While no one has waved an official checkered flag in the Sino-American race for artificial intelligence ( AI ) supremacy, the markets are betting that the United States will prevail. The chipmaker Nvidia recently became the world’s first US$4 trillion company ( and its CEO, Jensen Huang, has acquired global rock star status ). Microsoft, the biggest investor in OpenAI’s for-profit entity, is not far behind, with a valuation of US$3.7 trillion.

But early leadership does not guarantee victory, especially when it comes to innovation. Hardly a day goes by without a new report about China’s extraordinary AI gains. The US may have broken new ground with OpenAI’s ChatGPT, but China’s DeepSeek shocked the world earlier this year with the cost and processing efficiency of its R1 large language model. And just this month, the Chinese start-up Moonshot AI released its impressive Kimi K2 model, which outperforms Western competitors on several key benchmarks.

Many factors influence the AI race – not only Nvidia’s powerful chips, but also talent, software and strategic focus. For now, semiconductors are an obvious strategic chokepoint working to America’s advantage. Under its “small yard, high fence” policy, the Biden administration imposed stringent restrictions on advanced semiconductor exports. Yet this has backfired, encouraging China’s aggressive pursuit to develop its own AI chips.

In the end, I suspect that the AI race will be determined less by hardware than by strategic breakthroughs in software. Notwithstanding US President Donald Trump’s newly announced AI Action Plan, China is well-positioned for the long haul. The Global Innovation Index 2024 ( GII ), which gauges the innovation performance of 133 countries on 78 separate indicators, ranked China 11th – a dramatic rise from 15 years ago, when it was 43rd. Meanwhile, the US has remained around third place.

While the GII framework provides a comprehensive overview of the ebb and flow of innovation around the world, it misses one key piece of the puzzle: basic theoretical research. Government stewardship plays a crucial role here. Unlike private actors, who are motivated by commercial returns, public support gives scientists and other researchers the leeway to push the seemingly abstract frontiers of knowledge.

On this measure, the US has fallen dangerously short. Based on official statistics from the National Science Foundation ( NSF ), the federal-government share of total US spending on research and development has been on a downward trend since the post-Sputnik peak of 1964. For basic research, in particular, the federal government’s share of total spending fell from just under 30% in the late 1970s to around 10% in 2023.

Even more disconcerting is the Trump administration’s assault on scientific research and higher education ( ostensibly to abolish diversity, equity and inclusion programmes ), as well as the anti-collaborative mindset fostered by America’s increasingly worrisome Sinophobia. According to a detailed research and development ( R&D ) assessment recently published by the American Association for the Advancement of Science, Trump’s budget proposal for fiscal year 2026 will likely slash federal funding for basic research to just US$30 billion, a 34% drop from the US$45 billion projected for fiscal year 2025. Based on NSF metrics, that would mark a return to levels last seen in 2002.

By contrast, China has poured money into advancing its ambitious science and technology agenda, accounting for 28% of global R&D investment in 2023 – only slightly behind the US, which accounted for 29%. With Chinese R&D spending increasing at an average annual rate of nearly 14% over the past 10 years, more than three and a half times the 3.7% pace of the US, convergence most likely occurred in 2024.

While comparable country figures for basic research are not readily available, Jimmy Goodrich, a non-resident expert at the Center for Strategic and International Studies, has taken a stab at calculating them. His extrapolation of trend growth in Chinese R&D produces the staggering conclusion that the Trump administration is in the process of ceding America’s long-standing lead in government-supported basic research.

Why? The same question could be raised about many of the policy reversals of Trump 2.0, from tariffs to foreign-aid cuts and rollbacks of clean energy initiatives. Most of these actions were outlined in the Heritage Foundation’s Project 2025, the conservative agenda for Trump’s second term. But one of the blueprint’s main objectives was supposedly to “champion, engage and focus the American innovation ecosystem”. The subsequent gutting of basic research is anything but that. On the contrary, it borders on economic and competitive suicide.

Chinese President Xi Jinping is of precisely the opposite mind. Continuing his predecessor’s focus on “scientific development”, Xi has long stressed the importance of basic research as a pillar of Chinese innovation. In early 2023, he argued that “strengthening basic research is an urgent requirement for achieving greater self-reliance and strength in science and technology, and it is the only way to become a world leader in science and technology.”

Today’s global battle for AI supremacy is often positioned as a conflict between two systems: America’s market-driven model and China’s state-supported industrial policy. But basic research is the great leveller. Irrespective of whether the public or the private sector drives the system, innovation ultimately flows from discovery.

“Genesis: Artificial Intelligence, Hope and the Human Spirit”, a book by the late Henry Kissinger, Craig Mundie and Eric Schmidt, asserts that “discovery may be the single most exhilarating capacity of the human species.” Maintaining a culture of discovery requires support for basic research which is not only abstract and theoretical but also casts a wide net. As the inventors of gunpowder and paper, the Chinese have long taken this lesson to heart. Unfortunately, America may be about to re-learn it the hard way.

Stephen S. Roach is a faculty member at Yale University and a former chairman of Morgan Stanley Asia.

Copyright: Project Syndicate