China is winning one AI race, the US another – but either might pull ahead

During the latter half of the 20th century, the race to build nuclear weapons consumed the intellectual efforts of the US and the Soviet Union. Today, the focus has shifted to a different kind of competition, one centered on technological supremacy. The US now faces a new rival: China. The stakes are no less intense, but the battlefield has evolved, with the goal of mastering artificial intelligence becoming central to this contest.

This rivalry spans research facilities, academic institutions, and the boardrooms of tech startups. Global leaders from Fortune 500 companies and government officials are closely monitoring the progress. The financial investment is colossal, measured in trillions of dollars. Each side claims a distinct advantage, as noted by Nick Wright, a cognitive neuroscience researcher at University College London (UCL). He describes the contest as a struggle between “brains” and “bodies” in the realm of AI.

The US has historically excelled in the development of AI “brains”—chatbots, microchips, and large language models (LLMs). China, meanwhile, has demonstrated strength in AI “bodies,” particularly in the creation of humanoid robots that resemble humans. However, with both nations striving to outpace the other, these lead positions may not endure indefinitely, and the contest could take new directions in the near future.

In late November 2022, OpenAI unveiled its latest chatbot, ChatGPT, in a brief six-sentence message. The tool was designed to engage in conversational interactions, sparking immediate global fascination. Bloomberg’s Parmy Olson, author of Supremacy: AI, ChatGPT, and the race that will change the world, recalls the tech world’s reaction: “You could go on any social platform and see a surge of posts discussing the various ways people were using this new text box online.” This marked the debut of a mainstream large language model, capable of processing and learning from vast amounts of internet data.

Experts now largely concur that the US holds an edge in AI “brains.” OpenAI reports over 900 million weekly users of ChatGPT, nearly one in eight of Earth’s population. Competitors like Anthropic, Google, and Perplexity quickly followed, pouring billions into developing their own LLMs. These firms recognize that success in this area could automate tasks in white-collar sectors, translating into substantial profits.

Yet, Washington’s priorities extend beyond algorithmic innovation. A senior US official told the BBC that the country’s strategic edge lies in its control of high-performance microchips, crucial for powering LLMs. Most of these advanced chips, used by Silicon Valley firms, are American-designed, with Nvidia being a dominant player. In October 2022, Nvidia surpassed a valuation of $5 trillion, potentially making it the most valuable company in history, according to Stephen Witt, author of The Thinking Machine.

The US leverages strict export regulations to limit China’s access to these microchips. This policy traces back to the 1950s, when the US restricted electronic exports to Soviet-aligned nations. It was intensified under President Joe Biden in 2022 as the AI race intensified. Despite most of these chips being manufactured in Taiwan—a US ally—by TSMC, the US ensures few reach China through its “foreign direct product rule.” This law compels foreign firms to comply with US standards if their products incorporate American technology or are based on US innovations.

TSMC’s factory is nearly visible from mainland China, highlighting the island’s strategic importance. Why haven’t Chinese manufacturers developed their own high-end chips? The answer lies in the complexity of the process, which requires advanced engineering and infrastructure. This challenge underscores the ongoing battle for AI leadership, where both nations are vying for supremacy in different domains.