Artificial Intelligence is not stopping any time and it is very well-known that AI runs on high-performance professional GPUs.
According to the latest report by DigiTimes, the demand for Nvidia GPUs has increased drastically since the usage of AI technologies like ChatGPT and other such AI models.
DigiTimes says that Nvidia has seen a ramp-up in orders for its A100 and H100 AI GPUs, leading to an increase in wafer starts at TSMC. Unlike video games, AI models tend to rely on high-performance computing graphics cards that can deliver significantly faster computing performance with better accuracy and low power consumption.
Unlike AMD, Nvidia is currently the leading manufacturer of such GPUs like DGX A100 chip that is built on TSMC’s 7nm process node and features 80GB of HBM2 memory. Similarly, the DGX H100 GPU is particularly more in demand as it uses the refined 5nm process node and features significantly higher performance than the A100.
The demand for AI GPUs is even higher in China where the company is selling modified versions of these GPUs to bypass the US sanctions and Nvidia is easily able to charge 37.5% higher prices there according to ITHome.
It is reported that Nvidia is not able to keep up with the demand for AI GPUs but it is still trying to ramp up the GPU production as it has to fulfill the orders outside of China too.