On a Tuesday in May, Michigan Attorney General Dana Nessel filed a challenge to a major data center project near Ann Arbor. The target: a facility backed by DigitalBridge, the infrastructure giant that has bet billions on AI-driven demand for compute power.
Nessel's argument is direct: ratepayers deserve transparency about the contracts and potential costs before public subsidies are locked in. The move strikes at the heart of the model that has fueled the data center construction boom—municipal incentives, tax abatements, and utility rate deals negotiated behind closed doors.
DigitalBridge CEO Marc Ganzi has positioned the firm at the center of the AI infrastructure wave. The firm manages over $80 billion in assets, with data centers as a core focus. Ganzi has argued that the industry can grow responsibly, but the Michigan challenge suggests that local communities are not convinced.
The Ann Arbor project is not an outlier. Across the U.S., data center developers have secured billions in tax breaks and utility concessions, often with minimal public scrutiny. The pitch: AI growth will create jobs and tax revenue. The pushback: those benefits may not offset the costs imposed on residential and small-business ratepayers.
Nessel's legal theory centers on the Michigan Public Service Commission's approval process. She argues that the commission failed to require sufficient disclosure of the long-term financial arrangements between the developer and the utility. If she prevails, the ruling could force a new standard for transparency in data center incentive deals nationwide.
The stakes are enormous. Global data center investment is projected to exceed $1 trillion over the next five years, per McKinsey. Much of that capital depends on access to cheap power and favorable regulatory treatment. A legal precedent requiring full ratepayer impact disclosure would slow deal velocity and increase developer costs.
DigitalBridge and its peers have structured their business models around long-term, inflation-protected contracts with hyperscale tenants. Those contracts assume stable or declining power costs. If local regulators begin to push back, the cost assumptions underpinning those deals come under pressure.
The tension is structural. AI data centers consume 10 to 50 times the power of a typical commercial building. They require dedicated substations, new transmission lines, and often, rate structures that shift costs onto other ratepayers. The industry has treated these as engineering problems. Nessel is reframing them as governance problems.
This is not the first challenge to data center subsidies. In Virginia, home to the world's largest data center market, legislators have debated bills requiring economic impact studies before approving tax breaks. In Ohio, a proposed data center near Columbus drew opposition over water usage. But Michigan is the first state attorney general to directly challenge a utility commission's approval on transparency grounds.
The outcome will be watched closely by institutional investors. Pension funds and insurance companies have allocated billions to data center debt and equity, attracted by the narrative of AI-driven demand. A regulatory setback in Michigan could trigger repricing of risk in the sector, particularly for projects that rely on public subsidies.
For DigitalBridge, the timing is awkward. The firm is in the midst of raising its latest infrastructure fund, targeting $20 billion. Investor due diligence will now include a question that was not on the table six months ago: what happens if the subsidy model breaks?
Nessel's challenge is a reminder that the AI infrastructure boom is not just a technology story or a real estate story. It is a political story. And in politics, the people who pay the bills eventually get a vote.