While federal proposals on data center energy consumption remain fragmented, states are taking the lead. Public utility commissions in Georgia, Ohio, and Michigan are implementing "large-load tariffs" to force data centers, not households, to bear the costs of necessary grid upgrades.
To address voter concerns about affordability, the administration may pivot on seemingly unrelated policies like trade. A potential Supreme Court ruling limiting presidential tariff authority could be framed as an opportunity to pursue a lighter-touch tariff policy, alleviating cost pressures exacerbated by the AI buildout.
Public opposition to AI data centers is materializing in key states where voters directly link the infrastructure buildout to higher personal electricity costs. This tangible affordability issue is proving more potent politically than general concerns about AI's impact on employment, influencing local votes on new projects.