Marc Benioff, Ray Dalio, and the cofounder of Alibaba talk US and China AI
"China is behind, but not by a lot, in the best chips," said Ray Dalio at a Wednesday conference in Singapore.
admin
Jonathan Raa/NurPhoto
- US and Chinese AI advancements took center stage at a major conference on Wednesday in Singapore.
- Salesforce's CEO criticized Big Tech's spending on data centers and AI development.
- Alibaba's cofounder said open-sourced models could spur innovation outside Big Tech companies.
Some of the biggest names in tech and finance talked AI in Singapore on Wednesday, highlighting the hot debate about American and Chinese tech advancements.
At the CNBC conference, Salesforce's CEO, Marc Benioff, said some tech leaders have fallen into a "hypnosis" about the number of data centers and the level of training needed to create AI models. The emergence of seemingly low-cost, high-tech Chinese models like DeepSeek's R1 shows that these investments might not be necessary, he said.
"It has to be rethought. Exactly what are you doing and why are you doing this?" he said of the Big Tech companies spending hundreds of billions of dollars on hardware.
Benioff has been vocal about his dislike of how much capital Big Tech is earmarking for data centers and AI advancement.
During the same panel, Bridgewater Associates founder Ray Dalio said that the US is still winning against China in designing leading chips. But he said China has the upper hand in actually using AI.
"China is behind, but not by a lot, in the best chips," Dalio said.
Joe Tsai, Alibaba's cofounder, said the "so-called DeepSeek moment" is more about the open-source movement than which country has the best AI. Open-source models allow for the free and open sharing of software to anyone for any purpose.
Tsai predicted a rush of development on top of existing open-source models — which won't just benefit Big Tech companies.
With 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," he said.
In January, DeepSeek, a Chinese startup, launched an open-sourced AI model that rattled US tech and AI companies. Third-party tests showed the model outperformed its peers from OpenAI, Meta, and other top developers in some tasks, and the company said it was built for less money.
