Meta (META) unveiled three agreements Friday to secure up to 6.6 GW of nuclear power by 2035 for its data centers that are among the most sweeping and ambitious so far between tech companies and nuclear power providers.
Meta (META) announced a deal with Vistra (VST) to purchase electricity from three existing nuclear plants in Ohio and Pennsylvania, and will back new small modular reactor projects that nuclear developers Oklo (OKLO) and TerraPower are planning to build over the next decade; the deals follow an agreement announced in June to obtain energy from a Constellation Energy nuclear site.
Vistra (VST) +14.2% and Oklo (OKLO) +17% pre-market.
Meta’s (META) 20-year agreements to buy power from Vistra’s (VST) Perry and Davis-Besse plants in Ohio and Beaver Valley plant in Pennsylvania will provide more than 2,600 MW of energy, including 2,176 MW of operating generation beginning later this year and an additional 433 MW of combined power output increases, which Vistra said will comprise the largest nuclear uprates supported by a corporate customer in the U.S.
Vistra (VST) said the PPAs provide certainty to pursue subsequent license renewals for each of the reactors, which would extend each license an additional 20 years; currently, Beaver Valley Unit 1 is licensed through 2036, Davis-Besse is licensed through 2037, Perry is licensed through 2046, and Beaver Valley Unit 2 is licensed through 2047.
Oklo (OKO) said its agreement with Meta (META) advances its plans to develop a 1.2 GW power campus in Pike County, Ohio, and provides a mechanism for Meta to prepay for power and provide funding to advance project certainty for Oklo’s Aurora powerhouse deployment; financial terms were not disclosed.
Meta (META) also will help fund TerraPower’s development of two reactors to generate up to 690 MW of power as early as 2032, and provides Meta with rights for energy from as many as six other TerraPower reactors by 2035; financial terms were not disclosed.
Meta’s (META) new deals follow CEO Mark Zuckerberg’s plans to invest hundreds of billions of dollars through the end of the decade on AI and the infrastructure needed to support it.