【By Bear Chao Ran, Observer Group】The competition between China and the United States in the field of artificial intelligence (AI) is intensifying. At this time, American media suddenly changed its perspective, using sports as an entry point to observe the strength comparison in this rivalry.

On December 29 local time, the Wall Street Journal compared the US-China AI competition to a football game, interpreting the situation using touchdowns and field goal scores in this sport. After citing the views of several analysts, the report concluded that at halftime, the United States was temporarily leading China with a score of 24 to 18. However, what changes will occur in the second half?

The report then pointed out that entering the second half, although the United States still maintained the lead, China was currently on a strong upward trend, especially after President Trump made a so-called "risky deal," which gave China a new "quarterback."

July 27, 2025, Shanghai, China. Visual China

The Wall Street Journal explained in its report that this was not a real football match, there was no fourth-quarter timer, and both sides could claim victory. Moreover, the prize they were competing for was not a trophy but rather the leadership in economic and military fields.

To make this competition easier to understand, the Wall Street Journal introduced the concept of American football, an extremely popular sport among Americans, and invited professionals to interpret the AI competition using touchdowns and field goal scores in football.

Regarding the score at halftime, the report believes that the United States scored 24 points and China scored 18 points.

United States:

ChatGPT returned the kickoff for a touchdown, this chatbot opened the game with a decisive score. The United States earned 7 points.

NVIDIA's 77-yard touchdown run, this quarterback successfully connected Claude, Gemini, and Grok. The United States earned 14 points.

The defense created an easy field goal opportunity, the U.S. export controls forced the opponent to fumble and recover the ball. The United States earned 17 points.

NVIDIA's 22-yard rushing touchdown, relying on the quarterback carrying the ball into the end zone. The United States earned 24 points.

China:

DeepSeek's 80-yard reception touchdown, it caught a short pass, broke three tackles, and successfully reached the end zone. China earned 7 points.

Huawei's touchdown and two-point conversion, this quarterback steadily led the team forward toward the opponent's half. China earned 15 points.

A field goal narrowed the gap, Alibaba and ByteDance's receptions helped bring the score within one possession. China earned 18 points.

However, despite the way the media interpreted the situation, experts have different opinions on the closeness of the game.

For example, Chris Miller, the author of "Chip War," believes that the United States leads China by a large margin of 24 to 12. He argues that China has more power but is sending brainpower (talent) to the United States, while only U.S. companies can convert AI into profit, showing stronger monetization capabilities.

Meanwhile, Deepika Giri, head of AI research at IDC in Singapore, gave the most closely contested score, with the United States slightly leading China 21 to 19. She believes that while the United States dominates the advanced AI chip sector through companies like NVIDIA, China is rapidly rising through chatbots like DeepSeek and open-source innovation.

In other analyses, Tarun Chhabra, the director of national security policy at U.S. AI startup Anthropic, gave a score of 21 to 14. He believes that the United States leads in models, chips, global talent, and allies at the top of the supply chain, and claims that strong export controls are key to maintaining this advantage.

Saif Khan, a researcher at the U.S. Progressive Policy Institute, interprets the score as 24 to 17. He believes that the United States maintains its lead through its computing power advantages, but China still has a chance in this game with its strong algorithms, data, and energy resources.

Professor John Villasenor from UCLA believes the current score is 24 to 21, and if the United States does not continue to perform well, it may lose its slight lead. At the same time, as China accelerates its energy development, the United States must build clean energy faster.

Kevin Xu, founder of the U.S. hedge fund Interconnected Capital, believes the current score is 29 to 25. He points out that in the five key areas of AI, China leads in energy and infrastructure capabilities, while the United States leads in models, applications, chips, and computing power.

The Wall Street Journal pointed out that amid the intense and close competition, President Trump recently also completed a so-called "risky deal." This month, Trump said he would allow NVIDIA to sell an older but still powerful chip to China, thus relaxing previous restrictions preventing top U.S. technology from flowing to China.

Both political parties in the U.S. have exaggerated and portrayed this as Trump handing a "star player" to the competitor, while Trump supporters defended it as a strategic long-term move.

In this AI competition, the Wall Street Journal focuses on two key factors—chips and chatbots—and provides analysis on each.

Chips: Quarterback

To understand Trump's bet on chip exports, consider the 1993 San Francisco 49ers.

This team had the best quarterback at the time, Steve Young, and the legendary Joe Montana, who was replaced by him. The 49ers believed that Montana had little value sitting on the bench, so they traded him to the Kansas City Chiefs in exchange for a draft pick that could select a player to assist Young.

In this case, Montana, the veteran, refers to NVIDIA's H200 chip, which recently received export approval from the Trump administration. This chip was released in mid-2024, one generation behind the company's most advanced Blackwell chip.

Saif Khan analyzed that the cost-effectiveness and performance of the H200 chip are still higher than China's self-developed chips. His team stated that if the H200 chip is approved for sale to China, the U.S. computing power advantage will narrow to less than seven times.

However, U.S. national security "hawks" warn that an older "veteran" like Montana is still stronger than any quarterback China has developed.

But NVIDIA executives have a different view. A company spokesperson said that U.S. customers will continue to get more computing and AI capabilities than anyone else.

Corporate leaders worry that blocking sales will force China to build a parallel AI ecosystem around Chinese companies, which the U.S. cannot control. They argue that it is better for China to continue using NVIDIA chips so that if Chinese startups like DeepSeek make breakthroughs, the technology they rely on is something U.S. companies can easily replicate.

February 9, 2025, New Orleans, Louisiana, USA. Visual China

Chatbots: Wide Receiver

The Wall Street Journal described that if chips play the role of a quarterback, chatbots are the wide receivers who can turn a perfectly thrown spiral into a touchdown.

In fact, chatbots are already doing practical work, and both China and the United States have greater ambitions. U.S. developers say the ultimate goal is general AI, a system that may eventually surpass humans through independent thinking. The more top chatbots a country has, the more likely it is to make breakthroughs first.

The LMArena leaderboard for global chatbots shows that the United States is constantly scoring. Early this month, just four U.S. companies — Google, xAI, Anthropic, and OpenAI — had various models that once occupied the top 20 spots.

However, in the top 30 list, many Chinese tech giants, including Alibaba, Baidu, and the rising star DeepSeek, are also present.

Especially DeepSeek, which earlier this year used second-tier NVIDIA chips to create world-class chatbots, causing a market sell-off worth up to $1 trillion. This is like a wide receiver catching a wobbly pass from a backup quarterback, breaking through three defenders, and running 80 yards for a touchdown.

Barrett Woodside, an AI entrepreneur based in San Francisco who previously worked at NVIDIA, said that although U.S. labs are secretive, Chinese companies publicly share their research to prove that even with inferior hardware, they still have competitiveness.

"The question is, when Chinese companies get more and better hardware, will they quickly surpass the United States?" he asked.

In other words, Chinese companies have been training harder to compensate for the lack of a quarterback. What achievements could they achieve if they had a top quarterback like 37-year-old Montana, an American football legend?

This is the question left for people to think about after the Wall Street Journal's thorough analysis and comparison.

This article is exclusive to Observer Group. Unauthorized reproduction is prohibited.

Original: toutiao.com/article/7589617814671983130/

Disclaimer: The article represents the personal views of the author.