Operators evaluating dedicated power generation for a large data center campus will eventually land on a comparison between compact fission and combined-cycle gas turbines, or combustion turbine generators deployed in multi-unit configurations. The comparison is not as simple as reading off levelized cost of energy figures, because the cost structures have almost nothing in common. One technology front-loads almost all of its cost into capital and construction. The other spreads cost across operating years through fuel procurement, which introduces a variable that no financial model can accurately project over a 20-year horizon.
We spend time working through this comparison with prospective partners, and the analysis requires being honest about where the numbers are firm and where they are assumptions carrying real risk.
Capital Cost Structure: Front-Load vs. Distributed
A gas turbine generator installation for a 50 to 100 MW data center application is not a cheap project. A multi-unit combustion turbine configuration with associated gas supply infrastructure, SCR emissions controls where required, and grid interconnection work carries a capital cost in the range of $800 to $1,400 per installed kilowatt depending on location, configuration, and site conditions. For 80 MW of capacity, that is $64M to $112M in upfront capital before you account for fuel supply contracts.
Compact fission at comparable capacity has a higher capital cost per kilowatt at this stage of technology development. We are not pretending otherwise. The cost of the reactor vessel, containment, safety systems, and balance of plant is substantial, and the absence of a long track record for compact commercial units means contingency margins are larger. The capital cost difference is real and needs to be reflected honestly in any comparison.
What the capital comparison misses is the operating cost structure that follows. A gas turbine at 80 MW running at high capacity factors will burn through several hundred million cubic feet of natural gas per year. Fuel cost is the dominant operating expense and it is fully exposed to spot market and contract pricing, which has historically shown swings of 4x to 6x across market cycles. A facility that locked in a 15-year fuel supply contract at $3.50 per MMBtu in 2019 looked financially conservative by 2020. By late 2022 it looked prescient. By 2024 it looked expensive again. You cannot model this away.
Fuel Exposure and the Long-Duration Risk Problem
Infrastructure decisions for large data centers are 20-year planning exercises. The facility design, cooling infrastructure, and power architecture you build in year one constrains your options through year twenty. Gas turbine operations commit you to fuel cost exposure for the full operational life of the asset, with whatever hedging instruments your treasury team can put in place.
Uranium fuel for a fission unit is a categorically different cost profile. Fuel costs for a small fission plant are a small fraction of total operating cost, and uranium prices, while not immune to market variation, do not move in the same patterns as natural gas. More importantly, a fission unit's fuel consumption per unit of output is low enough that even a significant move in uranium spot pricing has limited impact on the total cost of energy delivered. The cost exposure simply does not compound across twenty years the way gas does.
The point here is not that fission is cheap and gas is expensive. It is that they carry different types of cost risk, and the appropriate comparison depends heavily on how you account for uncertainty over a long horizon. A financial model that uses a fixed gas price assumption across twenty years is not a useful tool. You need to run the comparison across price scenarios, not just point estimates.
Carbon and Regulatory Exposure
Gas combustion at the scale required to power a large data center generates significant CO2 emissions per year. For facilities with corporate sustainability targets, scope 2 or scope 1 emissions commitments, or exposure to carbon pricing mechanisms in their jurisdiction, this is a financial liability that does not always appear in upfront infrastructure cost comparisons.
California, where some of our early conversations are concentrated, has an active carbon cap-and-trade program with permit prices that have been rising. A data center in California running dedicated gas turbines accumulates a carbon liability that is real and increasing. Other jurisdictions are at earlier stages of carbon pricing implementation, but the policy direction in most major markets is toward pricing carbon, not away from it. A 20-year asset decision made in 2026 will spend most of its operating life in a regulatory environment that is different from today's.
Fission generation has near-zero direct carbon emissions at the point of generation. There are lifecycle emissions associated with mining, enrichment, and fabrication of nuclear fuel, but the operational CO2 is negligible. For operators managing carbon commitments, this is not a minor footnote in the comparison.
Where Gas Turbines Still Win
There are real advantages to gas turbines that should not be discounted in this comparison. The technology is mature, supply chains are well-established, and maintenance logistics are predictable. Finding qualified O&M contractors for combustion turbine assets is straightforward in most industrial regions. Lead times from commitment to first power for a gas turbine installation are meaningfully shorter than for a compact fission plant at our current stage of licensing.
For operators who need dedicated power capacity within a three-to-four-year window and cannot wait for advanced reactor licensing timelines to mature, gas turbines are the practical answer. We are not arguing that compact fission should replace gas for every application. We are arguing that for operators with planning horizons of eight years or longer, high capacity factors, and real carbon exposure, the TCO comparison shifts in ways that are worth working through carefully.
The Framework for the Actual Comparison
A credible TCO comparison between these two technologies requires at minimum: capital cost including site preparation and interconnection; projected O&M over 20 years; fuel cost modeled as a range of scenarios rather than a point estimate; carbon cost at current and projected pricing in the relevant jurisdiction; and the value of avoided grid exposure, including demand-response constraints and interconnection queue risk if the alternative is a new grid connection rather than dedicated generation.
When we run this comparison with potential partners, we build the scenarios together using their specific location, load profile, and planning timeline as inputs. The answer is different for a 20 MW facility in Texas than for an 80 MW facility in California, and different again for a facility in a jurisdiction with aggressive carbon pricing. There is no universal answer, which is why we approach these conversations as planning exercises rather than product pitches.
What we know from working through these models is that the crossover point, where 20-year compact fission TCO is competitive with gas even accounting for the higher upfront capital, is reachable at moderate gas price scenarios for large, high-capacity-factor facilities. At aggressive gas price scenarios or with carbon pricing factored in, the crossover happens earlier and the margin grows. That is the honest answer, not a confident assertion that fission always wins on cost.
