"One of the most amazing breakthroughs": How DeepSeek’s R1 model is disrupting the AI landscape
"One of the most amazing breakthroughs": How DeepSeek’s R1 model is disrupting the AI landscape
The Chinese startup's breakthrough is challenging U.S. dominance and shaking the AI market.
“One of the most amazing and impressive breakthroughs I've ever seen,” venture capitalist Marc Andreessen said about the new AI model launched by Chinese startup DeepSeek. The R1 model, released last week, has already climbed to the top of Apple’s App Store download charts and disrupted tech indexes in capital markets. Its launch has sparked questions about the continued technological dominance of the United States.
The model has been praised for its relatively low development costs and reduced reliance on complex chips. In a document accompanying the product launch, DeepSeek explained how it developed a large language model on a limited budget that can learn and improve autonomously without human supervision. “DeepSeek shows that it is possible to develop powerful AI models that cost less,” Vey-Sern Ling, managing director at Union Bancaire Privee, told Bloomberg. “It can potentially derail the investment case for the entire AI supply chain, which is driven by high spending from a small handful of hyperscalers.”
The company was founded by hedge fund manager Liang Wenfeng, and its R1 model is seen as a direct competitor to products from OpenAI and Meta. Industry experts suggest DeepSeek's focus on research makes it a formidable challenger because it shares its breakthroughs instead of guarding them for commercial gain. Despite its growing influence, the company has not yet sought outside funding or monetized its models.
The launch of R1 has already sent shockwaves through the AI industry. Shares of suppliers in the AI supply chain fell sharply in Monday morning trading as the company challenged assumptions about the necessity of enormous computing power and high energy consumption for advanced AI. Nvidia shares dropped in early trading, while Advantest Corp, a major supplier to Nvidia, dropped more than 8% in Tokyo trading.
Another assumption upended by DeepSeek’s breakthrough concerns American dominance in the field of AI. Previously, China was thought to be several years behind the United States due to U.S. government restrictions on exporting advanced chips to China. However, DeepSeek built its model using open-source technology that was relatively easy to access. “While current leaders like Nvidia have a strong foothold, it is a reminder that AI dominance cannot be taken for granted,” Charu Chanana, chief investment strategist at Saxo Markets, told Bloomberg. “The emergence of China’s DeepSeek indicates that competition is intensifying, and although it may not pose a significant threat now, future competitors will evolve faster and challenge the established companies more quickly.”
DeepSeek’s success has also brought attention to its enigmatic founder, Liang. According to the Financial Times, Liang began purchasing thousands of Nvidia chips for an AI side project in 2021, a move many in the industry initially dismissed as eccentric. “When we first met him, he was this very nerdy guy with a terrible hairstyle talking about building a 10,000-chip cluster to train his own models. We didn’t take him seriously,” one of Liang’s business partners told the Financial Times. “He couldn’t articulate his vision other than saying: I want to build this, and it will be a game change. We thought this was only possible from giants like ByteDance and Alibaba.”
Liang officially launched DeepSeek in 2023, declaring his ambition to develop human-level AI. Reports suggest he remains deeply involved in the company’s research and uses his hedge fund income to fund operations, including paying some of the highest salaries for AI engineers in China. “The team believes in Liang’s vision: to show the world that the Chinese can be creative and build something from zero,” the business partner added.
