According to Alexander Wang, CEO of Scale AI, the United States may have led China in the artificial intelligence race for the past decade, but on Christmas Day everything changed.
Wang, whose company provides training data to key AI players including OpenAI, Google AND Metasaid on Thursday at the World Economic Forum in Davos, Switzerland, that DeepSeek, China's leading artificial intelligence lab, released an “earth-shaking model” on Christmas Day and then developed the powerful, reasoning-centric artificial intelligence model DeepSeek-R1, which competes with the recently the released o1 OpenAI model.
“We found that DeepSeek… has the best performance or is about on par with the best American models,” Wang said.
In an interview with CNBC, Wang described the U.S.-China AI race as a “war on artificial intelligence,” adding that he believed China had achieved much more Nvidia H100 GPUs – AI chips that are widely used to create leading-edge, powerful AI models – than you might think, especially given US export controls.
Wang also said he thinks the AI sector will reach $1 trillion, matching estimates that the generative AI market will be ready highest $1 trillion revenues within a decade.
“The United States will need enormous computing power and enormous amounts of infrastructure,” Wang said, later adding: “We must unleash American energy to enable the artificial intelligence boom.”
Earlier this week, Mr. President Donald Trump announced A joint venture With OpenAI, Oracle AND SoftBank invest billions of dollars in America's artificial intelligence infrastructure. 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 early technology partners will include Microsoft, Nvidia and Oracle, as well as a semiconductor company Arm. They announced that they would invest $100 billion initially and up to $500 billion over the next four years.
In an interview Thursday, Wang said he believes achieving artificial general intelligence, or AGI, a widely cited but vaguely defined benchmark used in the artificial intelligence sector to define a branch of artificial intelligence that uses technology that equals or exceeds human intellect in many tasks. AGI is a hotly debated topic, with some leaders saying we are close to achieving it, and some saying it is impossible at all. Wang said his own definition of AGI is “powerful artificial intelligence systems that can use a computer just like you or me… and can essentially work remotely in the most efficient way.”
Anthropic, AmazonAI-backed startup AI, founded by former OpenAI research directors, has accelerated technology development throughout the past year, and in October the startup he said that its AI agents were able to use computers like humans to perform complex tasks. Anthropic's Computer Use feature allows its technology to interpret computer screen content, select buttons, enter text, navigate websites and perform tasks using any software, and browse the Internet in real time, the startup says.
The tool can “use computers basically the same way we do,” Jared Kaplan, Anthropic's chief science officer, told CNBC in an interview at the time. He said he could perform tasks consisting of “dozens or even hundreds of steps.”
OpenAI apparently plans introduce a similar feature soon.
When asked which U.S. AI startups are currently leading the AI race, Wang said each model has its own strengths — for example, OpenAI models are great at reasoning, while Anthropic models are great 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-centric AI model, DeepSeek-R1.