Recently, at the China-Europe Fund 2026 Investment Strategy Meeting, China-Europe Fund Manager Du Houliang predicted that demand for computing power in domestic and foreign markets during training and reasoning will rise sharply in 2026. He pointed out that the current core bottleneck is on the supply side. Currently, domestic chip streaming still relies on overseas foundries. It is expected that the mismatch between supply and demand may become apparent in 2026, and high-quality domestic foundries will need to accelerate technological breakthroughs and capacity building. Du Houliang said that the potential for AI applications, new software, and physical world applications in 2026 may become a “gray rhinoceros.” He believes that currently, technology companies' resources are concentrated in the core field of AI, and non-core businesses are shrinking, and the supply chain may be squeezed early next year, driving up upstream cost pressure. He judged that terminal and application innovation is likely to “suppress first and then improve”. After the base model matures, the empowerment of scenarios such as robotics and smart driving will accelerate, and the application layer is expected to explode.

Recently, at the China-Europe Fund 2026 Investment Strategy Meeting, China-Europe Fund Manager Du Houliang predicted that demand for computing power in domestic and foreign markets during training and reasoning will rise sharply in 2026. He pointed out that the current core bottleneck is on the supply side. Currently, domestic chip streaming still relies on overseas foundries. It is expected that the mismatch between supply and demand may become apparent in 2026, and high-quality domestic foundries will need to accelerate technological breakthroughs and capacity building. Du Houliang said that the potential for AI applications, new software, and physical world applications in 2026 may become a “gray rhinoceros.” He believes that currently, technology companies' resources are concentrated in the core field of AI, and non-core businesses are shrinking, and the supply chain may be squeezed early next year, driving up upstream cost pressure. He judged that terminal and application innovation is likely to “suppress first and then improve”. After the base model matures, the empowerment of scenarios such as robotics and smart driving will accelerate, and the application layer is expected to explode.