Donald Trump has characterized the debut of a chatbot by China’s DeepSeek as a “wake-up call” for US tech companies in the global competition for leadership in artificial intelligence. The launch of DeepSeek, which has developed its R1 model chatbot at a significantly lower cost than competitors like OpenAI’s ChatGPT and Google’s Gemini, caused a staggering $1 trillion (£800 billion) loss in value from the leading US tech index on Monday. Nvidia, a top manufacturer of computer chips that has thrived amid the AI boom, experienced a loss of $600 billion in market value, marking the largest single-day decline in US stock market history. “The introduction of DeepSeek, an AI from a Chinese firm, should alert our industries to the necessity of maintaining a sharp focus on winning the competition, ” Trump asserted. He highlighted DeepSeek’s capability to seemingly match the performance of existing AI models while utilizing significantly fewer resources, posing a challenge to the US-led AI growth. “That’s advantageous, as it reduces expenditure, ” Trump remarked. “I see that as a positive factor—an asset. ” On Monday, the DeepSeek assistant surpassed ChatGPT in download figures from Apple’s app store. Sam Altman, the CEO of OpenAI, acknowledged his admiration for DeepSeek but emphasized that the US industry would accelerate its development efforts. “DeepSeek’s R1 is an impressive model, especially concerning the performance they provide at that price, ” he noted. “We will undoubtedly release superior models, and having a new competitor is indeed refreshing.

We will expedite some releases. ” The stock market drop triggered by DeepSeek’s announcement in the US extended to Asia on Tuesday, with Japan’s Nikkei share average declining by 1. 3%. Tech stocks in Japan fell, with the manufacturer Advantest down by 11%, Tokyo Electron off nearly 6%, and Disco Corporation slipping nearly 3%. Shares of tech investor SoftBank plunged over 5%. Most other major Asian markets were closed in observance of the lunar new year holiday. Marc Andreessen, a prominent US venture capitalist, likened the launch of DeepSeek’s R1 model to a crucial transition in the US-USSR space race, stating on X that it represented AI’s “Sputnik moment”—referring to the moment when the Soviet Union surprised its Cold War counterpart by launching a satellite into orbit. DeepSeek claims its R1 model excels over OpenAI’s o1-mini model across “various benchmarks, ” while research conducted by Artificial Analysis positions it favorably against models from Google, Meta, and Anthropic in terms of overall quality. Founded by entrepreneur Liang Wenfeng, who operates the hedge fund High-Flyer Capital that employs AI to analyze stock price patterns, DeepSeek reportedly began with Liang purchasing Nvidia chips in 2021 to develop AI models as a hobby, financed by his hedge fund. In 2023, he established DeepSeek, headquartered in the eastern Chinese city of Hangzhou. The firm prioritizes research over commercial products—the DeepSeek assistant and its base code are available for free download, and its models are also less costly to operate than OpenAI’s o1.