Google’s newest energy partnership reflects a shift in how the company plans to power its fast-expanding AI infrastructure. The tech giant has agreed to support the Broadwing Energy Center, a 400-megawatt natural-gas plant in Decatur, Illinois that will use carbon capture and storage (CCS) to trap roughly 90% of its emissions. For a company that has spent years touting wind and solar, the change is notable. The move highlights both the surging energy appetite of artificial intelligence and the political trends influencing clean energy in the U.S.
The Broadwing project offers a mix of old and new energy strategy. Natural gas provides the dependable baseload power that hyperscale data centers increasingly require. CCS technology, still controversial and uneven in execution, is intended to prevent that energy from pushing Google’s emissions higher. If it works as planned, Google would secure a steady supply of low-carbon electricity for its Midwest data centers without relying on a power grid already strained by record demand.
Urgent Demand, Major Critics
For Google, which expects its global electricity use to climb sharply this decade, the urgency is real. Generative AI has transformed the company’s infrastructure needs. A hyperscale data center can demand well over 100 megawatts of power, or roughly the output of a small utility-scale power plant, just to keep the GPUs humming and the cooling systems running. When multiplied across Google’s fleet of U.S. facilities, the power requirements are staggering.
Under the agreement, Google will purchase most of Broadwing’s output once operations begin in the early 2030s. The construction effort alone is expected to create 650 union jobs, along with 100 management roles.
Still, the project raises fundamental questions. CCS has a track record marked by stalled facilities and ballooning project costs. A 2021 U.S. Government Accountability Office report found that of six federally backed CCS projects at coal plants, five never reached operation. Even successful efforts have required creative economics, often involving the sale of captured carbon to oil producers for enhanced recovery, an approach Broadwing is not pursuing.
The Illinois project plans instead to inject CO₂ deep into a saline aquifer, a geological formation considered well suited for long-term storage. Such reservoirs span much of the Midwest and have the advantage of high capacity. But the process demands precision and continuous monitoring. If CCS is to become a meaningful tool for decarbonization, projects like Broadwing will have to show that the science, the economics, and the regulatory oversight can align at scale.
Critics argue the project still locks in dependence on fossil fuels. Natural gas leaks methane, a far more potent greenhouse gas than CO₂, across wells and pipelines. Capturing carbon at the smokestack doesn’t mitigate that upstream risk. Others note that solar and wind remain cheaper than gas paired with CCS and avoids the health impacts of gas-plant emissions on nearby communities.
The Politics of Energy
The political environment plays a role in this shift. With a federal government that has pulled back support for renewables and left CCS tax incentives intact, corporate clean-energy playbooks are changing. Google has long been one of the country’s largest buyers of wind and solar power. The company’s expansion into geothermal, hydropower, and now carbon-capturing natural gas suggests a strategy shaped as much by necessity as ambition.
AI is driving that necessity. Hyperscalers are seeking gigawatts of new electricity, which is a demand level not seen since the industrial buildouts of the 20th century. Against that backdrop, Google’s carbon-capture experiment may be less an endorsement of CCS than an acknowledgement that the clean-energy landscape is shifting, and that reliable, low-carbon power—whatever the source—has become a strategic constraint.
