China is winning one AI race, the US another – but either might pull ahead
China is winning one AI race, the US another – but either might pull ahead
The Shift from Nuclear Arms to AI Competition
During the latter half of the 20th Century, the global focus centered on the development of nuclear capabilities, with the U.S. and Soviet Union vying for supremacy. Today, the competition has evolved into a technological showdown, with the U.S. and China locked in a battle for AI dominance. This contest spans research facilities, academic institutions, and innovative startups, closely monitored by corporate leaders and government officials. The financial stakes are immense, measured in trillions of dollars. Nick Wright, a cognitive neuroscience researcher at University College London (UCL), frames the rivalry as a clash between “brains” and “bodies” in AI.
The Rise of Large Language Models
In November 2022, OpenAI, a California-based tech company, unveiled its new chatbot, ChatGPT. With a concise six-sentence announcement, the firm highlighted its ability to engage in conversational interactions. Bloomberg’s Parmy Olson, author of *Supremacy: AI, ChatGPT, and the race that will change the world*, notes the immediate global reaction: “You could go on any sort of social network and there was just this flood of posts from people talking about all the different ways that they were using this new little text box that had appeared on the internet.” This marked the dawn of the first widely adopted large language model (LLM), which processes vast internet data to identify patterns in human communication.
Experts now largely agree that the U.S. holds an edge in AI “brains”—the realm of chatbots, microchips, and large language models. OpenAI claims ChatGPT is used by over 900 million people weekly, representing nearly 12% of the global population. Competitors like Anthropic, Google, and Perplexity have invested billions to match this progress, driven by the potential for LLMs to automate white-collar tasks and generate substantial profits.
Strategic Hardware and Export Controls
Yet, U.S. strategists emphasize that the race extends beyond software. A senior official told the BBC that the country’s advantage hinges on the hardware enabling advanced computing: microchips. Most high-performance chips, critical for powering Silicon Valley’s AI systems, are controlled by American entities. Nvidia, a California-based firm, designs the majority of these chips and became the first company to surpass a $5 trillion valuation in October. Stephen Witt, author of *The Thinking Machine*, highlights Nvidia’s significance in this context.
Washington employs stringent export controls to limit China’s access to these chips. While the policy originated in the 1950s to restrict Soviet allies, it was intensified in 2022 under President Joe Biden. Even though many chips are produced in Taiwan, a U.S. ally, the country ensures few reach China through its “foreign direct product rule.” This regulation compels foreign firms to comply with American standards if their products incorporate U.S.-made components or are based on U.S. technology.
Taiwan Semiconductor Manufacturing Corporation operates just across the strait from mainland China. Its proximity makes the island a strategic target for Beijing. However, replicating the advanced microchips requires specialized knowledge and infrastructure, which China has yet to fully achieve. The balance of power in this AI contest may shift further as both nations push their respective strengths to the limit.
