1. United States
2. Wash.
3. Letter

I am writing to urge you to support legislation requiring transparency and regulation of artificial intelligence energy consumption. The AI industry's explosive growth is occurring without public accountability for its environmental and economic impacts on constituents. Recent research from Lawrence Berkeley National Laboratory reveals that AI-specific servers consumed 53 to 76 terawatt-hours in 2024, enough electricity to power 7.2 million homes. By 2028, AI could consume 165 to 326 terawatt-hours annually, potentially tripling data centers' share of US electricity from 4.4% to 12%. These projections exclude emerging AI agents and reasoning models that require 43 times more energy than current simple queries. The fundamental problem is that major AI companies including OpenAI, Google, and Anthropic refuse to disclose energy consumption data, claiming trade secret protections. Without transparency from these companies, researchers cannot accurately assess AI's true environmental impact or plan for future energy needs. The US Energy Information Administration does not even treat AI as its own sector, leaving detailed data nonexistent. This opacity has direct consequences for your constituents. Harvard research found that utility agreements with tech companies often include discounts that raise consumer electricity rates. Virginia ratepayers, for example, could pay an additional $37.50 monthly for data center energy costs. Meanwhile, data centers use electricity that is 48% more carbon-intensive than the US average because they cluster in regions with dirtier grids and operate around the clock. Tech giants are investing unprecedented sums in AI infrastructure. SoftBank, OpenAI, Oracle, and MGX plan $500 billion over four years for US data centers. Google allocated $75 billion for AI infrastructure in 2025 alone. These investments will lock in decades of energy demand without public input. I urge you to support legislation mandating that AI companies disclose energy consumption per query and model, require environmental impact assessments before major data center construction, and ensure ratepayers are not subsidizing corporate AI infrastructure. Constituents deserve transparency about how AI is reshaping our energy grid and who bears the costs.