OpenAI has published a new set of commitments for future data centres that it owns and operates, promising to pay the full cost of electricity and supporting infrastructure required by its facilities. The company’s National Community Compact, announced on November 5, also covers water use, public reporting, construction impacts, local employment and the use of nondisclosure agreements with public officials. The policy arrives as the rapid expansion of AI computing has intensified debate over electricity demand, water resources and the effect large facilities can have on nearby communities.
The central promise is that communities should not be left paying for the additional electricity and grid infrastructure required by an AI campus. OpenAI says it will cover the full cost of the power it uses and the infrastructure needed to serve its facilities. Before major spending begins, the company says it will provide financial assurances intended to protect other electricity customers if a project is delayed, reduced or cancelled. The commitment applies to future compute campuses that OpenAI owns and operates.
The distinction between a company pledge and a binding utility arrangement matters. Data centres can require substantial new generation, transmission capacity, substations and local distribution upgrades. If those costs are not assigned clearly, utilities may recover some expenses through rates paid by households and other businesses. OpenAI’s compact says the company will pay its way, but the practical effect will depend on project contracts, utility rules, financial guarantees and the ability of regulators to enforce the relevant obligations.
The compact covers power, water and public accountability
OpenAI is also promising a water plan for each campus and says it will prioritise closed-loop or low-water cooling systems. Water consumption has become a major point of concern in areas where large computing facilities are proposed, particularly when communities already face pressure on local water supplies. The company says it will publish site-level water and electricity use. It also commits to publishing construction plans and independent noise studies before operations begin and annually afterward.
The reporting provisions aim to give residents information that is often difficult to obtain during early planning. A data centre project can affect local roads, construction traffic, noise, electricity demand and water systems long before the facility begins processing AI workloads. Publishing plans and impact information could make it easier for residents and local authorities to assess a proposal. The compact does not, however, provide a single national numerical limit for water consumption, noise or electricity use; the commitments can be adapted to each site.
OpenAI wants local communities involved before construction
OpenAI says every future campus it owns and operates will have a local compact. The company intends to engage residents early enough for their input to influence project plans, appoint a local point of contact throughout the life of the campus and publish an annual account of its performance. It also says it will avoid nondisclosure agreements with public officials, except for narrow protections covering legitimate security information, trade secrets and proprietary or competitively sensitive material.
The NDA commitment is significant because confidentiality agreements can limit what elected officials and public agencies can discuss publicly during negotiations with developers. OpenAI’s position is that local communities should have a meaningful voice in decisions about projects that can reshape infrastructure and land use. The company’s pledge is voluntary and applies to its own projects. It is not a general prohibition on NDAs across the data-centre industry, although the policy blueprint urges lawmakers to establish clearer limits and disclosure rules.
The commitments will be tested at future campuses
The compact also includes commitments around local employment and procurement. OpenAI says it will publish expected job numbers, work with unions and training providers to prepare residents for available roles, and report annually on hiring and apprenticeships. It also intends to give local suppliers enough information about purchasing plans to compete for contracts. These measures cover more than the construction phase: operating a large data centre requires workers who maintain electrical systems, cooling equipment, networks and physical security over many years.