What is at the heart of Nvidia’s earnings achievement?
Graphics and AI juggernaut Nvidia continues to exceed expectations, while setting loftier guidance for the coming quarter. With the earnings for the 2nd quarter of its 2024 fiscal year, which corresponds to the three-month period ending July 30, 2023, the company has lived up to its own hype and then some. Revenue guidance for the quarter was $11 billion, which Nvidia exceeded by $2.5 billion, giving the company just more than twice the quarterly revenue than the same period last year. The company guided $16 billion revenue for the next quarter, which if achieved, would represent a 170% increase from the same period last year.
Furthermore, Yole Group will release soon a detailed analysis of Nvidia H100 Tensor Core GPU (Coming soon – More info.). Make sure to get this latest analysis and stay tuned on www.yolegroup.com.
What is the reason for this upswell of business?
It comes from the explosion of demand for cloud-based AI acceleration, especially the kinds of generative AI which use large-language-models or some other transformer model for real-time content generation. Among high profile examples like Open. AI’s ChatGPT or Dall-E, there are thousands of companies examining how generative AI can improve, or even define, their business model. Just one such model might require thousands of accelerators running for weeks or months to train, and because the model is so large, the user-experience (aka inference) is also managed by AI accelerators. These AI accelerators, for training and inference, are the solutions that Nvidia is bringing to the market, made more flexible to the customers’ needs by their CUDA software platform.
Q2, 2024 results: so, what…
One conclusion from the recent quarter’s results and next quarter guidance, is that Nvidia is not significantly constrained in their supply chain. Relative to application processors and CPUs for PC, data center GPUs are not very hungry for foundry capacity. According to Yole Group’s Quarterly Processor Market Monitor, GPUs accounted for less than 11% of foundry and IDM processor wafer volume in 2022. However, Nvidia’s AI hardware carries additional complexities in packaging and memory integration, as they use TSMC’s CoWoS to interconnect the GPU die with stacks of high-bandwidth memory from Samsung, SK Hynix, or Micron. While none have appeared yet, the wafer production, advanced packaging, or HBM could be sources of future bottlenecks.
Is anything posing a legitimate threat to Nvidia’s success? In the immediate future, very little. However, Nvidia customers might not pay 80%+ margins in perpetuity, and competitors AMD and Intel have excellent solutions of their own.
John Lorenz
Senior Technology & Market Analyst, Computing & Software
Nvidia’s differentiation of a well-developed AI processing ecosystem might give way to similar efforts from the competition at the right price. We have already seen Nvidia sidestep a geopolitical risk when their A100 and H100 GPUs were barred from sale to the Chinese market, to which the company responded with scaled-down versions, resulting in even higher unit demand. At the same time, this increases the pressure for a home-grown AI acceleration supplier to rise in China.
Altogether, Nvidia is essentially the only boat riding on a majestic wave, supplying the exponentially growing generative AI demand. The company positioned itself well, building on their solutions for accelerating graphics. As their success has grown, so does the potential prize for the competition. As the company’s market capitalization nears the $1.2 trillion threshold, the whole industry will look for any cracks in the hull – for now, we see none. Stay in touch with us as we are updating quarterly our analysis, both at market side and supply chain side.
About the author
John Lorenz is a Senior Technology and Market Analyst, Computing & Software within the Semiconductor, Memory & Computing division at Yole Intelligence, part of Yole Group.
John is engaged in the development of market and technology monitors for the logic segment of advanced semiconductors, with a primary focus on processors. Prior to joining Yole, John held various engineering and strategic planning roles over 15 years at Micron Technology.
John has a Bachelor of Science degree in Mechanical Engineering from the University of Illinois Urbana-Champaign (USA), with a focus on MEMS devices.
This article has been developed in collaboration with Emilie Jolivet, Director, Semiconductor, Memory & Computing division at Yole Intelligence, part of Yole Group.
John Lorenz Senior Technology & Market Analyst, Computing & SoftwareNvidia’s differentiation of a well-developed AI processing ecosystem might give way to similar efforts from the competition at the right price. We have already seen Nvidia sidestep a geopolitical risk when their A100 and H100 GPUs were barred from sale to the Chinese market, to which the company responded with scaled-down versions, resulting in even higher unit demand. At the same time, this increases the pressure for a home-grown AI acceleration supplier to rise in China.
