Computing pioneer David Patterson discusses the AI boom, the memory chip shortage and the future of chip design | Ian Cutress fireside - GamesBeat

At the AI Infrastructure Summit in September 2026, computing pioneer David Patterson engaged in a deep conversation with semiconductor analyst Ian Cutress. The discussion covered the impact of the AI ... Read original →

The AI boom is reshaping chip design paradigms, with memory bandwidth emerging as the primary constraint. Patterson's insights highlight that, post-Moore's Law, the industry is shifting toward efficiency and bandwidth optimization, pressuring leaders like TSMC and NVIDIA. The ongoing DRAM shortage is inflating data center costs, pushing firms toward high-bandwidth memory and optical interconnects. Evolving IEEE floating-point standards are influencing scientific computing performance, reshaping global compute competitiveness. From a policy standpoint, tech decoupling between China and the US intensifies supply chain risks, especially in critical materials and advanced processes. Market dynamics see Google potentially accelerating in-house chip development to reduce reliance, while TSMC must deepen vertical integration with AI firms. Over the next 12–24 months, breakthroughs in storage technology will define whether AI infrastructure can support large-scale model training, creating a new industry divide.