Su also touched on the use of AI within AMD, stating that it is being used more than people might think, but clarified that AI is not going to be designing chips all by itself anytime soon. She also discussed the potential for regulation in the AI industry, stating that while there is no silver bullet, a combination of chip, software, and model development could provide safeguards. Su also mentioned the importance of geographical diversity in manufacturing and the potential for custom chips in the future.
Key takeaways:
- AMD CEO Lisa Su discussed the company's focus on AI and the chip supply chain, noting that the balance of supply and demand is in a good place, except for high-end GPUs powering large AI models.
- AMD is working on a new chip called the MI300, which is expected to compete with Nvidia’s H100 chip. The MI300 is designed for large language model training and inference.
- Su highlighted the importance of geographical diversity in chip manufacturing and expressed support for the CHIPS and Science Act, which aims to build chip manufacturing capacity in the United States.
- She also discussed the use of AI within AMD, stating that it saves time in designing, testing, and validating chips, and emphasized the need for safeguards in AI development and deployment.