Zuckerberg's Higher Bid Fails: Insider Details on Google's $650 Million Acquisition of DeepMind Revealed

According to 1M AI News, journalist Sebastian Mallaby’s new book “The Infinity Machine: Demis Hassabis, DeepMind and the Quest for Superintelligence” is set to be published on March 31. The Wall Street Journal has released an exclusive excerpt, providing the first detailed account of the 2013 competition between Google and Facebook (now Meta) for the acquisition of DeepMind. This book is based on over 30 hours of interviews with Hassabis, as well as dozens of conversations with DeepMind colleagues, investors, and others involved in the acquisition. In June 2013, then-Google CEO Larry Page expressed his interest in acquiring DeepMind to founder Demis Hassabis at Elon Musk’s birthday party: “Your real mission is to create AGI, why not leverage the resources I’ve accumulated?” Hassabis recalled that this statement convinced him: “I was tired of running around raising money. I went to Google to get a bunch of computational resources and then solve the intelligence problem.” Facebook CEO Mark Zuckerberg was also involved in the bidding. Facebook’s head of corporate development, Amin Zoufonoun, proposed a plan that would make the founders wealthier by lowering the equity acquisition price but offering substantial signing bonuses to the founders and core members. However, Zoufonoun dismissed the AI governance issues raised by DeepMind co-founder Mustafa Suleyman (now CEO of Microsoft’s AI division). Hassabis later dined at Zuckerberg’s home, deliberately steering the conversation from AI to virtual reality, augmented reality, and 3D printing, discovering that Zuckerberg was equally excited about all technologies. “That told me everything I needed to know,” Hassabis later said, “Facebook offered a higher bid, but I wanted someone who truly understood why AI was more important than anything else.” During negotiations, Suleyman used his poker instincts to bluff, emphasizing to Google that DeepMind had billionaire investors like Peter Thiel and Musk backing them (“Of course, those people weren’t really supporting us”). Hassabis set several conditions for the sale: DeepMind would remain in London, military applications would be prohibited, and an independent ethics and safety review board composed of external scientists and philosophers would be established to dilute Google’s control over the technology. Google’s chief negotiator Don Harrison called these conditions “a big issue for me,” but ultimately conceded because “if we didn’t absolutely believe that Demis represented the future of our AI strategy, we couldn’t agree to this framework.” By the end of January 2014, Google completed the acquisition for $650 million. Following the rejection, Zuckerberg recruited deep learning pioneer and NYU professor Yann LeCun to establish Facebook’s AI lab, and LeCun immediately attempted to poach core researchers from DeepMind. Mallaby describes this acquisition in the book as “a bargain by today’s standards,” with the true returns gradually materializing over the next decade as Google invested billions of dollars in research funding for DeepMind.