The subtitle to Ted Gioia’s recent post, “Anti-AI Laws Now Get Unanimous Support from Left & Right,” in his Substack, The Honest Broker, pretty much says it all in his subtitle: “In just three years, Silicon Valley has destroyed its entire support network. It’s still in denial—but that won’t last.” The major AI companies that have become synonymous with Silicon Valley have indeed managed to frighten and/or annoy huge swaths of the public both here and abroad.

As Gioia writes:

AI backlash is a huge national issue, but it’s also the hottest local issue right now. Across the US, more than 300 bills restricting data centers have been introduced in states and communities. At least 14 states are considering total moratoriums. And in places where politicians hesitate, the public is stepping in—putting voter initiatives on the ballot.

Silicon Valley has totally lost public support, and now will get punished brutally with legislation. But litigation will be just as painful. Tech companies are now routinely losing jury trials—they are so disliked by the public that going to trial almost always ensures a loss.

Any AI babies being thrown out with the bath water can be blamed on the AI would-be tech lords, but let’s also admit there’s a level of Anti-AI hysteria running through the populace that is, as hysteria tends to be, more triggered response than substantive understanding.

First, a Moment of Housekeeping

Ted Gioia’s The Honest Broker is a refreshing Substack, partly because of the quality of the writing and partly because of his wide-ranging purview focused, as it were, on the environments and practices of human creativity. How much do I like this Substack? Well, here in America we say “I love you” with money, and I’m a paying subscriber.

There’s another The Honest Broker Substack, written by Roger Pielke Jr. and I’d guess this one is better known in the climate change circles that I closely follow. Pielke is smart, but he sometimes irks me in his criticisms of climate science, where—spoiler alert!—sometimes someone gets something wrong. Pielke is especially on a crusade about imperfections in climate research, and this sometimes gives aid and succor to climate deniers and doubters, even though Pielke is not trying to do this; he’s trying to correct the record of what we actually know about climate change and how we know it. He’s also easily upset with the efforts and assumptions of many about climate change attribution, and he has a point that extreme weather happens and any specific instance of extreme weather cannot be definitively attributed to changes in climate. But in the grand scheme this is a minor point, since the evidence, to raise just one example, of the tie between climate change and worsening extreme weather and rates of extreme weather is in aggregate undeniable. Tomorrow’s massive deluge? Maybe, maybe not, but that warmer air holds more moisture and thus, with rising global temperature averages, more moisture means more deluges, that’s not a “maybe.” That’s simple enough to grasp even without a PhD in oceanic and atmospheric sciences. I wish he’d spend less time proving that he’s right and others are wrong, especially on particular points that the larger perspectives make moot. I’ve written about Roger Pielke Jr. a number of times. The following posts will be more than enough to satisfy your curiosity: “Let’s Not Worry So Much About ‘Climate Thought Police’,” from November 19, 2025, and “Statistics, Damned Statistics, and Common Sense, from July 3, 2024.

Now, back to my regularly scheduled programing.

Hubris, Tone-Deafness, Secrecy Dreams, and How to Foster Hysteria

You gotta admire the AI tech overlords, and the public electricity and gas utilities, and the fossil fuel industry: their intense focus on short-term gains blinds them to achieving their long-term success—or in the case of Big Oil, the long-term success of the Earth as a lovely place for human habitation.

The AI Tech Overlords

In the case of the AI tech overlords, their lust for massive and plentiful data centers has driven a market message that AI needs more and more and faster and faster. It doesn’t take a genius to see gaping holes in their wishes. For one thing, the general AI platforms of today still do not demonstrate successful business models. The primary argument is that AI can theoretically (or may someday) replace many humans who require salaries and paychecks, thus fulfilling the wet dreams of capitalists with huge cost savings. No one should be surprised that the vast majority of those who work aren’t thrilled by this vision. The AI dream, and especially the human-replacement part, is so fervid that AI corporations pursue staggering market valuations, while running huge development costs mentioned over and over again by the very same AI corporations as they seek to raise more and more capital. The sums chased are so large that they are regularly in the news and to such an extent that economic bubbles are also regularly in the news. Also in the news? The question of when such a bubble may burst and bring down the economy due to the huge sums of money being tossed at AI.

It is argued by some impressive people—I’d point to Tony Fadell in a recent Podbean podcast episode, “Breakthrough Technology  Dialogues with Tony Fadell,” appearing on Wednesday, April 15, 2026—that the current fixation on massive data centers is a problem. If you don’t know Fadell, look him up, but here’s the CliffsNotes: Godfather of Apple’s iPod; Co-creator of the iPhone; Co-Founder of Nest (The Learning Thermostat), which sold to Google for $3.2 billion, give or take some millions; perennial possible Apple CEO candidate (although never the bride); and as of 2017, Fadell has been running a venture fund originally called Future Shape, now called Build Collective. [Full disclosure: I worked for Tony at Fuse, a 1998 start-up aimed at network-connected consumer electronics that went away when the venture capitalists vanished during Dot.Com, when shortly he was off to Apple and the iPod. Years later, my daughter started with Tony at Nest as Assistant to the Founders, leaving the company sometime post-Google purchase, having become a product manager for the Nest thermostat.]

Podbean has Fadell talking about AI “on the edge,” by which he means, as best I can follow, that with powerful processing in our own devices, much of what we want from AI will be handled less in data centers and more in what is in your pocket or your corporate server. This may be related to the I/O challenge of data center AI, or I may be conflating the podcast here with some article or other read recently, but if you’re coming to me for analysis of computing technology and practice, stop reading now. The main point I’m making is that there are some very real questions about how AI will present in the world, but you wouldn’t know this if all you go by is the current AI corporations’ frenzy to establish huge fleets of data centers.

The Public Power Utilities

While the news from the kingdom of AI is all about fierce expansion of data centers to run all those servers stacked with  Nvidia chips so that the AI companies can charge for usage, all the while requiring significant electricity supply and water supply for cooling. For public utilities—both electricity and natural gas—all this represents gobsmacking opportunities. Not that power utilities can deliver the power in a sufficient or timely way, at least not with the way utilities have been deliveing electricity  for decades upon decades. Today, power utilities face grids constraints, electricity capacity hard limits, interconnect planning delays, and supply chain problems, but that doesn’t stop the utilities from dreaming. Of course, what the utilities really want is to build more generators, more distribution and transmission lines, and more other capital-intensive infrastructure to meet the new demand, because the capital spending they undertake gets them a guaranteed return on investment. This model is still today’s bread-and-butter profit mechanism for these public monopolies, especially investor-owned utilities.

By the way, there are plenty of well-understood and “shovel-ready” ways to increase actual electricity capacity that are less capital intensive, mainly through grid and supply digital management to extract huge gains in efficiency and flexibility. These include smart meters, smart panels, smart appliances, or the various ways to gradually incorporate distributed energy recourses (e.g., community and utility-scale solar, wind, and batteries systems, electric vehicles as batteries, etc.). There are also better ways to manage grid responsiveness and stability via grid-facing inverters, digital transformers, utility-level real-time sensors, and approaches like virtual power plants, where home solar/batteries are aggregated to serve as electricity generation for the grid for the benefit of contributing households and for improved overall grid supply. If you really want to learn about all this stuff and the lion’s share I haven’t mentioned, check out David Roberts’s Volts.

The point here is that there are lots of ways we can improve and expand the grid without huge spending, but the current hype about AI fits nicely into the way utilities prefer to make money today.

Big Oil is Hoping AI Goes as Big as Possible and to Hell with the Climate

Back on January 29, 2026, I posted “AI is Giving Me Gas.” The subtitle: “Behold the wonder of climate denial in the planned expansion of new gas generation plants.” The opening paragraph follows:

I’ve been saying that AI’s projected electricity demand is celebrated by fossil fuel companies because this growth in electricity demand provides an anchor for Big Oil to keep selling natural gas for decades to come.

Big AI is in bed with Big Oil and a glance at the headlines just about any day shows this clearly. Well, “clearly” may not be the best descriptor, since Musk and Zuckerberg and who knows who else seem happy enough to get their data centers rolling without waiting for the grid to supply the needed juice. There are lawsuits and entirely reasonable climate justice fights about data centers polluting neighborhoods through the use of fleets of big diesel-fueled generators and natural gas generators, whether for primary electricity source or for backup.

Google’s AI Overview for the query “Court cases involving diesel fueled generators for data centers” answers this:

Court cases and legal battles involving diesel generators at data centers primarily revolve around federal environmental xAI violations, local permitting disputes over emissions, and litigation brought by residents and environmental groups regarding public health. [1, 2, 3, 4, 5]

Notable Cases and Legal Disputes

• • xAI Colossus 2 Power Plant Lawsuit (Southaven, Mississippi): The Southern Environmental Law Center (SELC) and Earthjustice filed a major lawsuit against xAI, challenging the company’s use of unpermitted gas-turbine power plants to run its data center. The coalition (representing the NAACP) argued xAI bypassed federal Clean Air Act requirements and created a de facto power plant that emits smog, soot, and formaldehyde without proper public notice or health mitigation.
  • Markley Data Center Expansion Lawsuits (Lowell, Massachusetts): Residents and environmental justice advocates launched the first data center lawsuit in Massachusetts, challenging a state permit that allows a data center expansion to operate backup diesel generators. Filed with the help of the Yale Law School Environmental Justice Clinic, the suit charges the state with ignoring carcinogenic diesel exhaust and failing to conduct cumulative health impact analyses in an Environmental Justice population.
  • Independence AI Data Center Contract Battles (Independence, Missouri): A sprawling legal battle involves a $6.6 billion AI data center proposed on a 400-acre site. Neighbors and local groups are opposing the development—which is slated to rely on extensive diesel generation—bringing zoning, environmental, and development contract battles into circuit and county courts.
  • Quantum Loophole / Aligned Data Center Exemption Denials (Maryland): State regulators and local councils denied Aligned Data Centers an exemption that would have allowed them to operate 168 diesel backup generators at a massive data center campus in Frederick County, Maryland, prompting significant development and zoning pushback. [1, 2, 3, 4, 5, 6]

Regulatory Pushback and State Precedents

Beyond direct litigation, communities are actively fighting permitting processes and appealing to environmental boards regarding diesel use:

• • Port Washington (Wisconsin): Residents petitioned the Wisconsin Department of Natural Resources (DNR) to deny air quality permits for 45 backup diesel generators, citing concerns about localized particulate matter and nitrogen oxide emissions.
  • Hilliard (Ohio): The Ohio EPA faced pushback and held hearings over draft air permits allowing an Amazon data center to run 158 Tier 2 diesel generators. Residents and local officials have called for cleaner Tier 4 standards to limit carcinogenic emissions.
  • Bay Area (California): Environmental groups petitioned the Bay Area Air Quality Management District to tighten regulations, arguing that decades-old emergency rules for diesel generators are inadequate for the massive scale of modern hyperscale data center operations. [1, 2, 3, 4, 5, 6]

Resources for Tracking Ongoing Cases

You can follow these developments and access case documents through environmental advocacy and legal tracking sites: [1]

• • Southern Environmental Law Center: Provides ongoing litigation updates, news releases, and filings regarding the xAI turbine case in Mississippi.
  • Yale Law School News: Offers details and court documents concerning the Lowell, Massachusetts environmental justice lawsuit.

There’s also a trend toward behind-the-meter AI data centers that aren’t connected to the grid at all, like some of the examples listed above. One worry is that gas-fired generators will get built to support the data centers by producing private electricity without having to wait for the public grid to develop the means to supply the power and/or without raising the utility’s rate for existing customers. The issue of “affordability” is also as a big play for behind-the-meter approaches. The affordability issue, however, doesn’t really get addressed by the behind-the-meter approach, although it sure sounds good, doesn’t it?

Keep in mind that while behind-the-meter power supply doesn’t have to be gas-fired and could be solar, wind, and/or batteries, there are a hell of a lot more plans in the works today for fossil fuel-based behind-the-meter generation. Here’s the AI Overview to the query “With behind the meter data center electricity, how many being planned use fossil fuels and how many use renewable energy?”:

Recent analysis of behind-the-meter (off-grid) data center developments indicates a heavy dominance of fossil fuel generation in immediate construction plans, despite widespread corporate sustainability pledges. [1]

Fossil Fuels vs. Renewables in Behind-the-Meter Plans

• • Fossil Fuel Dominance (~75%): An analysis of permit documents and equipment deals for upcoming behind-the-meter projects identified approximately 23 gigawatts (GW) of generation capacity. Of this identified equipment, nearly 75% is powered by natural gas.
  • Renewable “Exceptions”: Genuine renewables-first behind-the-meter projects are currently rare exceptions rather than the rule. The most prominent example is Switch’s Citadel Campus in Nevada, which utilizes 127 MW of solar paired with 240 MWh of battery storage.
  • “Gas Now, Clean Later” Strategy: While many developers publicize “all-of-the-above” strategies involving nuclear or renewables, the actual infrastructure being installed for 2025 and 2026 operations is almost entirely gas-fired. Commitments for renewable capacity are typically scheduled for 2028 or later, with nuclear options often projected as being a decade away. [1, 2]

Specific Behind-the-Meter Fossil Fuel Projects

Several massive gas-powered projects have been announced or are in development to bypass grid connection delays:

• • Texas (Permian Basin): FO Permian Partners and Hivolt Energy announced a 5 GW off-grid natural gas power solution for data centers.
  • Wyoming: Prometheus Hyperscale and Engie are planning a 2 GW facility with onsite gas-fired generation.
  • New Mexico: Oracle and OpenAI are developing the “Stargate” project, which is expected to rely on massive natural gas systems for off-grid power.
  • Pennsylvania: International Electric Power is planning a 944 MW gas plant specifically to power a data center while avoiding the PJM grid interconnection but not convenient for the rest of us, though. queue. [1, 2]

    

Well, isn’t that convenient? And do you think such fossil fuel-based generation won’t run on for decades, even if in the long run renewables will prove far cheaper? But then, of course, these data centers will have built the gas generators and won’t likely wish to build a whole new renewable generation plant. The whole data center power supply scheme seems custom built for Big Oil to sell more and more fossil fuel for more and more decades.

Maybe Public Hysteria about Data Centers Has a Point

People are right to worry. I was on a Zoom call with others in a climate action group and the prospect of adding a new project to organize a statewide AI ban was on the agenda. Not only are increased pollution and higher greenhouse gas emissions and global warming likely consequences if most of the planned the behind-the-meter electricity supply for data centers go forward, but there are real concerns for water resources becoming compromised and communities made worse, including with noise pollution.

But there are other economic costs for the public that put lie to the behind-the-meter “affordability” issue. Our current grid—in most places, anyway—is under-used. By adding digital intelligence (oh, the seeming irony!) and distributed energy resources to the grid, there are more users of the public grid when data centers get their electricity from it. More users mean more widely distributed costs carried over more parties, which lowers each user’s own bill.

Weird, huh?

The more data centers opt out of participating in the grid, the less efficiently the grid is used, bringing higher grid and generation costs to be paid by fewer customers. If there is a legislative program, perhaps banning data centers outright would miss an opportunity to have them located fairly and properly, accounting for projected water use by developing additional water resources, or even tying data center cooling demands to thermal heating and storage networks to reduce heating loads in cold climates. Data centers could also be part of establishing more clean energy, which is cheaper to build these days and far cheaper to run over its expected lifecycle compared to a fossil fuel generation plant. Data centers can also contribute to advances in and implementation of grid efficiency and digital management of loads and demands, reliability, and grid balance and voltage regulation.

Of course, the AI tech lords will tell you that data centers must be built today and up and running tomorrow. Since AI is racking up tremendous investment liabilities with every new dollar added to the cause, I can see why they’re impatient to shake a leg. I’m less clear about why we need to put ourselves on their schedule, especially with all the existentially negative aspects of rushing.

Weird, huh?

Jerks R Us Interuptus

What may be the biggest prod to the widespread public AI backlash, as Gioia writes, is the AI tech leaders’ tone-deafness and astounding isolation from the real world. “These folks are living in an echo chamber—and they reveal their disconnect with the public almost every time they open their mouths,” he says. Commencement speakers mentioning AI are literally getting laughed out of the room.

Gioia continues:

In an odd twist, it’s actually getting harder for politicians to run on this issue—because their opponents are running on the same issue. In 2026, only oblivious tech bros are willing to stand up and praise AI. And when that happens, the booing begins instantly.

Let me put this in economic terms. Tech companies have spent more than a trillion dollars on AI and have only succeeded in becoming the most hated businesses in the world. Spending another trillion dollars—or (most likely) more—will only intensify the public’s antipathy.

It’s easy to see where this is leading.

History teaches us that you can only push people so far. Most of the time, they let things slide. But there are limits—and Silicon Valley has just slammed into them. People aren’t going to sit by silently while tech companies eliminate their jobs, hurt their kids, slopify the culture, and use up scarce resources.

They will fight back—and that’s the phase we’ve just entered in 2026.

Sooner or later, even the most clueless tech CEOs will realize that they need to slow down and address public concerns. My best guess is that we’re still 12-18 months away from that. But that’s only because Silicon Valley is so out-of-touch with the rest of world—they block out the boos and live in their own fantasy world.

Amen, brother.

Of course, it may turn out that in 12-18 months our understanding of AI and how best to platform it may present another world the tech bros can ignore for a while more. Or perhaps in 12-18 months’ time, today’s staggeringly large AI investments will be past due and the dear AI tech lords will be busy dealing with market corrections of their own making.

Meanwhile, and from here on out, let’s give a hoot and don’t pollute, and I don’t just mean the carbon emissions factor but also our very culture with AI slop threatening to become a self-reinforcing GI/GO problem, but don’t worry: the Tech AI lords are paying some people $20 an hour to keep this from happening, so relax.

Or maybe stay tense. The obtuseness of AI tech lords, the old lazy habits of power utilities, and the never-ceasing greed of Big Oil needs to be countered less with hysteria and more with a clearer and cleaner sense of how to build our world’s better future.

Can I have an Amen?

The post Let’s Do Anti-AI Backlash Right first appeared on David Guenette.

This is the first paragraph of one of my posts from earlier in the year, “AI is Giving Me Gas: The Collision of Tech Hype and the Carbon Budget.” It may be bad form to start a Substack post citing another Substack post, but clearly these two posts are related. The sub-title of the above referenced post” “We are scraping the bottom of the 1.5°C carbon budget. Big Oil’s response? Build 252 gigawatts of new gas power to feed the AI boom.”

I also wrote about this back in September of last year, in a post titled “New Gas Generator Plants and the Plan to Flood the (Electricity Demand Growth) Zone.” In this post there’s a link to an AI analysis I did in a report called “The Future of U.S. Natural Gas Power Generation: Projections, Accuracy, and the Confluence of Limiting Factors to 2030.”

The electricity demand growth tied to AI and data centers has Big Oil salivating, with plans—dreams?—of 100-plus new gas-fired generation plants in place by 2030. Not that the supply chain and turbine manufacturing capacity can deliver, but the explosion in small diesel or natural gas generators nonetheless seems a happy enough ending, boding well for Big Oil sales.

This doesn’t bode well for the rest of us, unfortunately. In the news of late is Elon Musk’s xAI company, which has used a large fleet of mobile, trailer-mounted gas turbines (rather than diesel generators) to power its “Colossus” AI data center in Memphis, Tennessee. These turbines are deployed as a temporary, “quick and dirty” solution to bypass power grid constraints while constructing the facility, which houses Nvidia H100 GPUs for training the Grok AI model.

To get the data center operational in just 122 days, xAI used mobile turbines (approximately 35 to 62, depending on the report and timeline). Each turbine is capable of providing 2.5 MW of power, with reports indicating a total capacity exceeding 35 MW to over 100 MW. There’s been pushback from local residents and environmental group. It turns out that neither noise pollution or emissions of nitrogen oxides and formaldehyde are being welcomed, and in January 2026, the EPA ruled that xAI violated the law by operating dozens of these, at times, unpermitted, gas generators. I’m guessing any fines actually levied against xAI will be just part of the cost of doing business. The broader AI data center industry is facing a shortage of power, with many companies increasingly using on-site, fossil-fuel-based generators to bridge the gap.

By the way, if you want to dive deeper into the issue of power strategies and developments for AI data centers, check out Distilled, on Substack. Distilled is written by Michael Thomas and he’s undertaken as series of articles on the issue of AI data centers and approaches being pursued for power, including data centers building their own power plants.

Wet Dreams and Sloppy Seconds

The accuracy of future demand predictions itself is highly questionable, considering the wide-ranging numbers and, among other issues, the double, triple, or greater duplicate counting of generation sources among data center hyperscalers. These eager corporations reach out to more than one potential generation source to cover their bets. Accurate forecasting seems hindered by some combination of wishful thinking and double counting. Here’s an AI summary of this issue:

The “duplication issue” (often called “phantom load” or “speculative queuing”) refers to the practice where data center developers submit multiple applications for electrical grid interconnection for the same single project. 

Because securing power is now the primary bottleneck for AI and hyperscale facilities, developers “spam” the queue to hedge their bets. They might file requests for the same 500 MW project in three different states (or three different sites within the same utility territory) to see which one gets approved first. Once one is approved, the others are withdrawn, but in the meantime, they clog the study queue and artificially inflate demand forecasts. 

Grid operators warn that these “phantom” requests make it impossible to accurately plan for new power plants, as the requested demand on paper is often 5x to 10x higher than what will actually be built. 

Estimate Ranges of Duplication

Industry data suggests that the vast majority of current interconnection requests are duplicate or speculative. 

• • Overall “Phantom” Rate: Experts estimate that 80% to 90% of the data center capacity currently in US interconnection queues will never be built.
  • Realization Rate: Utilities often project that only 10% to 20% of the requested data center load in their pipelines will actually materialize.
  • Speculative Ratios (Firm vs. Requested):
    • AEP (American Electric Power):Reported 24 GW of firm commitments but has requests for 190 GW of additional load—a ratio of nearly 8:1 (speculative to firm).
    • Oncor (Texas Utility):Reported a queue of 186 GW of data center requests. For context, the utility’s entire current peak demand for all customers is only ~50 GW, suggesting the queue is inflated by nearly 400% of the grid’s total existing capacity.
    • PSE&G (New Jersey):Reported a 9.4 GW large load pipeline but expects only 10–20% of those inquiries to result in actual projects.
    • ERCOT (Texas Grid):Has received requests for over 220 GW of new load by 2030 (mostly data centers), which is more than double the state’s all-time peak demand record. 

Why This Is a Problem

1. 1. Planning Paralysis: Utilities cannot distinguish real projects from “zombie” projects. If they build transmission lines for all 190 GW (in AEP’s case), they would bankrupt ratepayers. If they wait to see which are real, they risk being too slow for the 24 GW that is real.
   2. Queue Backlogs: The “phantom” requests force grid engineers to perform complex impact studies for projects that don’t exist, delaying the connection of viable power plants and real factories by years.
   3. Artificial Scarcity: The illusion of zero capacity drives up power prices and panic-buying of land, further fueling the cycle of speculative multiple-filing. 

Why is this a Problem?

You might also ask, from the perspective of Big Oil, “Why is this an opportunity?”

Big Oil loves the high estimates of power demand and the expanded market for their products, and not just more volume, but more over the next several decades, just when we need to reduce carbon emissions, not raise them. Big Oil is in a frenzy to keep their business going for decades more, despite the counter need for this industry to decline.

There’s temptation, too, for the utilities who contract or build new generation capacity. While solar/wind/batteries can meet new energy needs (and be quicker and cheaper), power utilities remain drawn to action that follows business-as-usual thinking, and for many utilities, especially without governmental and regulatory guidance, that means more power plants.

The opportunity is now for Big Oil to future-proof the industry.

Does the Trump Administration strike you as leaning on governance and regulation to push for clean energy?

I don’t think so, but then maybe I’m as wrong as all those wild estimates about power requests to feed the AI industry.

The post Fossil Fuel Demand Growth Uber Alles first appeared on David Guenette.

I’ve been saying that AI’s projected electricity demand is celebrated by fossil fuel companies because this growth in electricity demand provides an anchor for Big Oil to keep selling natural gas for decades to come.

What me worry?

Sure I worry. A much clearer and devasting understanding about the “carbon budget” mankind faces has recently been recalculated and the general gist is that we have a smaller amount of carbon we can dump into the atmosphere before we push beyond 1.5C.

Or, as AI Summary puts it:

The 2025 Global Carbon Budget reports that the remaining budget to limit global warming to 1.5°C is “virtually exhausted”. Fossil fuel emissions continue to rise, projected to reach a record 38.1 billion tonnes in 2025, driven by, for example, high demand. While emissions from land-use change have declined, total global emissions remain at record highs. 

Key Findings on Recent Carbon Budget Changes 

• • 5°C Budget Exhaustion: Scientists warn that at the current rate of emissions, the budget for limiting warming to 1.5°C could be exhausted in approximately three years, with only about 130 billion tonnes of CO2 left to emit.
  • Rising Emissions (2024-2025): Fossil fuel emissions have continued to grow, with 2024 seeing a 0.8% increase. The 2025 projection is a 1.1% increase in fossil fuel CO2 emissions.
  • Total Emissions Flat: Despite rising fossil fuel emissions, the total CO2 emissions (including land-use changes) for 2025 are projected to remain relatively flat compared to 2024, due to a decrease in emissions from deforestation.
  • Weakened Sinks: The Global Carbon Project (GCP) notes that climate change has weakened natural land and ocean sinks, accounting for 8% of the rise in atmospheric CO2 concentration since 1960.
  • The “Carbon Clock”: The Potsdam Institute for Climate Impact Research continues to track the rapid shrinking of the remaining carbon budget for both 1.5°C and 2°C, highlighting the urgency of the situation.

The 2025 Global Carbon Budget, often highlighted by sources like Carbon Brief and Euronews, indicates that while some nations are transitioning to cleaner energy, the overall global trajectory is not yet declining fast enough to meet international climate targets. 

Want to worry more? Many scientists are now projecting that we are already on the path to or actually at 1.5C, with higher temperature increases in the average global temperature moving faster than previously considered.

“US leads record global surge in gas-fired power driven by AI demands, with big costs for the climate,” written by Oliver Milman and published in The Guardian on January 29, 2026, puts the issue of overspending our carbon budget squarely on gas generation and the new and planned gas generation planned to address the electricity-hungry AI data centers. The main projected culprit for the voracious spending down of said carbon budget largely rests with the U.S.

The article reports that “Much of this new capacity will be devoted to the vast electricity needs of AI, with a third of the 252 gigawatts of gas power in development set to be situated on site at datacenters.” I wrote about this back in September of last year, in a post titled “New Gas Generator Plants and the Plan to Flood the (Electricity Demand Growth) Zone.” In this post there’s a link to an AI analysis I did in a report called “The Future of U.S. Natural Gas Power Generation: Projections, Accuracy, and the Confluence of Limiting Factors to 2030.” Here’s that report’s Executive Summary:

Executive Summary

1.1 Overview of Projections and Core Findings

An analysis of U.S. energy market trends and projections indicates a notable ambition for future natural gas power generation. A key projection from the firm Enverus suggests the United States is on a trajectory to construct 80 new natural gas power plants by 2030, which would add an estimated 46 gigawatts (GW) of new capacity. This figure is a focal point for assessing the future of the nation’s energy infrastructure. However, a comprehensive review of the current market and regulatory landscape reveals that this aggressive projection is highly speculative. It is a needs-based assessment rather than a realistic forecast of what can be built, as its feasibility is called into question by a complex and multi-faceted set of constraints.

Looks like Big Oil has been busy selling the idea of new gas generation for AI and other growing electricity demand. Forty-six GW has blossomed to 252 GW, although the first number is U.S. back in 2025, and the second number is worldwide in 2026.

You have to admire the ambition of the fossil fuel industry. The amount of additional carbon emissions these new plants will spew is staggering… and staggeringly dangerous. Here’s a quote from The Guardian article:

The gas projects in development in the US will, if all completed, cause 12.1bn tonnes in carbon dioxide emissions over their lifetimes, which is double the current annual emissions coming from all sources in the US. Worldwide, the planned gas boom will cause 53.2bn tonnes of emissions over projects’ lifetimes if fulfilled, pushing the planet towards even worse heatwaves, droughts, floods and other climate impacts.

Of course, as I discovered in my analysis, the desire to build new gas plants is tempered by ability, especially in terms of supply chains, where turbines from the two main manufacturers of gas generators—GE Vernova and Siemens Energy—are mightily backlogged. Building out manufacturing capacity for such complex machines is no fast undertaking. Unfortunately, a recent development has entered the market, as described in the AI Summary of the search prompt “Using jet turbines for new gas generator plants and capacity”:

Using jet turbines, specifically “aeroderivative” gas turbines derived from aircraft engines, has emerged as a critical, fast-deployable solution for new, high-demand gas generator plants, particularly for data centers and AI-driven power needs. Companies like ProEnergy are retrofitting used military and commercial jet engines (e.g., GE CF6-80C2) into 48-megawatt power generators to provide rapid, “behind-the-meter” electricity. 

This approach addresses severe bottlenecks in securing new, large-scale utility turbines, which often face 3- to 7-year wait times. 

There’s a lot more to find with this prompt, but nothing to keep you calm if you worry about carbon emissions.

Here’s what it all means: All efforts to restrict and reduce carbon emissions will be more than offset by all these goddam new gas generators if they come to be. Offset and then some.

A Brief Interruption About My AI Overlords

I use AI as an effective productivity tool. My purview for climate change is a wide one. I write climate fiction that seeks to be based on realistic and accurate information about the causes and consequences of climate change, which means I’ve been hard at work understanding the science (enough to tell the difference between truth and bullshit). I give talks about climate change, and I have little interest in providing information that is wrong, so I make an ongoing effort to get things right and up to date. I find myself researching climate policy, political realities, and economic benefits of the clean energy transition. I’m active within the climate fiction world, reading widely in the genre and critiquing the various approaches to it. I’ve been at all of this for many years.

The world of climate change is extremely wide and multi-faceted. Keeping up on the latest findings of science and technology and policy proposals and economic and political realities is gained by triangulated a diverse and wide range of information resources. I’ve found that AI can be useful as a research agent, where, in response to a thoughtful prompt, AI ranges far and wide across the Web to collect and then collate and then analyze the relevant sources and then synthesize these findings into well-produced reports.

I have high confidence in these reports, and I haven’t found hallucinations to be a problem, but that’s because I know the subject well enough and broadly enough to cast a critical eye on sources and am able to review AI’s findings. What I can’t so easily do—although, of course, I’ve done this all too often and with all too much effort—is to search the Web high and low for the information, explanation, and opinion I seek.

For me, AI is a useful tool. It is not my buddy, nor do I spend much time generating funny images. For me, AI’s current capabilities are impressive.

This is not to say that the present hysteria about AI—the very hysteria that supports Big Oil’s play to flood the world with greenhouse gas-generating electricity—is sensible. My take is that there is wild financing being accumulated mainly in the hope of the debt accumulators being first to market. I believe that AI’s market will actually take a long time to develop as a deeply useful and pervasive embedded tool—evolving human culture is still a slow undertaking. I believe that there’s likely to be something on the order of a crash or bubble because of the disconnect between all that money and all that non-market.

I also believe that the rush for adding huge amounts of electricity generation is hysterical, albeit not in a funny way. I believe that whatever additional capacity may be required can be more cost-effective and more quickly produced with clean energy and by bringing digital management to the grids that already contain huge amounts of spare capacity that currently cannot be managed well.

And now back to our regularly scheduled programming.

Big Oil is the Enemy

I guess I shouldn’t be surprised that the fossil fuel industry doesn’t care about carbon emissions. If Big Oil did care, Big Oil would have a lot to care about.

There’s no rational explanation for the fossil fuel industry not knowing that carbon emissions must come down, not go up. Denial doesn’t cut it, not with the now long-held and clear scientific consensus on global warming and all the resulting data collection. Denial doesn’t cut it, not with the growing evidence in the form of extreme weather and not with the majority of individuals’ personal experiences. It is as if a competent adult would argue that he didn’t know that someone could be killed if he pointed a loaded gun at that person and pulled the trigger. The “Go figure” argument is as absurd for Big Oil as it is for our proverbial idiot.

Speaking of idiots, in America, Trump is the gaslighter-in-chief in regard to climate and the contributory role of fossil fuels, just as he is on so many other topics, including, of late, the ridiculous and entirely and patently demonstrable falsehoods around the ICE/CBP killings and related inciting behaviors. That’s why the connection between fighting for the American democracy and the climate fight are one in the same. The old order of Big Oil, along with a rogue’s collection of other “Bigs,” has placed its bet on a rising fascist state, damn the consequences.

I’m placing a different bet.

The post AI is Giving Me Gas first appeared on David Guenette.