China Narrows AI Gap With US as Open-Source Models Reshape Valuations: George Yeo
Source: The Business Times
Former Singapore Foreign Minister George Yeo warned that the rapid rise of open-source AI models is narrowing China's gap with the US and could erode the proprietary 'moats' that support Big Tech valuations. Speaking at a Standard Chartered event in Singapore, he also called for more AI power centres beyond the US and China.
Former Singapore Foreign Minister George Yeo has warned that the rapid rise of open-source artificial intelligence models is fundamentally reshaping the competitive dynamics between the US and China, with significant implications for Big Tech valuations and global power structures. Speaking at a Standard Chartered H2 Global Market Outlook event in Singapore on July 9, Yeo said the accelerating shift toward open-source LLMs threatens to commoditise AI software and erode the proprietary "moats" that have sustained premium pricing power for major developers.
Yeo observed that China's AI capabilities have closed dramatically since the emergence of DeepSeek. "China appeared to be six to nine months behind the US, and then it is now three months... and then people wonder when China will overtake the US," he said during a fireside chat with StanChart's global CIO Steve Brice. He cited Zhipu AI's latest GLM-5.2 model, developed by a Tsinghua University-led team, as evidence of this fast-shrinking gap. China is now producing roughly half of the world's LLM experts, he noted, quoting Nvidia CEO Jensen Huang.
The former minister, who served as Singapore's foreign affairs minister from 2004 to 2011, framed AI as a civilisation-level transformation. "AI is a new phase in the story of man. It is as fundamental as the arrival of agriculture — everything in human society will change because of AI," he said. He added that every country will pursue sovereign autonomy in AI, with national security as the driving force, and suggested that India could emerge as a critical third pole to bring geopolitical balance to AI development beyond the current US-China duopoly.
On the investment side, StanChart's CIO team pushed back against the notion that AI is a bubble, arguing that ongoing capital expenditure in AI infrastructure will sustain growth for at least three to four years. The team had closed its overweight semiconductor position in May but is now reconsidering a re-entry point following the recent market sell-off. However, they remain cautious on physical AI — robotics and hardware integration — drawing parallels to the hype cycle that left hundreds of Chinese EV companies struggling for profitability.
Why it matters for Singapore: As a small, trade-dependent nation that has positioned itself as a neutral hub for AI governance and innovation, Singapore has a direct stake in how the US-China AI rivalry plays out. Yeo's analysis — delivered from a Singapore stage — underscores the city-state's unique vantage point in observing and navigating the geopolitics of AI. His call for more AI power centres beyond Washington and Beijing aligns with Singapore's longstanding advocacy for a multipolar global order where smaller nations retain agency in shaping technology governance.