Data center power bill stalls; the states already had the rule

The first significant congressional effort to address the consumer-cost side of the data center boom cleared the House on Wednesday by 417-3 and stalled in the Senate a day later, when an attempt by Senator Jon Husted, an Ohio Republican, to move it on a fast track was blocked by Democrats. Bisnow reads the block as leaving little chance of passage before November's midterms, which puts the Ratepayer Protection Act on hold for the rest of the year.

The margin is the number worth carrying into an investment committee. A 417-3 vote says the principle behind the bill — that data centers should carry the grid costs they create — commands broad bipartisan agreement. That is an inference from vote arithmetic, and it shapes capital allocation: political risk on power costs is now a state-by-state variable with bipartisan backing, the kind that a lobbying budget in Washington cannot settle.

Its substance is thinner than its vote: the bill would require states merely to consider a framework ensuring that data centers of at least 100 megawatts pay the full cost of the electricity needed to serve them, and nothing compels a state to act, while the criticism of the measure did not come only from Democrats.

ClearView Energy Partners told clients in a Thursday note, according to Utility Dive, that the legislation sits behind the regulatory curve, since the steps it asks states to consider are already being undertaken across most of the country. A majority of states — including data center hotbeds Virginia, Texas and Ohio — are already pursuing or have enacted measures meant to ensure data centers pay their own freight, which means the federal bill would have reinforced a trend already underway without starting it.

None of that reduces the pressure that produced the bill: customers on PJM Interconnection, the nation's largest power grid, have already seen their bills rise largely because of surging demand from data centers, and the report describes the potential for rapid development to push electricity costs higher as one of the primary flashpoints in a nationwide wave of opposition to the industry. Husted's support for data center development has become a liability with voters in his state, which suggests his fast-track attempt carried more political weight than the bill's text did.

The stall is less damaging to the sector than it looks, because the constraint it failed to address was already set at the state level. If state action is the binding rule, then a data center's diligence file has to price what the federal bill never touched: the host state's cost-allocation regime, the grid upgrades and new generation the local utility will need to serve the load, and whether the political temperature that produced those rules survives the next election.

As this publication has argued, industrial capital is now paying for land, credit and freight position rather than rent rolls, and the sector's repricing runs between assets priced on lease term and assets priced on optionality. Data center sites are where that test has the sharpest edge, because the optionality in question is regulatory: a campus in a state that has already assigned grid costs to large loads carries a quantifiable expense line, while one in a state still debating the question carries a contingency that no stabilized pro forma prices cleanly, and the power cost that makes a site work is a variable a legislature can move.

The politics around the measure cut the same way: Senate Democrats cast the bill, and the Republican urgency behind it, as election-season theater, with Minority Whip Dick Durbin of Illinois dismissing it, according to Politico, as last-minute legislation on a subject live in a Senate contest. Republicans countered that Democrats had blocked a bipartisan measure aimed at the public's most-cited concern about the artificial intelligence infrastructure boom.

Husted said he would keep pushing the issue, and the report does not close the door on eventual Senate passage, only on action before the midterms. The more consequential calendar is in the statehouses, where the cost-allocation measures ClearView says the federal bill would merely have reinforced are already moving. What failed was the uniformity, not the cost. The next underwriting file will have to price each statehouse's answer separately.