NVIDIA releases new generation AI chip H200, with up to 90% performance improvement.

 NVIDIA announced the launch of its new generation AI processor, the H200, tonight Beijing time. The H200 is a graphics processing unit (GPU) designed for training and deploying a variety of AI models.

The new H200 GPU is an upgrade to the current H100, integrating 141GB of memory. The performance of the H200 improves by 60% to 90% compared to the H100 when used for inference or generating answers to problems.

Systems equipped with the H200 are expected to be available in the second quarter of 2024 through NVIDIA's hardware partners and major cloud service providers, including Amazon Web Services, Google Cloud, and Microsoft Azure.

Ian Buck, NVIDIA's vice president of large-scale and high-performance computing, said that as the industry-leading end-to-end AI supercomputing platform, the H200 will help solve some of the world's most pressing challenges faster.

NVIDIA also stated that the H200 will be compatible with the H100, meaning that AI companies that have already trained on the previous generation of processors will not need to replace their server systems or software to use the new product.

Fueled by generative AI, NVIDIA's stock has risen by more than 230% so far this year. NVIDIA expects its revenue for the third quarter to reach approximately $16 billion, marking a year-on-year increase of about 170%.

Due to strong demand for GPUs, NVIDIA also announced last month that it will change its two-year architecture cycle to a one-year release model. Next year, NVIDIA will also release the B100 chip, which is based on the Blackwell architecture.