"Yin and Yang" OpenAI? Anthropic CEO: Some AI companies are taking on too much risk

Dario Amodei, CEO of the artificial intelligence startup Anthropic, warned that some tech companies are taking on excessive risks in their investments in the AI field. This statement is seen as a subtle criticism of competitors like OpenAI, who are spending lavishly, and indicates that concerns about the sustainability of the current investment frenzy within the AI industry are intensifying. At the DealBook summit held in New York on Wednesday, Amodei stated that the AI industry is facing a "real dilemma." He pointed out the necessity of making substantial investments in infrastructure such as data centers on one hand, and the uncertainty of how quickly AI technology can create economic value on the other. Amodei mentioned that some companies are "YOLO-ing" (meaning "you only live once," implying a reckless gamble). Although he did not name names, he clearly indicated that some participants "are pushing the risk meter too far," suggesting that such reckless spending could pose hidden dangers. These comments come against the backdrop of leading AI companies, including Meta, OpenAI, and Google, significantly increasing their investments in data centers and chips in recent months. Reports indicate that OpenAI alone has committed to investing $1.4 trillion in AI infrastructure projects, exacerbating market concerns about an AI bubble. Investment Dilemma: Uncertainty of Costs and Value Amodei's core point is that the investment required to build advanced AI systems is enormous and time-consuming. He noted in an interview that establishing data centers could take years, and it remains uncertain when and if these massive investments will yield corresponding returns through AI applications. He emphasized that the uncertainty between this investment and return constitutes the "real dilemma" for the industry. "As a company, we try to manage as responsibly as possible," Amodei said. This statement positions Anthropic as a cautious participant amidst the frenzy, contrasting with the aggressive strategies of other companies. Big Players Ramp Up: AI Spending Race Heats Up Amodei's cautious remarks come at a time when there is an intense spending race in the AI field. To build more advanced AI systems and promote technology adoption, industry leaders, including Meta Platforms Inc., OpenAI, and Alphabet Inc.'s Google, are significantly increasing their investments in computing infrastructure. Among these, the $1.4 trillion that OpenAI plans to invest is particularly noteworthy, as this figure is believed to be one of the targets of Amodei's "YOLO" comments and has become a key trigger for discussions about an "AI bubble." Investors and analysts are closely monitoring whether such massive capital expenditures can be transformed into sustainable business models and profits. Taking a Different Path: Anthropic's Cautious Expansion Compared to the trillion-dollar bets of its competitors, Anthropic has not halted its expansion but has adopted a more restrained approach. The company recently announced plans to invest $50 billion to build its first customized data centers in multiple locations across the United States This figure, while impressive, shows a starkly different risk appetite compared to OpenAI's plans. Anthropic was founded in 2021 by former OpenAI employees and has been committed to being a "more responsible AI steward" than its competitors since its inception. The company's focus is primarily on enterprise clients rather than the direct consumer market. According to Reuters, the company raised $13 billion in a funding round in September this year, with a valuation reaching $183 billion, demonstrating that its robust expansion strategy has also gained recognition in the capital markets