Mistral CEO: Europe has two years to build AI infrastructure

French AI startup Mistral CEO Arthur Mensch has warned that Europe has a two-year window to build its own AI infrastructure or risk becoming permanently dependent on American technology companies. Speaking during a hearing on digital sovereignty at France's National Assembly, Mensch argued that access to chips, computing power, and energy resources will determine competitiveness in the AI industry. He noted that the United States is deploying a trillion dollars next year into AI infrastructure, and warned that once supply is monopolized by American players, Europe will lose its ability to "transform electrons into tokens." The CEO, whose company is valued at approximately $13.6 billion and has positioned itself around digital sovereignty and open-source AI, said Mistral plans to build one gigawatt of AI computing capacity by 2029. He also criticized Europe's fragmented regulations and capital markets, saying they make it harder for startups to scale compared to the United States. In contrast to the concerns about technological dependency, macro strategist Jim Bianco has offered an optimistic perspective on AI's impact on the workforce. Writing in a lengthy post on X, Bianco argued that AI will serve as a new interface for enterprise software, consolidating multiple clunky programs into a single conversational prompt, similar to how the iPhone collapsed cameras, GPS devices, and music players into one device. He suggested companies will happily pay for an AI layer that eliminates the bloated costs of fragmented software stacks. Bianco cited the historical precedent of the electronic spreadsheet, which did not eliminate accounting jobs but instead created demand for more complex financial modeling. Drawing on Jevons Paradox, he argued that making resources more efficient increases total demand, and that AI will allow employees to focus on judgment and strategy rather than mundane tasks. While acknowledging we are near the peak of an AI hype cycle, Bianco compared the current moment to the internet in 1999, noting that the dot-com crash did not mean the internet was a failure. Meanwhile, South Korea is exploring how AI profits could be used to address growing anxiety about concentrated wealth among its youth, who feel driven to despair by fears that even a lifetime of work will never be enough to get ahead. A presidential adviser has proposed a "people's dividend" idea that would distribute AI-generated wealth more broadly. The developments in South Korea reflect a broader global conversation about how AI wealth and power should be managed and distributed, a tension that runs parallel to Europe's concerns about technological sovereignty and the debates sparked by Mensches warnings.