A data-center deal can look straightforward from the outside, with a site, a power requirement, a construction plan, and a projected investment figure. Behind those visible components, however, the commercial terms can determine how parties allocate financial exposure long before construction crews arrive. That makes confidentiality more than a legal preference because withheld information can affect how parties assess and negotiate the costs attached to power, land, tax incentives, construction capacity, and infrastructure upgrades. Recent reporting has put nondisclosure agreements around major AI data-center projects under renewed scrutiny, while industry representatives have argued that confidentiality can protect proprietary information and give developers room to negotiate without prematurely inflating costs.

The harder infrastructure question sits underneath that debate: does keeping a project commercially quiet change the economics of the project itself? If a developer negotiates separately with utilities, local authorities, landowners, contractors, and infrastructure providers, counterparties may have access to different portions of the broader commercial picture, potentially affecting the information available during each negotiation. That can protect commercially sensitive negotiations, but it can also make it harder for outside observers to determine how contractual responsibilities and infrastructure risks are allocated across a project. The issue therefore deserves to be examined as a pricing problem rather than simply as a secrecy problem.

NDAs Can Change the Negotiating Environment

Confidentiality can serve a practical purpose during early-stage infrastructure negotiations because developers may need to discuss future capacity, utility availability, incentives, financing structures, land arrangements, and construction plans before those details become public. Contractors can also have legitimate concerns that premature disclosure of a project could encourage competitors to increase bids when they know an operator has limited time or unusually high compute demand. Industry representatives have made versions of that argument, saying confidential negotiations can prevent proprietary information from spreading and help developers secure infrastructure without unnecessarily increasing costs. The problem begins when confidentiality extends beyond commercially sensitive information into the basic economic architecture of a deal.

If the public cannot determine the scale of an incentive, the structure of a power commitment, the conditions attached to land, or the responsibilities for supporting infrastructure, the market loses useful information about the true cost of development. That missing information matters to contractors because their pricing depends on schedule certainty, scope certainty, material exposure, and the customer’s ability to absorb changes. It matters to utilities because large-load commitments can require additional power-delivery infrastructure and other investments needed to serve major data-center loads. It also matters to end users because decisions about infrastructure costs can affect electricity customers when utilities recover investments associated with serving large data-center loads.

Confidentiality Can Become an Infrastructure Pricing Mechanism

The more interesting question is whether NDAs have effectively become another commercial instrument for setting infrastructure prices. A conventional contract establishes what each party receives and what each party must provide, while confidentiality terms can influence which information remains available to other participants before a broader transaction takes shape. Consider a developer negotiating simultaneously for a site, a major power connection, tax incentives, and construction capacity. If counterparties see only portions of the broader transaction, the developer may retain greater control over information about the project’s commercial structure, potentially affecting the negotiating environment. That does not automatically make the arrangement unfair, nor does it prove that any particular project has shifted costs improperly.

It does mean that information can become part of the negotiating leverage surrounding infrastructure, particularly where confidentiality covers land prices, utility availability, or competing incentives. Research and reporting on data-center deals have documented cases in which NDAs and confidential agreements covered project details involving land, incentives, utility use, and development plans, showing that the confidentiality question can reach beyond corporate trade secrets. For an industry developing facilities that can require hundreds of megawatts, even a relatively small change in the allocation of infrastructure costs can become material when applied across multiple projects. The pricing question therefore deserves attention before anyone reduces the debate to whether an NDA exists.

Disclosure Does Not Mean Publishing Every Commercial Secret

A stronger disclosure model need not require companies to publish proprietary engineering specifications, security information, workload details, bidding strategies, or commercially sensitive intellectual property. Those protections can have legitimate commercial value, particularly when operators compete for power capacity, sites, construction resources, and equipment, and industry representatives have cited proprietary information and pricing concerns as reasons for confidentiality. The useful boundary sits elsewhere, around information that determines how infrastructure costs and commitments are allocated. A project should be capable of revealing its material power obligations, major public incentives, infrastructure responsibilities, significant construction commitments, and the conditions that could transfer costs if schedules or capacity assumptions change. Such disclosure could give utilities, contractors, local authorities, investors, and end users a clearer view of a project’s economic structure without requiring every commercial detail to become public.

It could also create a more useful basis for comparing projects by placing the visible investment number alongside material infrastructure obligations supporting it. That matters as data-center financing involves developers, capital providers, utilities, technology companies, and long-duration contractual commitments, with recent projects illustrating how financing and contractual obligations can interact with power-delivery risks. Recent reporting on a large AI data-center project in New Mexico illustrates how power delays can interact with contractual payment obligations and financing arrangements, showing that infrastructure risk can continue accumulating even when a facility has not reached operation. Disclosure cannot eliminate that risk, but greater visibility into contractual obligations could make its allocation easier to understand.

The Price of AI Infrastructure Is Being Negotiated Before Construction

Data-center development increasingly combines large power requirements with complex land, financing, construction, and network dependencies. The economic structure of a project can therefore change depending on which party secures favorable terms before construction begins. Confidentiality can remain part of that process, particularly where genuine competitive information requires protection. But the infrastructure industry has a stronger reason to disclose the economic commitments that affect third parties and long-term cost allocation. The objective should not be to expose every negotiation, but to make the material consequences of those negotiations measurable.

Contractors need sufficient information to price construction risk, utilities need sufficient information to assess load commitments, and end users need sufficient information to understand which costs could move through the infrastructure chain. Without that visibility, AI infrastructure pricing could reflect differences in access to commercial information as well as differences in construction efficiency. The provocative issue is therefore not simply whether data-center deals contain NDAs. It is whether NDAs are becoming part of the negotiating mechanism that can influence the economics and ultimately the price of AI infrastructure.