Blue Energy, GE Vernova Hitachi advance Texas gas-plus-nuclear plant
Two gas turbines will feed a Crusoe data center by 2030, with small modular reactors adding capacity in 2032.
A power station in Victoria, Texas, is scheduled to deliver about one gigawatt to a data center being developed by Crusoe in 2030, but that first gigawatt is coming from natural gas. The nuclear half of the project, up to five BWRX-300 small modular reactors, is not expected to begin producing until 2032. That two-stage timeline is at the center of an agreement Blue Energy and GE Vernova Hitachi Nuclear Energy announced to move the 2.5GW hybrid plant forward in engineering design, licensing and safety analysis, Data Center Dynamics first reported.
The immediate plan is gas. Blue Energy will install two GE Vernova 7HA.02 gas turbines to supply roughly 1GW to the Crusoe campus by 2030. Starting in 2032, as many as five of GE Vernova Hitachi's BWRX-300 reactors would come on line, adding up to 1.5GW and bringing the total to 2.5GW. The entire build still hinges on a final investment decision expected in 2027.
Blue Energy is not a reactor manufacturer. The company designs modular nuclear plants that are prefabricated at shipyards and transported to the construction site, an approach it calls the 'Blue Way.' The model is meant to turn nuclear construction from one-off engineering into repeated factory-style builds. The BWRX-300 is a 300MW boiling water reactor, and the first unit is already under construction at Ontario Power Generation's Darlington site in Canada, with completion targeted by the end of the decade.
Jake Jurewicz, Blue Energy's CEO and co-founder, said the agreement keeps the company moving toward a nuclear production line that unlocks abundant energy. He described the 'Blue Way' as replacing the old way of building ever-larger reactors on unique schedules, compressing both cost and the time to power. Eric Gray, chief executive of GE Vernova's power segment, said the project pairs proven gas turbine technology with the reactor design. Meeting surging electricity demand, he said, requires scalable, integrated technologies, and the two companies are creating a blueprint for delivering reliable baseload power at the scale and speed customers need.
Gas buys time
Combustion turbines are a mature, off-the-shelf product, so they can be ordered and commissioned on a schedule that SMR licensing cannot match. The Crusoe data center gets power in 2030, two years before any reactor is due to operate. That also means the gas plant is likely to produce revenue before the nuclear half of the station produces a single electron.
The reactor schedule depends on a reference point that is still being built. Darlington is the first BWRX-300 anywhere, and its construction record will shape how investors judge Victoria. The overlap between Darlington's targeted completion and Victoria's 2027 final investment decision suggests the project's backers expect to learn from Darlington before committing to the reactor spend, though that connection is unconfirmed.
Crusoe's gigawatt
Crusoe is the anchor customer, though the announcement does not say whether a power purchase agreement is in place. The plant is expected to supply up to 1.5GW to a Crusoe data center, three-fifths of the station's eventual output. The gas turbines account for about 1GW of that total; the extra 500MW would only be available after the reactors start up. The announcement does not identify who would buy the remaining 1GW, or when that capacity would be contracted.
That unspoken slack capacity is itself a feature of the plan. The gas turbines are sized to serve the data center's near-term load, while the reactor build-out roughly doubles output. Whether that second phase finds a new customer, an expanded Crusoe contract, or merchant sales is an open question.
One underwriting or two
The 2027 final investment decision will reveal how the project's two halves are financed. A single underwriting would treat the gas plant and the reactors as one asset, with early gas revenue helping carry the nuclear construction. A split structure would fund the gas turbines first, then require the reactors to stand on their own economics once the turbines are online. The first path would make the hybrid model a bridge to a nuclear future; the second would leave the reactors to prove themselves as a separate bet.
For private-market investors, the phased approach is a case study in staging risk around an AI data center load. The gap between the 2030 gas start and the 2032 reactor start is where the project's financial logic will be tested. The 2027 FID will say whether the turbines and the reactors are one project or two.