The AI arms race between the United States and China took center stage at a high-profile conference in Singapore on Wednesday. A key point of contention was the role of Big Tech spending. Key figures from technology and finance debated the competitive landscape of advancements in artificial intelligence.
At the CNBC conference, Salesforce CEO Marc Benioff criticized the substantial investments made by major tech companies in data centers and AI model development. He questioned whether these expenditures are truly necessary, especially in light of emerging, cost-effective models from China, such as DeepSeek’s R1.
“It has to be rethought. Exactly what are you doing and why are you doing this?” Benioff said, referring to the hundreds of billions of dollars being spent on hardware by large tech companies. He has been a vocal critic of the capital allocation towards data centers and AI advancements by Big Tech.
During the same panel, Ray Dalio, the founder of Bridgewater Associates, offered a nuanced perspective. He stated that while the US still leads in the design of leading-edge chips, China may have an advantage in the practical application of AI.
“China is behind, but not by a lot, in the best chips,” Dalio said, highlighting the multifaceted nature of the competition.
Joe Tsai, Alibaba’s cofounder, shifted the focus towards the open-source model movement, which he referenced as “the so-called DeepSeek moment.” Open-source models provide free and open access to software, fostering innovation outside of the realm of Big Tech.
Tsai predicted a surge in developments based on existing open-source models that won’t just benefit the biggest companies, with the AI playing field being opened.
With these open-source models, “the AI game is not just left to the five richest companies in the world that can afford to invest $50 billion a year,” Tsai said.
DeepSeek, a Chinese startup, launched an open-source AI model in January, stirring significant interest among US tech and AI companies. Initial tests indicated that the model surpassed its counterparts from OpenAI, Meta, and other top developers in some tasks, and the company claimed that it was developed with a lower budget.
