Century Aluminum stalls a $4 billion Oklahoma smelter as 11 terawatt-hours of industrial load competes with datacenter interconnection
The joint venture with Emirates Global Aluminium would double domestic primary production, but the facility cannot break ground without a long-term power contract from AEP.
Eleven terawatt-hours of annual dispatch obligation and $4 billion of capital are waiting on a single interconnection agreement in Oklahoma. The joint venture between Century Aluminum and Emirates Global Aluminium, designed to double domestic primary aluminum production, has stalled at the grid edge. The structural reality is that the grid no longer has room for a 750,000-metric-ton industrial load without a dedicated expansion, and the utility is weighing that demand against a queue of datacenters.
The mechanism driving the delay is the shifting hierarchy of industrial power. Historically, aluminum production dictated regional grid planning because smelters provided massive, stable baseloadThe minimum continuous level of electricity demand on a grid over a 24-hour period. Historically, this was served by large coal or nuclear plants designed to run around the clock without stopping. demand. Today, Oklahoma Primary Aluminum is attempting to secure a competitive, long-term power contract from Public Service Company of Oklahoma — an AEP subsidiary — in a market where tech companies are willing to pay premiums for the exact same capacity. The smelter requires the equivalent of Boston’s annual electricity consumption, but it cannot absorb the interconnection delays or tariff premiums that hyperscalersMassive cloud computing providers that operate data centers at a global scale, predominantly Amazon Web Services, Microsoft Azure, and Google Cloud. Their infrastructure forms the physical backbone of the modern internet and artificial intelligence. routinely write off as capital expenditures.
The numbers define the bottleneck. The proposed facility, originally targeting a 2026–2030 construction window, is slated to produce 750,000 metric tons of primary aluminum annually. It is supported by a $500 million grant from the Department of Energy and hundreds of millions in state incentives. Yet the physical settlement of that power remains unresolved. Century Aluminum’s recent maneuvers illustrate the severity of the constraint: the company recently sold an idled Kentucky smelter to a datacenter operator specifically for its existing grid capacity, while Alcoa is negotiating similar offloads in New York.
The winners in this constrained environment are the utilities and the tech sector, which currently hold the leverage to dictate terms and absorb existing industrial power allocations. The losers are domestic heavy manufacturers attempting to reshore operations in response to global supply shocks and Middle East disruptions. Tariffs on imported metals improve the margins for existing domestic producers, but they do not generate the physical megawatts required to bring a new facility online.
What this impasse forecloses is the assumption that capital and federal incentives alone can force a manufacturing renaissance; the grid must actually be able to carry the load. What it opens is a zero-sum competition for the remaining high-capacity interconnection points in the United States. The next phase of industrial policy will not be decided by trade tariffs or tax credits, but by which sector secures the right to draw power from the wire.