Scale AI CEO Alexandr Wang said that for the past decade, the United States may have been ahead of China in the artificial intelligence race, but on Christmas Day, everything changed.

Wang’s company provides training data to major artificial intelligence players including OpenAI, Google and YuanDeepSeek, China’s leading artificial intelligence lab, released an “earth-shattering model” on Christmas Day and followed up with a powerful AI model focused on inference, DeepSeek said Thursday at the World Economic Forum in Davos, Switzerland. -R1, which competes with OpenAI’s recently released o1 model.

“We found that DeepSeek … performed best, or roughly on par with the best American models,” Wang said.

In an interview with CNBC, Wang described the AI ​​race between the United States and China as an “AI war,” adding that he believed China had more technology. NVIDIA The H100 GPU—an AI chip widely used to build leading powerful AI models—is more popular than one might think, especially given U.S. export controls.

Wang also said he believes the AI ​​field will reach trillions of dollars, matching estimates for the generative AI market. Up to $1 trillion Income over ten years.

“The United States is going to need a lot of computing power, a lot of infrastructure,” Wang said, later adding, “We need to unleash the energy of the United States to achieve the prosperity of artificial intelligence.”

Earlier this week, President Donald Trump announced joint venture and open artificial intelligence, Oracle and SoftBank Invest billions of dollars in U.S. AI infrastructure. The project, Stargate, 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 semiconductor companies arm. They said they would invest $100 billion at launch and $500 billion over the next four years.

In an interview on Thursday, Wang said he believed it would take two to four years to reach artificial general intelligence (AGI), a widely cited but vaguely defined benchmark in artificial intelligence that represents a branch of artificial intelligence that pursues Technology equals or surpasses human intelligence in many tasks. AGI is a hotly debated topic, with some leaders saying we’re close to achieving it while others say it’s simply impossible. Wang said his own definition of AGI is “a powerful artificial intelligence system that is able to use a computer like you and I do… and basically become a remote worker in the most capable way possible.”

Anthropic selection, AmazonThe artificial intelligence startup founded by a former OpenAI research executive has ramped up technology development over the past year and in October explain Its artificial intelligence agents are able to use computers just like humans to complete complex tasks. Anthropic’s computer usage capabilities enable its technology to interpret content on a computer screen, select buttons, enter text, browse websites and perform tasks from any software and real-time internet browsing, the startup said.

Jared Kaplan, Anthropic’s chief scientific officer, told CNBC at the time that the tool could “use computers in basically the same way that we do.” He said it could complete the task in “tens or even hundreds of steps.”

open artificial intelligence Reportedly planned A similar feature will be rolled out soon.

When asked which U.S. AI startups are currently leading the AI ​​race, Wang said each model has its own advantages — for example, OpenAI’s model is good at reasoning, while Anthropic’s model is good at coding.

“Competition in this space is getting more intense, not less,” he said.

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