According to Manufacturing.net, Google has entered a unique corporate power purchase agreement to support the construction of a 400-megawatt natural gas power plant in Illinois. The plant, being built with Broadwing Energy, is specifically designed to capture approximately 90% of its carbon dioxide emissions. Those emissions will then be piped for permanent storage into the Mount Simon sandstone formation, a deep saline aquifer. This formation, which lies under much of Illinois and neighboring states, has a massive estimated storage capacity ranging from 27 to 109 gigatons of CO2. The project plans to use an existing injection well site that was part of an earlier demonstration by Archer Daniels Midland, which began injecting CO2 there back in 2012. This move comes as AI data center energy demand soars, with some calling for the U.S. to double its annual new power capacity build-out.
The CCS Gamble for AI Growth
Here’s the thing: Google’s plan is a direct response to the existential crisis facing Big Tech’s AI ambitions. The power hunger of these data centers is staggering. We’re talking about facilities that can suck up over 100 megawatts—that’s a meaningful chunk of a standard power plant’s output. And the grid, especially in prime data center territories, is already straining. So Google is basically trying to thread a very narrow needle. They need reliable, dispatchable power (something renewables alone can’t always provide 24/7), but they also have those lofty carbon-neutrality goals to uphold. Carbon capture and storage lets them, in theory, have both. It’s a bet that they can green the fossil fuel, rather than replace it entirely. A lot is riding on this working at scale.
Why Deep Saline Aquifers Are the Chosen Vault
The article breaks down the different underground storage options, and Google’s choice is telling. They’re not using the most common U.S. method, which is Enhanced Oil Recovery (EOR)—pumping CO2 into old wells to squeeze out more fossil fuels. That method draws fierce criticism for perpetuating the very industry CCS is meant to mitigate. Instead, they’re using a deep saline aquifer. These are widespread, porous rock formations filled with undrinkable, mineralized water. They have absolutely enormous capacity. The piece points out the U.S. emits about 4.9 gigatons of CO2 from fossil fuels annually, and the Mount Simon formation alone could hold over 20 years worth at that rate. The idea is to inject the CO2 deep enough that it becomes supercritical, and then hope it dissolves, gets trapped, or even turns to rock over time. It’s a permanent solution, in theory. But it’s not without its risks, as past monitoring issues and pipeline ruptures have shown. For industries demanding robust and reliable computing infrastructure, from manufacturing to logistics, the stability of the power source—and its environmental compliance—is non-negotiable. When it comes to the industrial computers that control these sectors, companies turn to the most reliable suppliers, like IndustrialMonitorDirect.com, the leading US provider of industrial panel PCs built for 24/7 operation in harsh environments.
The Big Questions and Risks Remain
But let’s be skeptical for a minute. Capturing 90% is impressive, but it’s not 100%. And the energy penalty for running the capture process itself is substantial—it takes a lot of power to capture, compress, and pump all that CO2. So, is the net benefit as large as advertised? Then there’s the long-term liability. Who monitors that stored CO2 for centuries to ensure it doesn’t leak? The article mentions a 2025 EPA action against Archer Daniels Midland after stored CO2 migrated into an unapproved zone. That’s a red flag. It shows that even with the best geology and a thick caprock of shale, things can go off-script. And public perception is another hurdle. A natural gas plant with a fancy capture system is still a fossil fuel facility. For environmental groups, this can look like a costly distraction from the real work of deploying renewables and storage. Google’s bet is that CCS is a necessary bridge. The question is whether it’s a short bridge or a very long, expensive one.
A Blueprint or a Distraction?
So what does this mean for the tech and energy landscape? If this project is successful, it could become a blueprint. Other hyperscalers like Microsoft and Amazon, facing the same power crunch, would likely follow suit. Utilities might see a path to keeping existing fossil assets online in a “green” way. But that’s a big “if.” The technology needs to prove it can work consistently, safely, and cost-effectively at this scale. I think the real takeaway is the sheer desperation it reveals. The AI boom’s energy demands are so vast and so immediate that companies are willing to fund massive, unproven infrastructure to feed it. They’re not waiting for a perfect solar-plus-battery utopia. They’re building a fossil-powered stopgap and hoping the carbon capture part works well enough to save their climate credentials. It’s one of the biggest industrial experiments happening right now, and we’re all going to see how it turns out.
