Intel Partners with Top Chinese Tech Firms to Explore Hardware Solutions for AI Language Model ChatGPT.

The emergence of ChatGPT has not only propelled the development of AI technology to new heights but also brought changes and impacts that major technology companies cannot ignore. It is essential to recognize the implications ChatGPT may have on chip computing power and data centers. ChatGPT's training method, which involves interactive and reinforcement learning, has significantly increased the complexity of its training and doubled the amount of data, resulting in rapid growth in model size and training costs. OpenAI has estimated that computing power demand will double every 3.5 months and nearly 10 times a year, which will pose new challenges to many hardware systems. The emergence of ChatGPT presents an opportunity for GPU companies such as NVIDIA, which have made significant breakthroughs in the application of AI technology in various industries over the past six or seven years. However, as a new technology, ChatGPT is still in its early stages of development, and it will require exploration of suitable acceleration solutions among various types of hardware, including CPU, GPU, FPGA, and ASIC. Intel, a chip supplier, has a wide range of products, including CPUs, GPUs, FPGAs, and deep expertise in high-performance acceleration in data centers, providing a foundation for developing hardware solutions for AI applications such as ChatGPT. Intel has already engaged in extensive cooperation with clients, including Baidu and Alibaba, to explore the deployment of ChatGPT and similar AI technologies in enterprises and vertical fields.