News Nvidia 4Q23 Results

[Hitman928]

Nvidia follows a different fiscal schedule but I'm equating to the calendar year time frame that most others use.

Another monster quarter and projection from NV.

• Nvidia (NASDAQ:NVDA): Q4 Non-GAAP EPS of $5.16 beats by $0.52.
• Revenue of $22.1B (+265.3% Y/Y) beats by $1.55B.
• Data Center revenue for the fourth quarter was a record, up 409% from a year ago and up 27% sequentially.
• Gaming revenue was up 56% from a year ago and flat sequentially.
• Professional Visualization revenue was up 105% from a year ago and up 11% sequentially
• Automotive revenue was down 4% from a year ago and up 8% sequentially.
• Outlook for the first quarter of fiscal 2025 is as follows: • Revenue is expected to be $24.0 billion vs. $22.03B consensus, plus or minus 2%. • GAAP and non-GAAP gross margins are expected to be 76.3% and 77.0%, respectively, plus or minus 50 basis points. • GAAP and non-GAAP operating expenses are expected to be approximately $3.5 billion and $2.5 billion, respectively. • GAAP and non-GAAP other income and expense are expected to be an income of approximately $250 million, excluding gains and losses from non-affiliated investments. • GAAP and non-GAAP tax rates are expected to be 17.0%, plus or minus 1%, excluding any discrete items.

Nvidia Non-GAAP EPS of $5.16 beats by $0.52, revenue of $22.1B beats by $1.55B (NASDAQ:NVDA)

Nvidia (NVDA) reports impressive Q4 results with Non-GAAP EPS beating estimates by $0.52 and revenue exceeding expectations by $1.55B, but shares drop 3%.

seekingalpha.com

 

[tajoh111]

Impressive beat again. Freakish profits worthy of a company worth a trillion dollars(perhaps not two trillion). What adds to this feat is they beat outlooks that appeared to be generous to begin with and they have done this quarter after quarter.

What impresses me with the strong Q1 outlook is the numbers are like AMD is not even in the picture when they should be. Mi300 is supposed to steal marketshare, but with a outlook of 24 billion dollars, we are looking at over 20 billion dollars in data center revenue. Hence AMD is not effecting Nvidia's sales at all. The beat by analysts is the size of AMD's entire data center revenue for q1 data center.

This 20 billion is close to 4x AMD's entire revenue for q1 which is pegged at 5.4 billion(AMD's has not been beating expectations either so 5.4 billion is what your are likely to see in Q1). Is Q2 where we see the impact of MI300? I think Lisa's Su 3 billion dollar estimate for MI300 is actually on the accurate side not a lowball as some are suggesting. With Blackwell datacenter being unveiled at the end of Q2 with the launch sometime at the end of q3, MI300 is going to be squashed at this rate.

 

[Tigerick]

It's Intel's DC getting overthrown by NV. In term of revenue, Intel DC's revenue has been dropping every year since 2021, that's why AMD might overtake Intel's DC revenue this year if they are executed properly....

 

[CP5670]

I wonder how much nvidia will care about gaming cards in the future. The datacenter business is going to be the bulk of their business going forward. Gamers will only get cut down cards because the fully enabled chips can be sold for 20x as much to datacenters.

 

[Hitman928]

Impressive beat again. Freakish profits worthy of a company worth a trillion dollars(perhaps not two trillion). What adds to this feat is they beat outlooks that appeared to be generous to begin with and they have done this quarter after quarter.

What impresses me with the strong Q1 outlook is the numbers are like AMD is not even in the picture when they should be. Mi300 is supposed to steal marketshare, but with a outlook of 24 billion dollars, we are looking at over 20 billion dollars in data center revenue. Hence AMD is not effecting Nvidia's sales at all. The beat by analysts is the size of AMD's entire data center revenue for q1 data center.

This 20 billion is close to 4x AMD's entire revenue for q1 which is pegged at 5.4 billion(AMD's has not been beating expectations either so 5.4 billion is what your are likely to see in Q1). Is Q2 where we see the impact of MI300? I think Lisa's Su 3 billion dollar estimate for MI300 is actually on the accurate side not a lowball as some are suggesting. With Blackwell datacenter being unveiled at the end of Q2 with the launch sometime at the end of q3, MI300 is going to be squashed at this rate.

AMD's AI solution is just barely starting to ramp. They won't ship big numbers until the second half of the year. The competition for market share will start then and should really go into full swing when the next gen from both are out. I don't expect Blackwell will be shipping in large numbers until the very end of the year at the earliest. Even then, if industry predictions are correct, it seems like there will be so much demand that even both AMD and NV might have a hard time keeping up with it together. I expect even Intel will get some play because of this though in much smaller numbers than AMD, let alone NV.

 

[tajoh111]

IF AMD is only ramping up in 2nd of of this year, that's sign of trouble.

MI300 was supposed to be a Q1, 2023 product according to some here. If AMD can only deliver these product in high number 18 months later, any super large company would be hesitant on purchasing MI300 simply because timing is so critical for these companies along with having to do so much work on their own.

Nvidia track record with B2B has allowed them to take 90% of the marketshare of professional visualization, maybe more. What has allowed Nvidia to succeed on both business and game developers is support, being able to send their own people to help developers. This is likely to be happening with their data center product and along with their ability to ramp up and deliver products, provides roadblocks which I suspect why AMD revenue has not been creeping up that much.

 

[Hitman928]

IF AMD is only ramping up in 2nd of of this year, that's sign of trouble.

MI300 was supposed to be a Q1, 2023 product according to some here. If AMD can only deliver these product in high number 18 months later, any super large company would be hesitant on purchasing MI300 simply because timing is so critical for these companies along with having to do so much work on their own.

Nvidia track record with B2B has allowed them to take 90% of the marketshare of professional visualization, maybe more. What has allowed Nvidia to succeed on both business and game developers is support, being able to send their own people to help developers. This is likely to be happening with their data center product and along with their ability to ramp up and deliver products, provides roadblocks which I suspect why AMD revenue has not been creeping up that much.

MI300, especially X, was never a Q123 product. MI300X didn't release until Q124 and I don't believe was originally meant to be a mainline product initially, that was the MI300A which had interest from the HPC crowd. When AI had the sudden surge in demand, AMD pushed the MI300X to the premium positioning. Given how short of time they had to both sample the product for validation and ramp the product for extremely high volume (by AMD's standards), I don't see how you could see the release as being troubled. AMD will go from low hundreds of millions of dollars in DC GPU sales (in a good quarter) to over a billion a quarter by the end of the year, worst case. They are basically multiplying their revenue by 10x within 12 - 15 months and it will be AMD's fastest ramping product ever and it's not even close. It will still be well below NV's at the end of the year thanks largely to both having products ready when the boom happened and their years of groundwork leading up to the boom, but to paint AMD's progress as troubled is just plain ignorant.

 

[IEC]

This amount of brute force GPU power will only be needed during AI infancy. I don't see this lasting forever (<5 years), though obviously nV and AMD will both benefit from this in the near term. nV more than anyone else.

Intel is looking irrelevant, which is a bad place to be when the AI TAM is basically bigger than everything else currently...

 

[tajoh111]

The time for this datacenter boom is likely short and every quarter where AMD doesn't not have amazing revenue is a lost opportunity for revenue which it needs to counteract the decline in consoles and client products. 3 billion dollars seems like a lot of money but puny compared to the 80 billion dollars Nvidia is going to be getting in the same market. More over, 3 billion isn't much considering what is happening with the rest of AMD's product stack.

X box sales have been bad and Playstation 5 sales are beginning to stumble. Client products's are also doing poorly.

PS5 Entering ‘The Latter Stage of Its Life Cycle’, Sony Says - IGN

After sales of the PlayStation 5 failed to meet expectations, Sony has said the console is entering “the latter stage of its life cycle.”

www.ign.com

Console chips represent about 5 to 6 billion of AMD's revenue at the moment and with client products not selling well, AMD needs something to take up the slack. Last year, client products dropped 1.6 billion dollars vs 2022 and caused a loss for AMD.

When you factor in these things, 3 billion in additional revenue may not be able to off set lost revenue from the rest of AMD. AMD needs more money to fight Nvidia in this market and just in general. Intel is seemingly getting their foundries in order and Nvidia is spending more at TSMC than AMD which is not good since AMD's has been heavily reliant on TSMC for it's current position in the market.

It would not surprise me if AMD 2024 revenue is basically the same as 2023. On the opposite side, with the financial horsepower Nvidia at it's disposal(100 billion in revenue vs 23 billion), they can sabotage AMD's efforts to get in the AI Datacenter. For example, I suspect all the HBM3e supply in the market has been purchased by Nvidia, Skhynix said it's total allocation of HBM is sold out for 2024. The reason for this is certainly Nvidia. If they do the same with HBM4 too, AMD will be at a critical disadvantage to compete with Nvidia.

Next quarter, Nvidia's cost of good sold is going to be greater than AMD's entire revenue. That is they will be spending 6 plus billion dollars when AMD's entire revenue is 5.4 billion dollars. The AI market is going to be a struggle to compete in simply because Nvidia is such a competent competitor that can do dubious things at times.

AMD has to be careful it does not pull an intel like move by pouring billion in a market(larrabee or the Intel android products), that will cripple it in the long run. 3 billion in annual income is not enough to change AMD future with what else is going on at AMD. If the AI bubble pops in 2025 sometime and the rest of AMD's products are held back because of the investment in instincts products, AMD may be in an Intel like position. Nvidia will also feel this but more so in it's stock price than anything.

Nvidia will get pounded in the stock market when the bubble bursts but with the 10's of billions of dollars of accumulated profits(not revenue, profit), they will be in a comfortable position for it's future.

 

[Hitman928]

The time for this datacenter boom is likely short and every quarter where AMD doesn't not have amazing revenue is a lost opportunity for revenue which it needs to counteract the decline in consoles and client products. 3 billion dollars seems like a lot of money but puny compared to the 80 billion dollars Nvidia is going to be getting in the same market. More over, 3 billion isn't much considering what is happening with the rest of AMD's product stack.

X box sales have been bad and Playstation 5 sales are beginning to stumble. Client products's are also doing poorly.

PS5 Entering ‘The Latter Stage of Its Life Cycle’, Sony Says - IGN

After sales of the PlayStation 5 failed to meet expectations, Sony has said the console is entering “the latter stage of its life cycle.”

www.ign.com

Console chips represent about 5 to 6 billion of AMD's revenue at the moment and with client products not selling well, AMD needs something to take up the slack. Last year, client products dropped 1.6 billion dollars vs 2022 and caused a loss for AMD.

When you factor in these things, 3 billion in additional revenue may not be able to off set lost revenue from the rest of AMD. AMD needs more money to fight Nvidia in this market and just in general. Intel is seemingly getting their foundries in order and Nvidia is spending more at TSMC than AMD which is not good since AMD's has been heavily reliant on TSMC for it's current position in the market.

It would not surprise me if AMD 2024 revenue is basically the same as 2023. On the opposite side, with the financial horsepower Nvidia at it's disposal(100 billion in revenue vs 23 billion), they can sabotage AMD's efforts to get in the AI Datacenter. For example, I suspect all the HBM3e supply in the market has been purchased by Nvidia, Skhynix said it's total allocation of HBM is sold out for 2024. The reason for this is certainly Nvidia. If they do the same with HBM4 too, AMD will be at a critical disadvantage to compete with Nvidia.

Next quarter, Nvidia's cost of good sold is going to be greater than AMD's entire revenue. That is they will be spending 6 plus billion dollars when AMD's entire revenue is 5.4 billion dollars. The AI market is going to be a struggle to compete in simply because Nvidia is such a competent competitor that can do dubious things at times.

AMD has to be careful it does not pull an intel like move by pouring billion in a market(larrabee or the Intel android products), that will cripple it in the long run. 3 billion in annual income is not enough to change AMD future with what else is going on at AMD. If the AI bubble pops in 2025 sometime and the rest of AMD's products are held back because of the investment in instincts products, AMD may be in an Intel like position. Nvidia will also feel this but more so in it's stock price than anything.

Nvidia will get pounded in the stock market when the bubble bursts but with the 10's of billions of dollars of accumulated profits(not revenue, profit), they will be in a comfortable position for it's future.

You seem to be very obsessed about trashing AMD in a thread created specifically about NV.