Scale AI’s CEO says China quickly caught up with the US with DeepSeek

According to Alexander Wang, CEO of Scale AI, the US may have led China in the artificial intelligence race for the past decade, but that all changed this Christmas.

Wang, whose company provides training data to key AI players including OpenAI, Google and Metasaid Thursday at the World Economic Forum in Davos, Switzerland, that DeepSeek, China’s leading artificial intelligence lab, released a “disruptive model” on Christmas Day, followed by the powerful reasoning-focused AI model DeepSeek-R1. , which competes with the recently released OpenAI o1 model.

“We found DeepSeek … to be the most effective, or about on par with the best US models,” Wang said.

In an interview with CNBC, Wang described the AI ​​race between the US and China as an “artificial intelligence war”, adding that he believes China has much more Nvidia The H100 GPUs are artificial intelligence chips widely used to build leading powerful AI models – beyond what people might think, especially with US export controls in mind.

Wang also said he believes the AI ​​sector will reach a trillion dollars, which is in line with estimates that the generative AI market is poised to reach top $1 trillion in income for ten years.

“The United States is going to need an enormous amount of computing power, an enormous amount of infrastructure,” Wang said, later adding, “We need to free up the energy of the U.S. to enable this AI boom.”

Earlier this week, the president Donald Trump announced a joint venture with OpenAI, Oracle and SoftBank invest billions of dollars in US AI infrastructure. The Stargate project was unveiled at the White House by Trump, SoftBank CEO Masayoshi Son, Oracle co-founder Larry Ellison and OpenAI CEO Sam Altman. Key initial technology partners will include Microsoft, Nvidia and Oracle, as well as a semiconductor company hand. They said they would invest $100 billion to start and up to $500 billion over the next four years.

In an interview Thursday, Wang said he thought it would take two to four years to reach artificial general intelligence, or AGI, the widely cited but vaguely defined benchmark used in the AI ​​sector to refer to the field of artificial intelligence. which deals with technology that equals or surpasses human intelligence in a wide range of tasks. The topic of AGI is hotly debated, with some leaders saying we’re close to achieving it and some saying it’s impossible at all. Wang said his own definition of AGI is “powerful artificial intelligence systems that are able to use a computer like you or I could … and basically be a remote worker in the most capable way.”

Anthropic, Art AmazonThe AI-powered startup, founded by former OpenAI research heads, has been ramping up technology development over the past year, and in October the startup said that its artificial intelligence agents were able to use computers like humans to perform complex tasks. The Anthropic Computer Use capability allows users to interpret what’s on a computer screen, select buttons, type text, navigate websites and perform tasks using any software and web browsing in real time, the startup said.

“The tool can ‘use computers basically the same way we do,'” Jared Kaplan, Anthropic’s chief scientific officer, told CNBC at the time. He said it can perform tasks with “dozens or even hundreds of steps.”

OpenAI reportedly plans will introduce a similar feature soon.

When asked which US AI startups are currently leading the AI ​​race, Wang said each model has its own strengths — for example, OpenAI’s models excel at reasoning and Anthropic’s at coding.

“The space is becoming more competitive, not less competitive,” he said.

Correction: This article has been updated to correct the name of DeepSeek’s reasoning-based AI model, DeepSeek-R1.