On Monday, Chinese artificial intelligence startup DeepSeek became the most-downloaded free app in the US on Apple’s App Store, dethroning rival OpenAI’s ChatGPT. The move marked a significant shift in the AI landscape, with global tech stocks selling off and wiping out billions in market capitalization.
DeepSeek’s AI Assistant has been praised for its performance and reasoning capabilities, and its open-source nature has made it accessible to developers. The startup’s R1 model, released earlier this month, was built at a fraction of the cost of rival models, with estimates suggesting that it was developed at a cost of around $5.6 million.
The news has sent shockwaves throughout the AI industry, with some tech leaders and analysts warning of a potential bubble. The rapid pace of innovation and funding in the AI sector has raised questions about the sustainability of the market.
OpenAI’s CEO Sam Altman tweeted a quote attributed to Napoleon, saying that a revolution can be neither made nor stopped, but only given direction by successful battles. Microsoft’s chief AI scientist, Yann LeCun, wrote on LinkedIn that DeepSeek’s success is indicative of changing tides in the AI sector favoring open-source technology.
The news comes as the US government has restricted chip exports to China, a move that could affect the development of AI models. Despite the challenges, analysts are optimistic about the potential of the AI sector, with predicted revenue of over $1 trillion within a decade.
Other tech giants, such as Meta, Google, and Microsoft, are also responding to the shift, with Meta starting “war rooms” to focus on potential competition from DeepSeek. Microsoft CEO Satya Nadella wrote on X that the DeepSeek phenomenon is an example of the Jevons paradox, where increased efficiency and accessibility lead to increased demand.
The news has also sparked debate about the future of the AI industry, with some experts warning of a potential bubble. Raymond James analysts detailed the questions plaguing the industry, including the implications of open-sourced vs. proprietary models, the effectiveness of throwing money at GPUs, and the potential consequences of US export restrictions.