This Week in Cleantech is a weekly podcast covering the most impactful stories in clean energy and climate featuring Paul Gerke of Factor This and Tigercomm’s Mike Casey. Jeremiah Karpowicz, Content Director at Clarion Events, is filling in for Paul this week.
This week’s episode features special guest Molly Taft from WIRED, who discusses the growing, cross-partisan opposition to AI data centers.
This week’s “Cleantecher of the Week” is Drew Baglino, founder and CEO of Heron Power. His company builds solid-state transformers that don’t rely on the steel causing shortages, and Heron just closed a $140 million round to build a U.S. factory capable of producing 40 gigawatts of the gear a year. Congratulations, Drew!
1. OpenAI Locks In Lease for Huge Data Center in Ohio With Backing From Nvidia — The Wall Street Journal
OpenAI has signed a 20-year, 10-gigawatt data center lease in southern Ohio with SoftBank’s SB Energy, the largest data center project announced to date. Nvidia is backstopping the value of the completed facility, to the tune of up to $105 billion, so SB Energy can borrow against the project without lenders worrying about what happens if OpenAI ever walks away.
That’s a much smaller ask than the roughly $250 billion all-up guarantee Nvidia was reportedly considering earlier this summer, before investor pushback over the exposure sent its stock down 5% in late July. In exchange, Nvidia becomes the exclusive chip supplier for that first half of the campus and is taking a $1.5 billion equity stake in SB Energy.
2. The Blackout That Could Devastate America — The New York Times
Modeling by FERC found that taking out just nine critical substations nationwide could black out the entire country. Large power transformers are truck-sized units that step down high-voltage electricity for local use, and are mostly bespoke and hand-built. They can take up to five years to source and cost as much as $14 million each. The U.S. only manufactures about 20% of what it buys.
A significant portion of the country’s large power transformers were already averaging 38-40 years old over a decade ago, meaning many are now well past that. Replacing this aging fleet and powering new data centers could double the demand for large power transformers by 2027.
If a coordinated attack knocked out enough substations, experts say the recovery could take years, not weeks, since replacement transformers have to be custom-built and are too heavy to move without specialized rail cars that barely exist in North America anymore.
3. Big Batteries to Bolster the Grid Are Stuck Waiting to Plug In — Bloomberg
Battery demand is surging as data centers and electrification push up electricity use, but connecting those batteries to the grid is getting slower. Nationwide, roughly 750 gigawatts of storage projects are sitting in interconnection queues waiting to get hooked up, and the median wait time has jumped from 1.5 years in 2015 to 5 years in 2025.
The holdup is mostly equipment, meaning transformers, circuit breakers, and other grid gear that are in such short supply that utilities can’t build the upgrades fast enough. PG&E says some circuit breakers now take nearly four years to procure.
4. DOE says we need more transmission, decides building less is fine — Factor This
The DOE is killing off the review process for the last three proposed National Interest Electric Transmission Corridors, or NIETCs. These are DOE-designated regions that are eligible for loans and permitting tools intended to speed up transmission development in high-congestion parts of the grid. All seven of the others had already been scrapped back in 2024.
These three corridors would have spanned more than 3,500 miles, connecting Canada to the PJM grid, linking Southwest Power Pool with WestConnect, and expanding tribal energy access. Secretary Chris Wright called the review process ineffective at improving reliability or cutting costs. The DOE points to other transmission investments instead, including $1.6 billion and $3.3 billion in loans to utility AEP and its Texas subsidiary, along with over $26 billion for Georgia Power and Alabama Power, subsidiaries of utility Southern Company.
5. How Data Centers Broke American Politics — WIRED
A growing, cross-partisan backlash against AI data centers is uniting right-wing populists and left-leaning environmentalists who otherwise agree on almost nothing. A recent Gallup poll found 70% of Americans would oppose data center construction in their area, and Facebook groups opposing data centers grew from under 100,000 members in December to over 500,000 by July.
The backlash has already led to political fallout, including data center moratoriums introduced in a dozen states and passed in New York, an attack on a pro-data-center Indianapolis council member, and protester arrests in multiple states.
Facts Only
* OpenAI signed a 20-year, 10-gigawatt data center lease in southern Ohio with SB Energy.
* Nvidia provided a value backstop of up to $105 billion for the Ohio facility.
* Nvidia took a $1.5 billion equity stake in SB Energy and became the exclusive chip supplier for the first half of the campus.
* FERC modeling indicates the loss of nine critical substations could cause a nationwide blackout.
* The U.S. manufactures approximately 20% of its large power transformers.
* 750 gigawatts of battery storage projects are currently in interconnection queues.
* The median wait time for grid interconnection increased from 1.5 years in 2015 to 5 years in 2025.
* The DOE terminated the review process for the final three National Interest Electric Transmission Corridors (NIETCs).
* Heron Power closed a $140 million funding round to build a U.S. factory for solid-state transformers.
* A Gallup poll indicates 70% of Americans oppose data center construction in their local area.
Executive Summary
The intersection of artificial intelligence expansion and aging electrical infrastructure has created a systemic bottleneck in the United States. Massive projects, such as OpenAI’s 10-gigawatt campus in Ohio supported by Nvidia, are driving an unprecedented surge in electricity demand. However, the physical grid is struggling to keep pace; critical components like large power transformers are plagued by long lead times, high costs, and a heavy reliance on foreign manufacturing. This hardware shortage is delaying the integration of battery storage and increasing the vulnerability of the national grid to catastrophic failure.
Simultaneously, the regulatory and social landscape is shifting. While the DOE has pivoted away from specific designated transmission corridors in favor of direct utility loans, a cross-partisan political backlash is emerging. This opposition, uniting environmentalists and populists, has led to data center moratoriums in several states. The situation presents a tension between the rapid scaling of AI capabilities and the physical and political realities of energy distribution and land use.
Full Take
The strongest version of this narrative highlights a critical "physicality gap": the collision between the ethereal speed of AI software scaling and the stubborn, slow-moving reality of heavy industrial hardware. It correctly identifies that the "cloud" is actually made of steel and silicon, and that this physical foundation is currently fragile.
The narrative employs a subtle Fear Appeal by juxtaposing the "devastating" potential of a nine-substation blackout with the aggressive expansion of AI data centers. This creates a causal link in the reader's mind—suggesting that AI's hunger for power is not just a resource challenge, but a national security liability—without explicitly stating that AI is the direct cause of the vulnerability.
Patterns detected: ARC-0014 Fear Appeal
The driving paradigm is one of resource scarcity and infrastructure decay. It assumes that the current trajectory of AI growth is inevitable and that the only variables are how we build the hardware and how we manage the public's anger. This echoes historical patterns of industrialization where technological leaps outpaced the regulatory and social frameworks intended to manage them.
The second-order consequence is a potential "infrastructure aristocracy," where only the largest players (Nvidia, OpenAI, SoftBank) have the capital to backstop the risks of grid expansion, effectively privatizing the stability of public utilities.
Bridge Questions:
1. If solid-state transformers can bypass steel shortages, does the bottleneck shift from material scarcity to regulatory permitting?
2. Is the cross-partisan opposition to data centers based on energy concerns, or is it a proxy for a broader anxiety regarding AI's societal impact?
Counterstrike Scan: An influence campaign would weaponize the "blackout" fear to lobby for specific protectionist trade policies or to sabotage AI competitors by amplifying local "NIMBY" protests. The current content remains a synthesis of reporting and does not match this tactical pattern.