Matthew Yglesias argues that the rapid growth of AI data centers is driving up electricity demand and carbon emissions, creating tension with existing climate targets. He suggests that rather than denying the value of AI, policymakers should consider a carbon tax specifically on data centers, which voters already dislike, to account for the global social cost of carbon. Yglesias notes that while data center emissions are not uniquely large compared to other sectors, new energy demand makes it harder to hit decarbonization goals. He frames the issue as a tradeoff between economic benefits from AI and the global harms of additional warming, especially for vulnerable populations in the tropics. The piece concludes by calling for accelerated deployment of clean energy and proposes the carbon tax as a politically palatable way to internalize the climate cost of data center growth, since the tax would not directly hit consumers.
