This episode is now available to stream on-demand using the form below.

Over recent years, there has been increasing competition in the AI chips market, with the number of companies developing GPUs increasing. Economies of scale that have traditionally favored CPUs can now be applied to GPUs, due to the rising demand for specialized applications such as AI. New AI chips are dramatically more effective for training than even state-of-the-art central processing units, and a recent study has estimated that by 2025, AI-related semiconductors are expected to account for almost 20% of all demand.

In this episode, expert speakers will consider how the data center industry can prepare for the wider-spread implementation of AI in their systems, and how the increasing number of AI chip vendors will drive innovation in the market. How specialized are chips going to become, and how can the industry move beyond conventional ideas of computing?