Columnist predicts AI slowdown in the US after data center votes — Fortune
In the United States, the development of artificial intelligence and the construction of data centers may temporarily slow after local votes on such facilities. This assessment was made by Mark Minevich, founder of Going Global Ventures, in a column for Fortune.
According to the author, 41 local ballot measures concerning data centers were put to a vote in 10 states. Voters sided with opponents of the industry in every case except one vote in Harwood, North Dakota. In most of these votes, the margin exceeded 70%.
Possible new rules
Minevich predicts that after the new US Congress convenes in January, rules for developers of advanced AI models may become stricter. Among the possible steps, he named mandatory model testing under government standards, stronger whistleblower protections, transparent incident reporting, and independent audits.
More current news is available on the UA.News Telegram channel Telegram.
In the author’s view, discussions are also possible regarding liability for the actions of AI agents, export restrictions on equipment, and regulation of so-called surveillance pricing. This refers to the use of data on website browsing, purchases, or a consumer’s geolocation to set a personalized price.
Requirements for data centers
Minevich believes that to gain public support, companies will have to pay for power grid upgrades needed to connect facilities themselves, sign long-term contracts, publish data on water and electricity use, and reduce the load on the grid during peak hours.
Andrey Danis, head of digital infrastructure at AlixPartners, suggested that restrictions on large campuses could strengthen the role of smaller regional data centers located where spare electricity capacity is available. According to him, data centers that already have all permits and are under construction may become more valuable, while projects without the necessary permits for electricity, water, and operations risk delays or cancellation.
