much impact AI has. I worry a bit that it may be mining-lite, creating a large demand for GPU's until specialized processors take over, which I think will happen.

But I don't expect a recovery in sales to consumers at these prices. Why would it?

I think this is unlikely. These big AI companies aren't going to start buying gaming GPUs for their AI systems. The amount of AI processing power in a gaming GPU is pretty low in comparison to a Tesla card which is entirely tensor cores.

 

[Dribble]

I think this is unlikely. These big AI companies aren't going to start buying gaming GPUs for their AI systems. The amount of AI processing power in a gaming GPU is pretty low in comparison to a Tesla card which is entirely tensor cores.

AI and datacenter will take fab space that won't be used to make gaming chips so gaming will be effected if there is a fab shortage which there often is on the high end nodes these things are built on. Less chips means they can get away with charging higher prices and still sell them all. That said it's better then Nvidia going into the red and having no money to develop chips.

As for gaming figures being bad vs a year ago, well you've got to discount mining sales, if you take them out I suspect gaming sales are actually pretty good. I don't think their % of the total market has dropped.

Tbh the results are impressive - I think many were expecting doom and gloom as the post mining figures would show huge drops for Nvidia, but infact they've covered them with the datacenter market, which unlike mining isn't a 1-2 year boom-bust market. You may not like him if you just want cheap gaming gear but you gotta respect JHH's ability to lead a company.

 

[Thibsie]

The only real ability I admit concerning JHH is his capacity to lie to investors and get away with it.

 

[scineram]

Gaming is already up 16% sequentially while AMD was down. All your copium tears will be delicious very soon.

 

[Aapje]

Gaming is already up 16% sequentially while AMD was down. All your copium tears will be delicious very soon.

The sequential numbers are pretty meaningless when you compare the holiday quarter to a historically much weaker quarter, Ada was launched in that quarter & Nvidia has been undershipping, so they could just have undershipped more in Q3.

 

[DooKey]

Everyone spent their free money over the last couple of years and either doesn't need/want a new gaming card or is totally broke. Welcome to the new normal. NV will continue to make bank on AI/data center and play silly games with the gaming market from now on. I don't expect AMD to do much more than they are right now either when it comes to discrete GPUs.

 

[Mopetar]

AMD trying to shake up the status quo depends on two important factors. The first is that they have a competitive card. RDNA2 showed they could compete with NVidia again, but RDNA3 stumbled out of the gate. If they can't compete with NVidia they have to accept the also ram position. The second is that they need to have production capacity to make sufficient quantities of chips. If they're wafer limited there's little sense in cutting prices to try to gain market share when they'll constantly not have any cards available to sell.

If AMD gets into a position where they have both a competitive product as well as the capability to produce it in significant volume, only then will they price aggressively to pick up extra market share. Otherwise it's not worth the effort and pricing below NVidia and collecting more money is a better financial move.

 

[Stuka87]

AMD trying to shake up the status quo depends on two important factors. The first is that they have a competitive card. RDNA2 showed they could compete with NVidia again, but RDNA3 stumbled out of the gate. If they can't compete with NVidia they have to accept the also ram position. The second is that they need to have production capacity to make sufficient quantities of chips. If they're wafer limited there's little sense in cutting prices to try to gain market share when they'll constantly not have any cards available to sell.

If AMD gets into a position where they have both a competitive product as well as the capability to produce it in significant volume, only then will they price aggressively to pick up extra market share. Otherwise it's not worth the effort and pricing below NVidia and collecting more money is a better financial move.

How did RDNA stumble out of the gate? The 7900s were never intended to compete with the 4090. And they compete well enough with the 4070Ti/4080. In some games RT performance on AMD is down, but in others it is equal or better.

The single biggest issue AMD has to overcome is mindshare. Just too many ignorant people that will buy nVidia because they have been told they are the best, regardless of what the facts actually say.

AMD had to blow intel out of the water on Price to Performance in order to gain the CPU marketshare that they have now. It would take at least that to claw back marketshare from nVidia.

 

[jpiniero]

The 7900s were never intended to compete with the 4090.

It absolutely was. As to why it's not remains to be seen.

 

[Aapje]

The 7900s were never intended to compete with the 4090.

I'm sure that in hindsight that was suddenly true...

But it makes no sense to think that AMD had the plan to fall further back, rather than gain ground.

 

[leoneazzurro]

It absolutely was. As to why it's not remains to be seen.

They used too little silicon for competing with the >600mm^2 GA102 which in turn was also on a denser process. Yes, AMD tried to do an extremely area optimized architecture but even so, they used a smaller total area over two different processes and half the 5N area. So it's quite possible AMD did not aim to overcome the top Nvidia chip, but to undercut it in costs. Top end cards costing >1,5K$ or more aren't the top priority in the market, they are good for "halo" effect but the BIAS is so high that no halo would have pushed AMD sales.

 

[Aapje]

@leoneazzurro

It's not just about the die size, especially since GA102 is also a professional product, so it's a compromise and not optimized fully for gaming. They might also have missed the frequency target. Or they may simply have underestimated how much Samsung was holding Nvidia's design back.

 

[leoneazzurro]

@leoneazzurro

It's not just about the die size, especially since GA102 is also a professional product, so it's a compromise and not optimized fully for gaming. They might also have missed the frequency target. Or they may simply have underestimated how much Samsung was holding Nvidia's design back.

It is as much of a professional product as Navi31 or GA103 are, all of these will lead to "Pro" CAD SKUs and GA102 shares the architecture with the GA103/104/106/107. I don't know of any "pro" feature that is GA102 only. And yes, AMD probably missed their clock targets but even if they met them, it is unlikely they could have matched the GA102 with just more than half the equivalent 5N area...