Wisconsin's $2.5B Data-Center Transmission Line: Who Pays When Private Gain Demands Public Wires

Urban Milwaukee reported in September 2026 that American Transmission Co. (ATC) had resubmitted a $2.5 billion transmission project to the Wisconsin Public Service Commission after the commission unanimously revoked its prior approval in August, citing more than 560 revisions that made oversight impossible^[1]. The line is meant to deliver power to Oracle's share of a $15 billion data-center campus in Port Washington, built by Vantage Data Centers for Oracle and OpenAI as part of the $500 billion Stargate artificial-intelligence infrastructure program^[5].

The mechanism at stake is a regulatory choice: how much of a private corporation's infrastructure cost the utility can shift to the general ratepayer base. ATC, a regulated monopoly transmission owner, earns a FERC-regulated return on every dollar of capital it deploys. That means the utility's financial interest is pure: build the line, rate-base it, and collect a guaranteed return forever. The larger the project, the larger the profit. The initial cost estimate was $1.4 billion to $1.6 billion; after the PSC sent it back, the revised proposal climbed to $2.5 billion to $2.7 billion, citing a new voltage-regulation facility^[1][9]. No independent study certified the increase was necessary; no competitive bidding process compared it to alternatives. ATC alone studied whether the line was needed, ATC alone designed it, and ATC alone will build and own it.

The question that has not yet been raised on the record is whether a 100-mile transmission corridor spanning six counties should be funded by Wisconsin's 3 million electricity customers, or whether Oracle, which will consume nearly one gigawatt of computing power and will pocket the margin between its costs and its revenue, should bear the full cost of the wires its operations require. Currently, the default is cost-sharing: ATC proposes the project, the PSC approves cost allocation, and the utility recovers its capital and operating costs, plus its allowed return on equity, from all ratepayers regardless of who benefits. The Citizens Utility Board, Wisconsin's ratepayer advocate, has not yet taken a public position on cost allocation, though its concerns about the project's scale are on record^[1].

Complicating the picture: Oracle has already negotiated financial security measures. The PSC declined to ease credit requirements on We Energies (the retailer serving the data center), meaning Oracle faces a potential $7 billion letter of credit guarantee to shield ratepayers from the risk that the company defaults or abandons the campus^[6]. Oracle has also stated it will offset additional electricity costs to prevent ratepayer bill impact^[8]. But those commitments are separate from the transmission line itself. The line is ATC's asset, not Oracle's; ATC will own the poles, the wire, and the right to earn a return for thirty years or more. If Oracle's guarantee covers the retailer's exposure but not the transmission owner's, the cost-allocation boundary is exactly where the regulatory fight belongs.

The path forward: Before the PSC approves either the line or its cost allocation, it should require ATC to file a alternatives analysis comparing the full cost (including the utility's return on capital) of the proposed build to genuine grid-enhancing technologies, dynamic line ratings on existing corridors, advanced power-flow control, or storage-as-transmission from battery plants Oracle itself could site near the data center. If Oracle's load truly requires new wires, the PSC should order competitive bidding under Wisconsin's equivalent of FERC Order 1000, forcing ATC to justify its bid against third-party proposals. And the commission should separate the cost allocation decision: Oracle should pay the line's full capital cost plus ongoing easement fees to landowners, with Wisconsin ratepayers covering only the interregional benefits, if any, that the line provides beyond Oracle's consumption. The current framework, one utility, no competition, full cost-allocation to captive ratepayers, is the mechanism that turns $1.6 billion into $2.7 billion with each revision cycle.