Log in to the Sina Finance APP and search for [Information Disclosure] to view more rating levels. The US Congress's scrutiny of AI data center construction is further expanding. Jamie Raskin, the ranking Democratic member of the House Judiciary Committee, has sent letters to Amazon, Google, Meta, and Oracle, demanding that the four companies disclose their use of non-disclosure agreements with government officials in data center projects and provide more detailed project information.
This investigation focuses not on AI models themselves, but on how many project terms, energy requirements, resource consumption, and government support arrangements are being kept out of public view through non-disclosure agreements when large technology companies advance data center construction at the local level.
Congress begins asking why data center projects need confidentiality
Raskin has asked the four companies to explain their data center project portfolios and the scope and specific ways in which non-disclosure agreements (NDAs) are used in related projects. His core concern is that if the basic terms of a project remain confidential for a long time, local residents will find it difficult to assess what impact a data center will have on electricity prices, water resources, local finances, and surrounding communities, and will also find it difficult to oversee the government officials responsible for approving the projects.
Amazon responded that the company no longer uses such NDAs with government partners involved in data center projects. Google, Meta, and Oracle did not immediately comment as of the time of publication.
Confidentiality issues are becoming an important part of opposition to data centers
For some time, disputes over AI data centers in many parts of the United States have clearly intensified. The issues of concern to local communities are mainly concentrated in several areas: large data centers may drive up the cost of power infrastructure and consume large amounts of water resources, while many projects also receive tax incentives or other local support. In this situation, if a large number of confidentiality agreements also exist in project negotiations and approval processes, residents often only learn the specific scale and resource needs after the project has been approved, which has gradually made "lack of transparency" one of the important reasons for opposing data center construction.
The technology industry believes that some confidentiality arrangements are commercially necessary. For example, disclosing a site and scale too early before a project is finalized may leak competitive information, and may also lead local contractors and suppliers to raise quotes after learning that a large technology company is involved. Therefore, the core of the current dispute is not whether all project details should be disclosed, but where the boundary should be drawn between companies' commercial confidentiality needs and the transparency of local government public decision-making.
Congress has begun pushing to restrict data center NDAs
This letter is not an isolated action. On September 24, Raskin, together with Democratic Representative Greg Landsman and Republican Representative Tom Barrett, jointly proposed the "No Secrets for Data Centers Act," which plans to restrict federal, state, and local government officials from using NDAs to conceal information related to data center projects. According to the bill's description, if government officials hide data center project details through confidentiality agreements, the relevant NDAs will be restricted or invalidated; if a data center receives government incentives and the incentive conditions include requiring the government side to keep project information confidential, that may also be restricted. This shows that the regulatory focus is gradually shifting from simply asking companies to "voluntarily increase transparency" to limiting confidentiality arrangements between governments and technology companies through legislation.
AI data centers are turning from industrial projects into political issues
AI infrastructure was previously mainly viewed as a technology and capital expenditure issue, but as the number and scale of data centers grow rapidly, it is increasingly directly affecting local power grids, land, water resources, and fiscal policy. This has also gradually brought what was originally a highly technical data center construction process into broader public policy discussion. Reports show that data centers have become an important issue in some states and electoral districts. Some localities have begun to re-examine tax incentives previously granted to projects, and members of Congress have also begun to demand that technology companies bear more of the cost of power infrastructure or improve project transparency. Therefore, what this inquiry into Amazon, Google, Meta, and Oracle reflects is a change: regulation around AI data centers in the United States is moving from "whether they can be built and how much power is needed" further toward "how projects were approved, who bears the cost, and how much local residents actually knew before approval." If this trend continues, future large AI data center projects may face constraints not only in electricity and land supply, but also in stricter information disclosure and local government transparency requirements.
Comments