NVIDIA's major customer Microsoft: "AI chip shortage" is improving
On Wednesday, Kevin Scott, the Chief Technology Officer of Microsoft, stated that the company has found it easier to obtain artificial intelligence chips produced by NVIDIA compared to a few months ago. Although the supply-demand imbalance remains tight, the situation has been improving on a weekly basis.
On Wednesday, September 27th, Kevin Scott, the Chief Technology Officer of Microsoft, stated that the company is finding it easier to obtain artificial intelligence chips produced by NVIDIA compared to a few months ago.
Scott mentioned that the supply of NVIDIA's graphics processing units (GPUs) is gradually increasing. Since the launch of ChatGPT, a chatbot developed by OpenAI and supported by Microsoft, there has been a significant demand for GPUs.
Scott said:
Currently, the demand for GPUs in the market still far exceeds the supply. The supply-demand imbalance remains tight, but every week brings more good news than bad news.
Following the success of ChatGPT, a language model developed by OpenAI, tech giants such as Google, Amazon, and Microsoft have increased their investment in the development and application of large-scale language models and AI chatbots. This has led to a sharp increase in demand for high-performance GPUs designed for data centers, such as NVIDIA's H100 and A100. Multiple vendors have been aggressively purchasing these GPUs, resulting in a supply shortage. Microsoft, Google, and other tech companies are rapidly integrating generative AI capabilities into their products and selling features with this technology to customers.
NVIDIA stated last month that it expects its revenue to increase by 170% compared to the same quarter last year. The company's control over the AI chip market has boosted its gross margin from 44% to 70% within a year. NVIDIA's stock price has surged by over 200% this year, far surpassing all other listed companies in the S&P 500 index.
In May of this year, Scott mentioned that one of his responsibilities was to control Microsoft's budget for GPUs. He described it as "a terrible job" and "painful for the past five years." However, he stated on Wednesday:
The situation has improved compared to before. At the beginning of the year, generative AI was still relatively new to the market and received widespread attention from the public.
He mentioned that the increase in GPU supply has made his job less complicated.
During last month's earnings conference call, Colette Kress, the Chief Financial Officer of NVIDIA, stated that the company expects to increase its supply every quarter until next year.
Like many Silicon Valley giants, Microsoft is actively developing its own AI chips to reduce the cost of AI development and operation. However, Scott did not provide any specific comments on this matter. He said:
I don't have any information to confirm, but I can say that we have made significant investments in this field for many years. What we need to do is ensure that we make the best choices for how to build these systems using any of our existing options. And in recent years, the best choice has been NVIDIA's GPUs.