Core Scientific closes Polaris deal, expands Muskogee campus to 1.5GW
The $444 million Polaris acquisition delivers 440MW of grid power and 40 acres, letting Core Scientific push its Oklahoma campus to 1.5GW.
Core Scientific has closed its $444 million cash purchase of Polaris DS, a bitcoin miner with an adjacent site in Muskogee, Oklahoma. The acquisition adds 440MW of grid-connected power. It also adds 40 acres of land. At that price, the grid capacity costs about $1 million per MW, with the acreage folded in. Data Center Dynamics reports the deal, first announced in May, transfers the existing electric service agreements Polaris holds with Oklahoma Gas & Electric.
With that interconnection in hand, Core Scientific can expand the Muskogee campus to 1.5GW of gross power. Of that, 1GW is leasable, using grid connections alongside behind-the-meter generation. The site's history goes back to an April 2022 announcement with the Muskogee City-County Port Authority for a 500MW blockchain and crypto mine at the John T. Griffin Industrial Park. Polaris followed with its own 2024 cryptomine plan on a nearby 42-acre parcel. Core Scientific started in 2017 as a cryptomining operator and has since turned to AI data center development; it runs facilities in Alabama, Georgia, Kentucky, North Carolina, North Dakota, Oklahoma, and Texas. That 1.5GW target triples the scale announced for the campus in 2022.
The first building at Muskogee is 70MW. It is scheduled to be handed over to CoreWeave in Q2 2026. The next building is 82MW, Core Scientific says. An unnamed customer is on track to receive it in the second half of 2027, more than a year after the CoreWeave transfer. CEO Adam Sullivan said the Polaris deal follows site acquisitions this year that added more than 600MW of leasable power. What $444 million buys is an existing seat at the grid, the hardest asset to replicate in the data center boom. Acquiring a neighbor's connection is faster than building one, which is why former miners with power contracts are being repurposed into AI data center capacity.