Unlocking the Potential of ChatGPT
Bernstein suggests that the semiconductor industry is set to benefit significantly from the AI craze fuelled by ChatGPT, with a potential multi-billion dollar annual market opportunity arising as result of driving AI queries. Analyst Stacy Rasgon estimated that almost 400 quadrillion operations are needed to accomplish a typical-sized ChatGPT query response, suggesting a GPU Total Addressable Market of multiple tens of billions of dollars annually. Microsoft and Alphabet are likely to be the main beneficiaries, however Nvidia is also well-positioned to profit from this trend. Monitor developments closely to stay informed!
Nvidia, a well-known leader in the semiconductor industry, experienced a 9% surge in its shares last week due to a slight beat in its fourth-quarter results, attributing to increased demand for its AI-powered chips for gaming and machine learning. CEO Jensen Huang believes AI is at an “inflection point,” leading to a 61% YTD growth in its stock.
Bernstein estimates that 100 million ChatGPT queries daily would require annual purchases of $1 to $2 billion worth of Nvidia’s H100 PCIe chips, while Rasgon predicted a potential worth of $10 to $20 billion in purchases to handle the large language model loads. This demonstrates the potential of the AI industry, with Nvidia’s Ampere (A100 generation chips) and Hopper (H100 generation chips) being highlighted as the preferred offerings.
Conjunct adverbially, it seems that the estimates of annual H100 graphics processing unit sales of worth $150 billion, when considering Google’s search queries of roughly 10 billion a day, are reasonable. As the adoption of large language models rises and the GPU market grows, their use in search and other applications is likely to be sustained. CNBC’s Michael Bloom reported that Adani also recognizes the potential and dangers of generative AI, similar to the advent of chips which enabled the development of modern warfare weapons.