Tuesday, November 21st 2023
![NVIDIA](https://tpucdn.com/images/news/nvidia-v1723816949774.png)
NVIDIA Announces Q3-FY24 Results; Earns 5 Times More Revenue from AI Chips than Gaming GPUs
NVIDIA (NASDAQ: NVDA) today reported revenue for the third quarter ended October 29, 2023, of $18.12 billion, up 206% from a year ago and up 34% from the previous quarter. GAAP earnings per diluted share for the quarter were $3.71, up more than 12x from a year ago and up 50% from the previous quarter. Non-GAAP earnings per diluted share were $4.02, up nearly 6x from a year ago and up 49% from the previous quarter.
"Our strong growth reflects the broad industry platform transition from general-purpose to accelerated computing and generative AI," said Jensen Huang, founder and CEO of NVIDIA. "Large language model startups, consumer internet companies and global cloud service providers were the first movers, and the next waves are starting to build. Nations and regional CSPs are investing in AI clouds to serve local demand, enterprise software companies are adding AI copilots and assistants to their platforms, and enterprises are creating custom AI to automate the world's largest industries."NVIDIA GPUs, CPUs, networking, AI foundry services and NVIDIA AI Enterprise software are all growth engines in full throttle. The era of generative AI is taking off," he said.
NVIDIA will pay its next quarterly cash dividend of $0.04 per share on December 28, 2023, to all shareholders of record on December 6, 2023.
Outlook
NVIDIA's outlook for the fourth quarter of fiscal 2024 is as follows:
NVIDIA achieved progress since its previous earnings announcement in these areas:
Data Center
Commentary on the quarter by Colette Kress, NVIDIA's executive vice president and chief financial officer, is available at https://investor.nvidia.com.
"Our strong growth reflects the broad industry platform transition from general-purpose to accelerated computing and generative AI," said Jensen Huang, founder and CEO of NVIDIA. "Large language model startups, consumer internet companies and global cloud service providers were the first movers, and the next waves are starting to build. Nations and regional CSPs are investing in AI clouds to serve local demand, enterprise software companies are adding AI copilots and assistants to their platforms, and enterprises are creating custom AI to automate the world's largest industries."NVIDIA GPUs, CPUs, networking, AI foundry services and NVIDIA AI Enterprise software are all growth engines in full throttle. The era of generative AI is taking off," he said.
NVIDIA will pay its next quarterly cash dividend of $0.04 per share on December 28, 2023, to all shareholders of record on December 6, 2023.
Outlook
NVIDIA's outlook for the fourth quarter of fiscal 2024 is as follows:
- Revenue is expected to be $20.00 billion, plus or minus 2%.
- GAAP and non-GAAP gross margins are expected to be 74.5% and 75.5%, respectively, plus or minus 50 basis points.
- GAAP and non-GAAP operating expenses are expected to be approximately $3.17 billion and $2.20 billion, respectively.
- GAAP and non-GAAP other income and expense are expected to be an income of approximately $200 million, excluding gains and losses from non-affiliated investments.
- GAAP and non-GAAP tax rates are expected to be 15.0%, plus or minus 1%, excluding any discrete items.
NVIDIA achieved progress since its previous earnings announcement in these areas:
Data Center
- Third-quarter revenue was a record $14.51 billion, up 41% from the previous quarter and up 279% from a year ago.
- Announced NVIDIA HGX H200 with the new NVIDIA H200 Tensor Core GPU, the first GPU with HBM3e memory, with systems expected to be available in the second quarter of next year.
- Introduced an AI foundry service—with NVIDIA AI Foundation Models, NVIDIA NeMo framework and NVIDIA DGX Cloud AI supercomputing — to accelerate the development and tuning of custom generative AI applications, first available on Microsoft Azure, with SAP and Amdocs among the first customers.
- Announced that the NVIDIA Spectrum-X Ethernet networking platform for AI will be integrated into servers from Dell Technologies, Hewlett Packard Enterprise and Lenovo in the first quarter of next year.
- Announced that NVIDIA GH200 Grace Hopper Superchips, including a new quad configuration, will power more than 40 new supercomputers, including the JUPITER system at Jülich Supercomputing Centre and Isambard-AI at the University of Bristol.
- Made advances with global cloud service providers:
- Google Cloud Platform made generally available new A3 instances powered by NVIDIA H100 Tensor Core GPUs and NVIDIA AI Enterprise software in Google Cloud Marketplace.
- Microsoft Azure will be offering customers access to NVIDIA Omniverse Cloud Services for accelerating automotive digitalization, as well as new instances featuring NVL H100 Tensor Core GPUs and H100 with confidential computing, with H200 GPUs coming next year.
- Oracle Cloud Infrastructure made NVIDIA DGX Cloud and NVIDIA AI Enterprise software available in Oracle Cloud Marketplace.
- Partnered with a range of leading companies on AI initiatives, including Amdocs, Dropbox, Foxconn, Genentech (member of Roche Group), Infosys, Lenovo, Reliance Industries, Scaleway and TATA Group.
- Announced record-setting performance in the latest two sets of MLPerf benchmarks for inference and training, with the NVIDIA Eos AI supercomputer training a GPT-3 model 3x faster than the previous record.
- Announced growing worldwide support for the NVIDIA CUDA Quantum platform, including new efforts in Israel, the Netherlands, the U.K. and the U.S.
- Third-quarter revenue was $2.86 billion, up 15% from the previous quarter and up 81% from a year ago.
- Launched DLSS 3.5 Ray Reconstruction, which creates high-quality ray-traced images for intensive ray-traced games and apps, including Alan Wake 2 and Cyberpunk 2077.
- Released TensorRT-LLM for Windows, speeding on-device LLM inference by up to 4x.
- Added 56 DLSS games and over 15 Reflex games, bringing the total number of RTX games and applications to over 475.
- Surpassed 1,700 games on GeForce NOW, including launches of Alan Wake 2, Baldur's Gate 3, Cyberpunk 2077: Phantom Liberty, Forza Motorsport and Starfield.
- Third-quarter revenue was $416 million, up 10% from the previous quarter and up 108% from a year ago.
- Announced that Mercedes-Benz is using NVIDIA Omniverse to create digital twins to help plan, design, build and operate its manufacturing and assembly facilities around the world.
- Announced a new line of desktop workstations with NVIDIA RTX 6000 Ada Generation GPUs and NVIDIA ConnectX smart interface cards for training smaller AI models, fine-tuning models and running inference locally.
- Third-quarter revenue was $261 million, up 3% from the previous quarter and up 4% from a year ago.
- Furthered its collaboration with Foxconn to develop next-generation electric vehicles for the global market, using the next-generation NVIDIA DRIVE Hyperion platform and NVIDIA DRIVE Thor system-on-a-chip.
Commentary on the quarter by Colette Kress, NVIDIA's executive vice president and chief financial officer, is available at https://investor.nvidia.com.
55 Comments on NVIDIA Announces Q3-FY24 Results; Earns 5 Times More Revenue from AI Chips than Gaming GPUs
Doubt it. I believe it is triple digit.
There are buyers, but let's not pretend gamers are buying RTX 4090 a year and a half after it's release for 10 - 20% over MSRP!
I bet anyone that Nvidia will fall below Intel and AMD in market value within the next year, they are just easily replaceable in the main markets they operate in. :P
The only possibility of Nvidia stop selling gaming GPU is when PC gaming died out, which honestly will not be happening any time soon.
Consumers paid for and beta tested Nvidia's AI hardware for years giving it the resources to develop this reality.
They're not giving that up.
Clearly, tongue-in-cheek. But sort of relevant.
Market cap
NVDA $1.22T
AMD $200B
INTC $185B
Nvidia would have to lose about 85% of its value to drop below AMD if the latter stayed even. However if NVDA dropped that much it would drag down the entire semiconductor sector as well as make a major dent in broad market indexes like the S&P 500, Nasdaq Composite and NASDAQ-100.
Nvidia does not function in a vacuum.
As for putting my money where my mouth is, I have indirect positions in all three via a variety of index funds.
AI isn’t going away, I will make money regardless who has market dominance a year from now.
Remember that revenues equals shipped products and paid services. Nvidia’s inbox is stacked with purchase orders, probably enough to keep TSMC busy for two, three years.
This isn’t Joe Gamer standing in front of a store shelf at Microcenter debating whether or not to buy a GeForce 4070.
Companies budget for these purchases in advance.
It's a bit like Toyota explaining that even if the west want to go full electric, there's a lot of places where it's just not going to work. You need alternative technologies to cover the weakness of the other one
Tensor cores, useful Eventually, occasionally.
Raytracing core's useful Eventually and occasionally
Dlss all upto version blah blah blah
Many other things time's IMHO.
So yeah as an owner/Beta tester yes.
In one regards it is really exciting, but then in another it is kinda scary. We will get Blackwell, but it could be limited and expensive. AMD is skimping out on RDNA4. AMD is going to focus on their CPUs and I'd imagine put their GPU resources behind CDNA. ATM, gaming isn't the juggernaut in terms of revenue that it used to be and it obviously isn't where the money is. So I can see options being reduced (everyone), prices going up (Nvidia), and development slowing down (AMD).
Maybe if GPU cores go chiplet it won't be too bad.
I choose not to engage in the seemingly endless debate of Nvidia vs AMD vs Intel. My conviction in my perspective led me to exit at $500. If you genuinely believe in the positive prospects, consider purchasing the stock at this price point. Opinion is subjective, reality is not, and it eventually imposes itself by subjugating individual opinions.
The future will show who was right. Good luck!
First of all, I addressed your statement that NVDA would drop below AMD and INTC in valuation within a year. I showed market cap and you won't accept raw numbers. That's fine. But this is ON YOU.
Second, yes, you chose to engage in the endless debate of Nvidia vs. everyone else via your previous post. You mentioned no exit at $500.
And guess what? Unless your dollars are fully invested in something that returns better than NVDA, well, you lose. I know. I bailed out of NFLX after owning it for 18 months and it climbed through the stratosphere. Sure, I made money but not the same amount if I kept it.
We all have to sleep though. If you can't handle the suspense, just say so. No one can blame you for bailing out to get enough shut eye. But don't misconstrue it with making more money.
Third, even if NVDA doesn't have the same price acceleration in the next twelve months as it has in the previous 12, it's still more than losing 85% of market cap value, isn't it?
My money is that that in 12 months NVDA's market cap will still be higher than that of AMD or INTC.
You proposed something else and I might just bookmark this discussion just to revisit it in a year to see where we stand. Let me know if you'd like me to ping this thread a year from now.
Good try trying to cover your sorry ass on what you posted. Since I included your post, you can't back out. Everyone reading this thread will understand what sort of odds you are up against.
Feel free to respond to this, you will only dig a deeper grave for yourself. Remember: you are on the record for proposing that NVDA will have a lower market capitalization on November 22, 2024 than AMD. Technically, you stated that NVDA would be lower than both AMD and INTC. But let's just start with AMD. We get plenty of garbage chat here at TPU but this one is worth highlighting just due to its complete ludicrousness.
Best of luck.
If we were going to take that route and follow that logic, I'd argue that FSR 1/2/3 and Antilag+, perhaps others - they're beta features the users are testing, just on a smaller scale because they lack the vision to foresee (most of) these prospective innovations. Consumers paid for and beta tested AMD's reactionary hardware and software solutions for years giving AMD the resources to just barely stay in the PC space. Seen a fair few Radeon's sold, or purchases delayed, on hopium that they can match Nvidia's hardware and features, so Nvidia's features can't be all that bad if AMD aspires to meet or exceed them.
Consumers rarely can tell or steer the market in how to innovate, they want more, faster, better, but largely of known quantities. A 5090 that has 48GB and is twice as fast in every metric to a 4090? sounds great right? (price excluded for point making purpose), A monitor that's higher res and faster refresh, awesome! But average Joe would be hard pressed to predict or ask for the newer features and capability GPU's have come out with over the years. RT, Upscaling, Mesh Shading, Tessellation, and so on.
The fact of the matter is that these features are the very example of "when life gives you lemons, make lemonade". NVIDIA chose to dedicate more die space on its GPUs to ML- and RT-accelerating hardware because they didn't want to have to endure the cost and complexity of producing different GPUs for consumer and professional workloads. This hardware took away from raster hardware, which created a problem that their consumer products might no longer be able to compete on sheer performance. Instead of sitting back and hoping that wouldn't bite them, NVIDIA came up with feature solutions to leverage that hardware to add value for consumers. And they've been so successful in that regard, that those features have become the industry standard such that every single one of their competitors now has their own version of said features. That's not because those competitors slavishly copy whatever NVIDIA does, it's because consumers now expect those features as standard.
Basically, NVIDIA invented an entire new category of consumer graphic features entirely because they wanted to save money, and they've utilised those features so successfully that they've increased their market leadership. That's the very opposite of "beta", that's "astounding success", and your irrational hatred doesn't change that; it just makes you look ridiculous.
I've said it before and I'll say it again - NVIDIA's self-proclaimed moniker of "World Leader in Visual Computing Technologies" is incredibly pompous but entirely justified. They lead, they succeed, and others follow.
Nvidia does not necessarily need to lose so much market value to fall below Intel and AMD, they just need to both grow while Nvidia continually weakens with the rise of competition and battles to make AI cheaper.
Furthermore, I just don't invest in what I don't believe in, It doesn't matter if banks or analysts say otherwise. If I lose or win, it will be by following what makes sense to me. GG.
Who do you think you are trolling.
Nvidia trolled me .
I sat through months of no use of features because f all used them, BeTA.
I also tested most iterations of dlss because they can't make one right first time and guess what none are at iterative one because we beta tested them found issues for Nvidia.
I'm sorry this insults your master but deal with it.
Neither Intel nor AMD have been able to compete effectively with NVIDIA in the ML space and there's zero indication they'll be able to anytime soon. So to claim that both of those companies will somehow be able to grow by a factor of more than three times (which is what would need to happen for them to surpass NVIDIA's market cap), isn't just laughable - it's so absurd that your post is literally hurting my brain. And that more than three times figure is also only relevant if NVIDIA themselves doesn't grow at all in the same period, which ain't gonna happen.
Now, I'd agree that NVIDIA is overvalued right now, and that there's going to be a correction eventually. But even if their stock drops to the ~$200 it was at the beginning of this year, they'd have a market cap of ~$488 billion which still massively exceeds AMD and Intel combined.
But go on, keep digging that hole. How exactly does it matter how many units were sold? Even if you sell only one unit at 100000000% profit, you still made a lot of money. I'm sorry that I wasted my time arguing in good faith, when you evidently aren't capable of doing the same.