Chip Constraints Slowing AI and Robotics, Says Hassabis

- The strain on the semiconductor supply chain is not limited to GPUs; there's a significant shortage of high-bandwidth memory (HBM) and even CPUs, as advanced manufacturing capacity is being prioritized for higher-margin AI accelerators. This directly impacts the development of agentic AI and complex robotic systems which require a mix of processing capabilities. - To counter foreign supply chain vulnerabilities, the U.S. government is investing heavily in domestic semiconductor production through the CHIPS and Science Act, which allocates $52.7 billion to boost manufacturing and research. This includes $2 billion specifically for the Department of Defense to fund microelectronics research and workforce training. - The competition for chips is creating a structural shift, with the AI data center market projected to consume over 50% of the entire chip market by 2030, deprioritizing other sectors like automotive. This trend could impact robotics companies that rely on foundational chips similar to those used in vehicles. - The defense industry faces significant risks from semiconductor obsolescence and supply chain disruptions, which can ground military equipment and delay deployments. The U.S. military's reliance on foreign-fabricated chips, particularly from Taiwan for FPGAs used in systems like the F-35 fighter jet, is a key national security concern. - Venture capital is flowing into AI chip startups to address the supply-demand imbalance, with U.S. semiconductor startup funding reaching a record $6.2 billion in 2025. Notable examples from early 2026 include Ricursive Intelligence, founded by creators of Google's AlphaChip, which raised a $300 million Series A at a $4 billion valuation. - While chip constraints are a bottleneck, the development of humanoid robots is also dependent on advancements in actuators, sensors, and AI algorithms for manipulation and interaction. To address labor shortages and increase efficiency in their own industry, some semiconductor manufacturers like STMicroelectronics are beginning to deploy humanoid robots in their fabrication plants. - Geopolitical tensions are a major factor, with U.S. export controls restricting China's access to advanced AI chips from companies like NVIDIA and AMD. This has prompted Chinese firms like Huawei to accelerate their own chip development, though they are still considered to be behind the most advanced U.S. offerings. - The lead time for building new semiconductor foundries means that significant new capacity from investments by companies like TSMC and Intel in the U.S. is not expected to come online until 2027-2028, suggesting the current constraints will persist. TSMC has already informed key customers like Nvidia and Broadcom that it cannot meet all of their demand for advanced AI processors.