The H100 GPU is considered the “new gold” in the technology industry because it is the preferred choice for providing computing power for generative artificial intelligence.
Which companies did NVIDIA’s best-selling H100 GPUs go to this year? The following chart clearly lists NVIDIA’s largest customers this year.
According to the latest report from market research firm Omdia Research, Meta and Microsoft are tied for first place with a purchase volume of 150,000 units of H100 GPUs.
Microsoft’s large order quantity is not surprising as the company needs computing power for its growing artificial intelligence products and its existing AI assistant Copilot (formerly Bing Chat).
Meta also has ambitions in the field of artificial intelligence. There have been reports in recent months that Meta continues to purchase AI chips and create data centers with the aim of developing a chatbot as powerful as ChatGPT. It is reported that Meta is developing a new large-scale language model that is expected to be several times more powerful than Llama 2, which was launched in July this year, and aims to start training new models next year.
However, there is a slight disparity in purchasing volume between other companies and Meta/Microsoft.
Google, Amazon, Oracle, and Tencent jointly rank third with a purchase volume of 50,000 units of H100 GPUs; cloud service provider CoreWeave estimates a purchase volume of 40,000 units; Baidu and Alibaba purchased 30,000 and 25,000 units respectively; ByteDance and cloud service provider Lambda Labs follow closely behind with a purchase volume of 20,000 units each; Tesla purchased 15,000 units.
Most went to cloud providers
Omdia Research’s report also shows that NVIDIA supplied most GPU servers to large cloud service providers. This may also indicate that NVIDIA is moving closer towards cloud services.
It should be noted that among NVIDIA’s top twelve clients for H100 GPUs only CoreWeave and Lambda Labs are start-up companies, but they have close ties with NVIDIA.
As early as the launch of the H100 chip this year, NVIDIA chose CoreWeave and Lambda as the first companies to use the chip. Moreover, NVIDIA’s investment can be seen in their financing processes: in April this year, NVIDIA participated in CoreWeave’s $221 million Series B financing; in July this year, NVIDIA invested $300 million in Lambda Labs.
Currently, CoreWeave is valued at approximately $2 billion and Lambda’s valuation will also exceed $1 billion.
Analysts previously pointed out that by selling its GPUs to these cloud start-ups, NVIDIA aims to expand its target customer base beyond just Amazon, Microsoft, Google and other cloud giants.
Future prospects
In addition, recent data released by Omdia shows that chip manufacturing giant NVIDIA shipped nearly 500 thousand units of H100 and A100 GPUs in the third quarter. The previous financial report showed that revenue from NVIDIA’s data center segment for this quarter was $14.5 billion, almost four times that of the same period last year.
The analysis firm also predicts that by the fourth quarter of 2023, sales of H100 and A100 GPUs from NVIDIA will exceed 500 thousand units.
However, almost all companies purchasing large quantities of NVIDIA’s H100 GPUs are developing custom chips for AI, HPC (high-performance computing), and video workloads. Therefore, over time as customers turn to their own chips instead of relying on NVIDIA hardware purchases may decrease.