H.R. 9372: Data Infrastructure Energy Measurement and Standards Act

This bill would direct the National Institute of Standards and Technology (NIST), working with the Department of Energy, to study and develop better ways to measure how much energy and water data centers use. It focuses on data centers broadly, including the energy use tied to artificial intelligence training and inference as well as other compute-heavy tasks.

What the bill would require

• Run a research program to improve how data center energy and water use is measured.
• Develop or identify best practices, definitions, methods, procedures, and technical standards for measuring and reporting that use.
• Study different kinds of data center workloads, including AI workloads, and how they affect total energy and water consumption.
• Consider how power use changes over time, including on-site power generation, grid-connected generation, and energy not sent to the grid.
• Account for differences in data center design, such as servers, storage, networking, cooling systems, power distribution, and backup power.
• Look at how location, climate, and other local factors affect electricity demand.
• Study what data is needed to improve energy and water demand forecasting, including gaps in data collection, public and private data availability, and researcher access.
• Support standardized metrics and data-sharing methods that could help researchers and stakeholders better forecast energy and water demand.
• Coordinate with the Energy Department to manage and archive the data appropriately and encourage sharing at the federal, state, academic, and industry levels where practical.
• Work with international partners and standards organizations to keep measurement standards aligned globally.
• Consult with industry, academics, nonprofits, standards groups, civil society, and federal agencies during the work.

Reporting and funding

• NIST would have to brief Congress on the program within one year of enactment and again within two years.
• The bill authorizes $10 million per year for fiscal years 2027 through 2029.

Who would be affected

The bill does not directly regulate data centers or require immediate reporting by private companies, but it could affect companies that build, operate, supply, or rely heavily on data centers by pushing toward new measurement standards and data practices.

Relevant Companies

• META — Large operator of data centers that support its digital services and AI workloads; may need to adapt to new measurement or reporting standards.
• GOOGL — Operates extensive data center infrastructure for cloud and AI services; could be affected by new energy and water measurement standards.
• AMZN — Through AWS, operates major data center facilities and may be affected by standards for measuring and reporting resource use.
• MSFT — Runs large-scale data centers for cloud computing and AI; could be impacted by standardized measurement and forecasting efforts.
• ORCL — Operates cloud data centers and could be affected by evolving measurement and reporting practices.
• AMD — Sells chips used in data centers and AI systems; indirect effects could come through changing data center efficiency and workload measurement practices.
• NVDA — Supplies GPUs widely used in AI and data center workloads; may be indirectly affected if measurement standards influence data center design and optimization.
• INTC — Supplies processors and other data center hardware; could face indirect effects from standards that shape data center energy-use reporting and infrastructure choices.
• DUK — Electric utility with significant data center load exposure in some service areas; improved forecasting standards could affect planning and demand estimation.
• NEE — Utility and energy infrastructure company that may be indirectly affected by data center power-demand forecasting and siting trends.