OpenAI says it will stop using non-disclosure agreements with public officials as it develops future data center projects, a policy shift that addresses a recurring source of concern in communities considering large computing facilities.
The company announced the commitment after a week in which data center expansion became a prominent political issue. Its new approach is intended to make discussions with government officials more open while preserving limited protections for information that is genuinely sensitive.
A narrower approach to confidentiality
OpenAI said it will not use NDAs with public officials. It will instead allow only narrow protections for legitimate security information, trade secrets, and proprietary or competitively sensitive information.
The distinction matters because NDAs can prevent local officials from discussing proposed projects, negotiations and impacts with residents. Large data centers can raise questions about land use, electricity demand, water, grid upgrades, employment and public incentives. When the terms governing a project are not publicly discussable, communities can have less opportunity to assess those tradeoffs.
OpenAI’s move follows a similar October announcement from Amazon, which said it would no longer use NDAs with government agencies it works with on data center projects. The changes point to increasing pressure on major technology companies to make the public-facing parts of infrastructure development more transparent.
Power and infrastructure commitments
OpenAI also said it will pay the full cost of the electricity it uses and the infrastructure needed to serve its facilities. That commitment directly addresses a central concern around the buildout of AI computing capacity: whether new, power-intensive campuses could shift costs onto existing households, businesses or public systems.
The company did not detail how those costs would be calculated, what projects the commitment already covers, or how the arrangements would be independently verified. Those implementation details will be important, because electricity and grid costs are typically handled through a mix of utility regulation, interconnection agreements, local approvals and contracts with developers.
Still, the statement sets a public standard OpenAI can be measured against. In practical terms, local governments and regulators will be able to ask whether a planned facility’s electricity use and required upgrades are being paid for by the project rather than transferred to other customers.
Local compacts for future campuses
For every future compute campus that OpenAI owns and operates, the company says it will establish a “local compact.” Each compact will include the new commitments and reflect local priorities.
OpenAI is using “compute campus” as its term for data centers. The framing emphasizes the facilities’ role in supplying computing power for AI systems, but it does not reduce the underlying local questions. A campus still depends on physical infrastructure: electrical capacity, network connections, construction activity and, depending on its design, cooling systems and related services.
The local-compact approach could give communities a more formal vehicle for defining priorities. The value of that framework will depend on what each agreement contains, whether its terms are public, and how residents and officials participate in setting them. OpenAI has not described a standard format or enforcement mechanism in its announcement.
Jobs and public reporting
The company also said it will publish the jobs it expects to create and report on hiring. That is significant because job creation is often one of the principal benefits presented for large infrastructure projects, while the number and type of ongoing positions can differ from the temporary construction workforce.
Publishing expectations and later reporting on hiring would create a clearer record for comparing promises with outcomes. The announcement does not specify which job categories will be included, whether contractors will be counted, or how frequently the figures will be updated.
For local leaders, those details may determine whether the reporting meaningfully captures the economic effect of a facility. For communities, it could help separate construction-period employment from the permanent staffing associated with operating a computing site.
A response to a changing debate
OpenAI’s commitments show that the debate over AI infrastructure is no longer confined to model performance and corporate investment. It is increasingly about the terms under which physical computing capacity is built: who gets information, who bears costs and what benefits remain in the host community.
The policy does not resolve every question attached to a proposed data center. It does, however, place three commitments on the record: no NDAs with public officials, full payment for the company’s electricity and needed infrastructure, and local compacts for future campuses it owns and operates.
As OpenAI pursues additional projects, the test will be whether those commitments are translated into accessible agreements, specific cost arrangements and reporting that communities can evaluate. The company’s decision makes transparency part of its stated infrastructure strategy; its project-level execution will determine how consequential that change becomes.
This article is based on reporting by The Verge. Read the original article.
Originally published on theverge.com
