OpenAI and Anthropic have won corporate valuations approaching $1 trillion from investors, but the steady emergence of cheaper Chinese artificial intelligence (AI) models is raising questions about such lofty price tags. Yet GIC, Singapore's sovereign wealth fund and a major AI investor, took a positive view, forecasting that the arrival of Chinese AI models would significantly reduce the global cost of AI adoption. At the same time, GIC acknowledged the strong growth of Chinese AI companies while maintaining a cautious stance on investing in Chinese AI startups.
Despite GIC's positive assessment, the rapid advance of Chinese AI models is heightening a sense of crisis within the United States. The administration of Donald Trump is weighing regulatory options against Chinese AI models, while some voices are raising concerns that such measures "would hamper competition and innovation."
AI Investment 'Big Hand' GIC: "Expecting Cost Savings From Chinese AI"
On the 23rd, GIC Chief Investment Officer Bryan Yeo said in an interview with the Financial Times (FT) that "open-weight AI models are increasingly emerging from various countries, including China," adding, "This is expected to bring cost savings and have a positive impact on the global AI ecosystem."
He predicted, "As costs fall, AI adoption will spread faster and more widely across more companies and industries, and as a result, AI usage will grow exponentially."
GIC has made AI a core part of its investment strategy, investing billions of dollars in the sector over the past several years. In February, it led Anthropic's $30 billion funding round, committing a large sum.
Yeo did not answer when asked whether open-weight models released by Chinese AI companies such as DeepSeek and Moonshot would pose a threat to Anthropic and OpenAI, which adhere to expensive, cutting-edge closed models.
He said GIC has observed the strong growth of Chinese AI companies but takes a cautious stance on investing in Chinese AI startups. "We need to examine much more closely the startups' capacity to invest in research and development (R&D) to grow their AI models," he said. Since 2024, GIC's asset allocation to the Americas has risen to 53% of its total portfolio from 44%, while its share in the Asia-Pacific region has declined to 22% from 28%.
'Kimi Shock' Follows DeepSeek... US AI Dominance Wavers
With the recent "Kimi shock" following the DeepSeek shock, doubts are growing over whether Silicon Valley's massive AI investments will translate into commensurate results. Bloomberg reported that "confidence in US AI leadership has also developed cracks."
Moonshot's self-assessment claimed that its new model "Kimi K3," unveiled on the 17th, outperforms all competing models in overall performance except Anthropic's "Claude Fable 5" and OpenAI's "GPT-5.6." K3 was assessed as being of an unusual scale for an open-weight model, with 2.8 trillion parameters. AI performance analysis firm Artificial Analysis assessed that K3 outperformed Anthropic's "Opus 4.8" in some frontier benchmarks — the first time a Chinese open-weight model has achieved such a result. Bloomberg said the K3 launch has shaken the conventional wisdom that the US leads China in AI.
In response, the US is considering effectively banning its companies' access to Chinese AI products. On the 20th, Axios cited a government source as saying the government is considering a "slow but persistent" approach to pressure US companies using Chinese AI models by applying procurement rules or designating them on trade restriction lists.
However, even within the US, concerns are being raised that restricting competition could leave the country behind in innovation. David Sacks, co-chair of the President's Council of Advisors on Science and Technology, recently said, "Two closed AI labs that have built an oligopoly want the government to eliminate their open-source competitors," adding, "It's time for the overwhelming majority of Silicon Valley companies that recognize the value of open competition to take action against this." Sacks has warned of the risk of "regulatory capture," in which a small group of regulated interests captures legislative and regulatory authorities to change the rules in their favor.
