In 2025 you made what you called a decisive strategic decision, moving from oxy-combustion to post-combustion capture. Walk me through that moment and why the decision became inevitable.
If you take a step back to when we took Net Power public in 2023, the premise was that the world is going to continue to use natural gas for power generation. As aspirational as it is to think we can move away from natural gas, the reality is that it is the backbone of a stable, functional grid system, and we'll be burning it for a long time, similar to oil. If we truly believe we can achieve net zero ambitions, we have to find a way to decarbonize natural gas because it's going to be a fixture in the energy scene for decades. Our approach was not predicated on the data center boom. We thought we'd have a longer runway, time to innovate, time to invest in R&D, time to invent this new natural gas cycle that inherently captures the CO2. It was going to require a lot of capital and time, and the cost would be more expensive than the carbon-emitting alternative, but with the market putting a premium on clean power, it's a price the market would be willing to pay.
Fast forward just two years to 2025 and the whole world has inverted. You have this AI data center boom that's disruptive in a good way to the economy. We've gone from a surplus of power generation to a deficit, and priorities have shifted away from clean at all costs. For the last decade, US load growth was flat, so most new generation, especially renewables, was just displacing existing capacity rather than adding to the system. Now we're using up just about all of the generation capacity and we need to add more baseload and natural gas generation. Customers aren't saying they need new power in 2030 or 2031, they're saying they need it next year.
The differentiator between winners and losers isn't going to be who has the chips or the best large language models, it's going to be who has access to the power.
This has become an arms race, and it's not just contained to Microsoft, Meta, Amazon, Google and the hyperscalers. It's a US versus China issue. Are we going to let China win this AI race? The biggest pinch point is power, which I think gets to the heart of your report.
Where exactly does the bottleneck sit in the energy-to-AI dynamic? Walk me through the value chain from the well to the turbine and tell me what has to be fixed for the data centers to get the energy they need.
The oil and gas industry is really well capitalized and financially healthy at this point. Coming out of the shale boom of the last decade, especially the public companies, they're generating excess free cash flow, returning it to shareholders, and keeping production where it needs to be. So it's not a question of whether they can produce more. The real question is whether they can transport the natural gas from the well to where the demand is. Historically, data centers clustered in Data Center Alley in northern Virginia, around McLean, mostly because of fiber-optic cables going to Europe. The Marcellus supply was coincidental. There was enough pipeline capacity to get gas from Texas up the eastern seaboard and from Appalachia. But that area is now tapped out in terms of capacity to add more, it's a land issue. So data center developers are saying, I need to build wherever I can get power.
What we're seeing is a massive shift toward locating data centers next to large-scale power generation, as close to the natural gas resource as possible. You see big campuses popping up in Pittsburgh and western Pennsylvania at the heart of the Marcellus, in Louisiana on the Gulf Coast near the Haynesville, and increasingly in West Texas. Texas has become the largest area for data center growth, both because of low-cost natural gas and because the permitting and regulatory process is so efficient. Natural gas in Texas is a third of the price in California for the same molecule. Speed is criteria number one, and Texas has taken the lead. But we still need more natural gas infrastructure. Gas at these volumes moves by pipeline, and the regulatory process has become so cumbersome that pipelines that used to take months now take years. For one of our pipelines at EQT, it took an act of Congress to approve, and eight years to build, a pipeline from western Pennsylvania down to Virginia. That didn't make the project safer, it just shows how the regulatory process is stymieing safe, accelerated construction. Permitting reform is a big thing the US has to figure out. We sit here saying China is kicking our butts, building 10 nuclear plants and 30 coal plants every month, but it's not a competence issue on our industry side. A lot of what they're doing is based on learnings from watching us decades ago. The only thing that's really changed is our regulatory process. That, more than anything else, is the governor of how quickly we can build to meet this demand.
You have a partnership with Entropy for capture technology, you get the turbine from Siemens, and Occidental is critical because it gives you the offtake. What is Net Power's actual moat?
The moat is more philosophical and strategic than technical. It's the ability to combine disparate elements and put them together to figure out what makes the most sense. When most people think about a clean energy future they think nuclear and renewables, looking purely through the zero-emissions lens. They're ignoring the two biggest criteria in energy selection, when can I get it and what does it cost. Nuclear isn't really available today to build in the United States, partly technology, partly regulatory, but a huge part is affordability. During the shale boom, natural gas went from a higher-cost form of power to the lowest cost, killing off coal and stopping nuclear development because we suddenly had a reliable baseload form of energy at a fraction of nuclear's cost. Energy is a commodity. People expect reliability and affordability. We're looking at this through the lens of how to generate the lowest-cost form of clean, firm, reliable power.
Most folks wouldn't say natural gas is clean power, but there are technologies that can make it clean. It'll be more expensive than the carbon-emitting version, but we're starting with the lowest-cost feedstock for power generation, so even a small premium for capture should be cheaper than nuclear or renewables paired with batteries for 24/7. Our real skill is pairing natural gas power generation with the subsurface. You have to do something with the CO2. You could turn it into industrial-grade CO2 for beverages, but a single 100-megawatt power plant with carbon capture would satisfy half the global beverage industry's carbonation needs. So the only real option is permanent underground storage, which means you need to understand the subsurface. You're effectively combining three maps, natural gas infrastructure, the power generation and transmission layer with data center demand, and the subsurface layer for sequestration. In an ideal world the bright spots on all three overlap. You can't do this everywhere. New England has gas pipelines and data center demand but no geology to store the CO2, so you'd need a pipeline from Boston to Texas, which would be cost-prohibitive. West Texas is one of the best places. There, CO2 is used in oil fields for enhanced oil recovery, helping produce oil that wouldn't otherwise be recovered while permanently sequestering the CO2 captured from the power plant. CO2 goes back underground forever and incremental oil is coaxed out of the existing well, displacing other oil in the global marketplace.
The very interesting thing is that even if you can't do this everywhere, the fact that you can do it in West Texas and across the Permian means this is scalable across the biggest shale reservoirs in the US. What's missing for this to scale, and what's the main challenge in front of you?
The nice part is none of these challenges are unsolvable, and it's not a technology challenge. The PCC amine-based technology has been proven over decades in other industrial applications, it's just never been used in a power-first application. That's the beauty of what the Entropy folks have done, they've taken that amine-based solvent and applied it to natural gas power generation. Natural gas is harder because the flue gas is predominantly oxygen and nitrogen and only 5 to 6 percent CO2, whereas an ethanol plant might be 50 or 60 percent. The more efficient the gas plant, the smaller the CO2 percentage and the harder it is to capture. Entropy has demonstrated over several years that their technology works on natural gas, which is excellent. So technology isn't the constraint.
The biggest challenge is sentiment among potential power buyers around natural gas and oil. A couple of years ago everybody was pledging to be off natural gas and oil and to be net zero. Today there's been a paradigm shift, and everybody's saying natural gas is absolutely critical for power generation. People are embracing gas again, but many of those same folks say they don't want to perpetuate the production of oil because of environmental pledges. Some of that is dogmatic rather than grounded in how the world works. Everybody is now starting to understand that fossil fuels are foundational to modern civilization, both for AI and for quality of life. There's a strong correlation between quality of life and energy consumption per capita. As people move from lower to middle class their energy use rises sharply. The amount of energy most people in the developing world consume in a year is roughly what a single American refrigerator uses. We don't have the right to deprive them of the same quality of life we have, but we do have the opportunity to produce the energy that enables it.
A byproduct of what we're doing on the power generation side is incremental oil production from existing wells, and there's real social benefit in that. If we can also decarbonize the grid with post-combustion capture, that's a good outcome for everybody. Data centers get the power they need, scope 2 emissions go down versus unabated gas, and as a byproduct we get greater domestic energy security and more oil to help our allies. Oil isn't just gasoline for vehicles. Oil is plastics. Oil is fertilizer. Oil is one of the pillars of modern civilization. Part of our challenge is breaking through and advocating for a sensible energy ecosystem. The US has so much natural gas that catalyzes all of this. We just need the right regulatory framework and the right commercial partnerships. Our focus now is getting the first project online by the end of this decade, which would make it the first clean gas power plant online in the US in a long time, and the first natural gas PCC facility online ever. We have the buyer of the CO2 lined up with Oxy, the missing link is the power buyer, and we expect to have strategic partnerships formed on that side by the end of the year. I'm really encouraged about what the future of energy looks like in the United States, and the US can be the poster child for how to do this right and help our allies do it the same way this decade.
