AI Arms Race Heats Up in Asia: US Seeks to Counter China’s Growing Influence

BEIJING – The intensifying artificial intelligence race between the United States and China is casting a long shadow across Asia, the world’s largest continent. Experts believe the American strategy is to prevent China from emerging as the dominant AI supplier not only for Asia but for the entire globe. While China’s AI solutions are often more cost-effective, the U.S. currently holds an advantage by offering a more comprehensive ecosystem, from advanced chips to sophisticated AI models.

However, this U.S. sales challenge was subtly underscored at the Asia-Pacific Economic Cooperation (APEC) "Digital Weeks" held in Chengdu this month. Despite earlier efforts by a U.S. official and a U.S. business representative to APEC to highlight AI in promoting the event, the U.S. left minimal public evidence of its participation. This subdued presence follows a period of upheaval for U.S. AI firms, including Anthropic’s delay in releasing its Fable AI model due to sudden U.S. policy shifts, and the recent launch of new Chinese AI models offering comparable capabilities at a significantly lower cost.

In contrast, the previous summer at the inaugural APEC AI meeting in South Korea, Michael Kratsios, then President Donald Trump’s chief science and technology policy advisor, actively promoted the U.S. AI Action Plan and the establishment of the American AI Exports Program. This was detailed in remarks shared by the White House.

Despite these initiatives, a recent report by Politico, citing three former officials, indicated that the Commerce Department has received a lower-than-anticipated number of applications – just 78 – for the American AI Exports Program. Further complicating the U.S. narrative, on July 24, Bill Guidera, deputy under secretary for innovation and engagement at the U.S. Department of Commerce, speaking at an APEC High-level Forum on AI organized by China’s cybersecurity regulator, continued to emphasize the AI exports program. He broadly called for Asia-Pacific partnerships, highlighting that buyers could procure either the complete U.S. technology stack or specific components through the program, describing it as a "brilliant design that shows the strength, security and capability of U.S. AI."

The International Trade Administration of the Commerce Department confirmed Guidera’s participation in Chengdu via a social media post on July 29. When questioned about the Politico report, an ITA spokesperson stated that the volume of applications "exceeded our expectations." The White House did not provide a comment on the matter.

American corporate participation was also notably limited. As of July 23, CNBC observed Google and Meta as the only U.S. companies with exhibition booths at APEC. These were positioned alongside booths from Thailand and China, which primarily featured Chengdu-based enterprises. The two U.S. tech giants focused on their respective advancements in molecular AI systems like AlphaFold and AI applications designed for small businesses, rather than showcasing large language models (LLMs). Wilson L. White, Google’s vice president of government affairs, made only brief mentions of Gemini during his speech on July 24. Following White, Cai Guangzhong, Vice President of Tencent, took the stage to highlight the increasing adoption of its Hunyuan LLM and a cloud project in Thailand.

Beijing’s AI Diplomacy Gains Traction

China’s hosting of APEC this year coincides with a critical juncture marked by heightened U.S.-China tensions and intense technological rivalry. Beijing has seized this opportunity to prominently feature its AI capabilities, which are largely characterized by open-source models, in contrast to the predominantly closed-source U.S. offerings.

At the World AI Conference in Shanghai on July 17, Chinese President Xi Jinping announced that China would provide developing countries with 5,000 training and seminar opportunities in AI. Furthermore, the nation pledged to establish AI application cooperation centers with Southeast Asia and other regions. Subsequently, a high-ranking Chinese official, Vice Premier Zhang Guoqing, advocated for the development of tech standards with other Asia-Pacific nations at the minister-level APEC Digital Weeks on July 23. By the end of that day, the 21 member economies, including the United States, reached an agreement to support open-source AI with "strong security."

Wei Sun, principal analyst for artificial intelligence at Counterpoint Research, commented that the "endorsement" of open-source models with robust security assurances lends greater regional legitimacy to China’s open-weight strategy, particularly among emerging Asian economies where deployment costs and technological sovereignty are significant considerations.

The Complexities of Integration in a Multilingual Landscape

However, analysts anticipate a future where AI adoption in Asia is not strictly divided into U.S. or Chinese spheres but rather involves a hybrid approach. The sheer linguistic diversity of Asia, with over 1,300 living languages in Southeast Asia alone, presents a challenge for the straightforward implementation of either U.S. or Chinese AI models.

Governments across Asia and beyond are investing "billions" in AI systems tailored to local languages, according to the privately funded startup Votee AI. CEO Pak-Sun Ting stated that his company is collaborating with at least five governments, including two in Southeast Asia, and is already generating over $10 million in annual revenue. Ting noted that entities in Southeast Asia often utilize Nvidia chips for AI training but may opt for different chip manufacturers for model execution. He also pointed out that Votee’s open-source model for Cantonese speakers was partly developed using Alibaba’s open-source Qwen model.

The ability of AI to generate economic returns remains a paramount concern, irrespective of its origin. Yue Su, principal economist at the Economist Intelligence Unit, observed that despite the fierce competition and distinct approaches of the U.S. and China in AI technology and diplomacy, complete decoupling is not feasible. She suggested that the U.S.’s relatively low profile at APEC was understandable given the scheduling of other significant events, such as an AI Summit in San Francisco on July 24.

South Korea’s tech ministry organized this San Francisco event, where President Lee Jae-myung aimed to leverage the nation’s substantial investments in chip and AI megaprojects. This involved meetings with leaders of frontier AI models, including Sam Altman of OpenAI, Dario Amodei of Anthropic, and Jensen Huang of Nvidia.