Rep. Ro Khanna, D-Calif., has introduced a Data Center Bill of Rights that would give local governments greater authority over the rapid growth of data centers supporting AI.
Khanna’s proposal would allow communities to prohibit data centers in residential areas and within 2,500 feet of homes, schools, childcare facilities, hospitals and nursing homes. Local governments could also reject proposed projects through a public review process and retain regulatory authority without state governments overriding their decisions.
The resolution addresses a nationwide conflict surrounding AI infrastructure: More than 4,400 data center facilities are listed across the US, while hyperscale facilities supporting compute-intensive AI workloads could directly consume as much as 33 billion gallons of water annually by 2028, according to projections cited by Khanna’s office.
Facing a still larger number of facilities as AI grows, communities are now frequently resisting data center construction, based on local quality of life issues and concern that the facilities will increase ratepayer utility bills.
As hyperscalers expand data centers, utilities may need additional generation and transmission capacity. Khanna’s resolution would enable states to pause new data center development until protections are established to prevent households and small businesses from absorbing higher electricity costs. It also calls for data centers to use clean, reliable energy and comply with limits governing noise and air pollution.
The resolution would also require operators to disclose projected water requirements to local utilities, report water consumption twice annually, and receive a compatibility determination before projects are approved.
Another provision calls for independent impact studies before construction is approved. These reports would examine public health, pollution, electricity demand, utility costs, and housing availability. Developers would also disclose the number and types of jobs expected from a facility and its projected economic impact over a 10-year period.
This provision could bring greater scrutiny to the economic case for data centers. While construction can generate substantial temporary employment, even large facilities can employ relatively few (sometimes less than 200) once the data facility is up and running.
Khanna also wants operators to negotiate enforceable community benefit agreements that could fund affordable housing, infrastructure improvements, workforce training, public computing resources and permanent tech jobs. Data center owners would additionally be expected to pay state and local taxes, while public subsidies would carry transparency requirements and provisions for recovering benefits if agreed conditions are not met.
Khanna’s proposal is yet another example of how AI infrastructure is becoming a political concern as well as a tech investment. A March Gallup survey found that 71% of Americans opposed construction of AI data centers in their local communities.
