As leaders from the United States and China prepare for a bilateral summit this week, a Bank of America analyst has emphasized that neither country can achieve artificial intelligence (AI) supremacy independently due to the interconnected nature of global technology supply chains.
Matty Zhao, co-head of China equity at Bank of America, highlighted the contrasting strengths and vulnerabilities facing both nations in the race for AI dominance. Speaking on the sidelines of the Bank of America 2026 Asia-Pacific Conference in Hong Kong, Zhao noted China’s advantages in sectors such as power grids and manufacturing equipment. However, she pointed out that China remains reliant on foreign suppliers for advanced semiconductors, memory chips, and certain high-end materials necessary for AI development.
Conversely, Zhao observed that the United States grapples with an aging power infrastructure, with its electrical grid averaging 30 to 40 years old, compared to China’s estimated 15 to 20 years. She said data centers in the US have encountered political and community opposition driven by concerns over escalating energy costs. While power constraints may limit US growth in AI, China still requires “breakthroughs” in high-end chip technology for further progress. “Various countries have various limitations,” Zhao said, underscoring the complexity of the AI ecosystem.
Zhao stressed the need for continued cooperation amid competition, citing South Korea’s expertise in memory chip production and Japan’s leadership in advanced materials such as copper foil and manufacturing equipment essential for printed circuit boards. “This race cannot be done by one country,” she said.
The comments come ahead of a summit in Washington where US President Donald Trump and Chinese President Xi Jinping are expected to discuss AI alongside trade and geopolitical issues. Recently, US Treasury Secretary Scott Bessent and Chinese Vice-Premier He Lifeng agreed to establish a formal dialogue on AI and create a bilateral trade body following extensive talks in New York. These initiatives aim to support ongoing engagement between the two powers and set a positive tone for the summit.
Despite these diplomatic efforts, tensions remain high. The US has repeatedly accused Chinese AI developers of replicating capabilities from American models, while China has criticized calls from the US to slow frontier AI development, viewing them as measures to maintain US dominance. Both countries have simultaneously sought to reduce technological dependence on each other; for instance, US officials have advocated for tighter export controls on advanced chips, while China encourages domestic firms to develop indigenous semiconductor technology.
Yet the fundamental infrastructure for AI—ranging from energy systems to specialized components—remains deeply globalized. China has emerged as a significant supplier of AI-related electrical infrastructure, including power equipment and energy storage systems. Its exports of electrical transformers, crucial for regulating voltage in power grids, increased by about 30 percent last year and have continued growing at a similar rate in 2026.
Nevertheless, China’s dependence on foreign technology persists, particularly in areas such as high-end Japanese manufacturing equipment and advanced chips and optical components from the US and other countries. This interconnectedness, Zhao concluded, means that “the AI race cannot be won alone” and underscores the intertwined challenges and opportunities facing the world’s two largest AI rivals.