As usual, Nvidia beat Wall Street expectations, both in terms of Q2 2024 results (the quarter ending July 31, 2024), but also provided guidance for the next quarter above market expectations.. and it has been the case for many quarters … see Figure 1.
Looking at the main market narratives and investor sentiment … regardless of the earnings beat, probably Nvidia’s stock price would have fallen anyway.
Two days after the results were released, Nvidia is down 4.97% (while the S&P500 is up 1.01%). Relatively, Nvidia is losing 5.97% to S&P500 - see Table 1.
Historically (after the last 6 quarterly earnings releases) Nvidia has always beaten S&P500 - counting to the end of a given quarter, i.e. before the publication of the next earnings - see Figure 2. Table 1 (last column) shows higher rates of return compared to S&P500 from 4% to 47% (on relative basis).
Key takeaways after the Nvidia’s earnings call:
1) It’s all about growth, growth, growth!
Colette M. Kress : “Hopper demand is strong, and Blackwell is widely sampling. (…) Blackwell production ramp is scheduled to begin in the fourth quarter and continue into fiscal year '26. In Q4, we expect to get several billion dollars in Blackwell revenue. Hopper shipments are expected to increase in the second half of fiscal 2025. Hopper supply and availability have improved. Demand for Blackwell platforms is well above supply, and we expect this to continue into next year”.
“NVIDIA H200 platform began ramping in Q2, shipping to large CSPs, consumer Internet, and enterprise company. The NVIDIA H200 builds upon the strength of our Hopper architecture and offering over 40% more memory bandwidth compared to the H100”.
Currently, Nvidia sells chips on the Hooper platform (H100 and H200), the next platform will be Blackwell, and the one after that Rubin.
2) The next AI demand is "Sovereign AI" - nothing new, but Nvidia has given numbers for the first time …
„Our sovereign AI opportunities continue to expand as countries recognize AI expertise and infrastructure at national imperatives for their society and industries. Japan's National Institute of Advanced Industrial Science and Technology is building its AI Bridging Cloud Infrastructure 3.0 supercomputer with NVIDIA. We believe sovereign AI revenue will reach low double-digit billions this year.”
3) According to Nvidia… generative AI is accelerating: more scale and compute; image generation; coding; general robotics; physical AI; synthetic data generation..
“And there are several things that are happening in generative AI. So, the first thing that's happening is the frontier models are growing in quite substantial scale. (…) And so, it's not unexpected to see that the next-generation models could take 10, 20, 40 times more compute than last generation. (…) The second is although it's below the tip of the iceberg, what we see are ChatGPT image generators. We see coding. We use generative AI for coding quite extensively here at NVIDIA now.
And then general robotics. The big transformation last year as we are able to now learn physical AI from watching video and human demonstration and synthetic data generation from reinforcement learning from systems like Omniverse, we are now able to work with just about every robotics companies now to start thinking about, start building general robotics. And so, you can see that there are just so many different directions that generative AI is going.
And so, we're actually seeing the momentum of generative AI accelerating”.
4) the world builds about $1 trillion worth of data centers NOW !
Jensen Huang:
“(…) remember, the world is moving from general-purpose computing to accelerated computing. And the world builds about $1 trillion worth of data centers.
$1 trillion worth of data centers in a few years will be all accelerated computing. In the past, no GPUs are in data centers, just CPUs. In the future, every single data center will have GPUs”.
5) Blackwell is a game changer
“Blackwell is going to be a complete game changer for the industry. And Blackwell is going to carry into the following year. (…) remember that computing is going through two platform transitions at the same time. (…) which is general-purpose computing is shifting to accelerated computing, and human-engineered software is going to transition to generative AI or artificial intelligence-learned software”.
“The Blackwell vision took nearly five years and seven one-of-a-kind chips to realize, the Gray CPU, the Blackwell dual GPU, and a colos package, ConnectX DPU for East-West traffic, BlueField DPU for North-South and storage traffic, NVLink switch for all-to-all GPU communications, and Quantum and Spectrum-X for both InfiniBand and Ethernet can support the massive traffic of AI”.
6) NVLink (5th generation) is a game changer
Nvidia’s webpage: “Unlocking the full potential of exascale computing and trillion-parameter AI models hinges on swift, seamless communication between every GPU within a server cluster. The fifth generation of NVIDIA® NVLink® is a scale–up interconnect that unleashes accelerated performance for trillion- and multi-trillion parameter AI models”.
Jensen Huang:
“This is a very big deal with its all-to-all GPU switch is game-changing.
The Blackwell system lets us connect 144 GPUs in 72 GB200 packages into one NVLink domain, with an aggregate NVLink bandwidth of 259 terabytes per second in one rack. Just to put that in perspective, that's about 10x higher than Hopper. 259 terabytes per second kind of makes sense because you need to boost the training of multitrillion-parameter models on trillions of tokens”.
コメント