Today's stories are about the practical limits on clean energy: who profits from keeping oil underground, what really drives power bills, and who controls local approval.
Picked from 48 new articles across 13 sources · arrow keys for other days
Today in 30 seconds
30s left
Peter Erickson finds each $70 barrel left unproduced shifts $60–$350 from consumers to producers, which matters for the Senate permitting deal. story 1 ↓
New York regulators say Climate Act programs made up 5–9.5% of a typical 2024 bill, a check on Blakeman's attack on Hochul. story 2 ↓
Tulsa exempted Host Digital from its data center moratorium 9-0 after it proposed retrofitting a 55-megawatt bitcoin mine. story 3 ↓
Greenlight America says it wins 79% of local clean energy hearings, with about 90 staff and roughly $20 million raised this year. story 4 ↓
China's biggest solar makers cut a fifth of their staff last year, and no EU firm can make top-grade transformer steel. story 5 ↓
Trump has reopened the fight over Orsted's Revolution Wind, another reminder of the policy risk facing US offshore wind.
Peter Erickson's research helped make the case against Keystone XL. He now argues that "no new oil and gas fields" rests on weak evidence and makes incumbent producers richer. He says growing clean energy beats blocking fossil fuel projects.
Why it matters · The Senate's bipartisan permitting bill trades easier gas pipelines for easier power lines, and Erickson's numbers bear directly on whether climate advocates should take that deal.
Know this first
Context from the editor, not the article · 34 sec
1 · Supply-side climate policy
Most climate policy cuts demand for fossil fuels through efficiency, electric cars and clean power. Supply-side policy instead limits production: blocking pipelines, refusing drilling leases, or phasing out extraction. Keystone XL made this a central activist strategy in the 2010s.
2 · Why one barrel matters globally
Oil trades on a single world market. Adding supply anywhere pushes the global price down a little, so people use more oil. It also squeezes out some higher-cost production elsewhere. How the effect splits depends on elasticities, which measure how strongly producers and consumers respond to price.
3 · The permitting bill
2/5Inside Climate News· By Lauren Dalban· Oct 14
Hochul and Blakeman clash over New York's rising electricity bills
In New York's only governor's debate, Democrat Kathy Hochul defended a mix of renewables and nuclear power. Republican Bruce Blakeman blamed clean energy for high bills, but state data puts climate programs at under a tenth of a bill.
Why it matters · The claim that "clean energy raises your bill" is now a standard campaign argument, and New York has published numbers that let you check it.
Know this first
Context from the editor, not the article · 28 sec
1 · New York's Climate Act
The 2019 Climate Leadership and Community Protection Act requires 70% renewable electricity by 2030, a zero-emission grid by 2040, and an 85% cut in greenhouse emissions by 2050. Some of the cost is paid through charges on utility bills.
2 · Why gas sets power prices
In wholesale power markets, the last plant needed to meet demand usually sets the price for every generator. In New York that plant is often gas-fired, so electricity prices rise and fall with natural gas prices.
3 · What a bill pays for
A residential electric bill pays for the electricity itself, for delivering it over transmission and distribution lines, and for smaller program charges. The state Public Service Commission regulates the utilities and approves their rates.
3/5Heatmap· Jael Holzman· Oct 16
Developer pitches small data centers in existing buildings to win local approval
Harmol Samra, CEO of Host Digital, told Heatmap his company wins local support by building 20–100 megawatt data centers, often in existing powered buildings, and funding community programs instead of asking for tax breaks.
Why it matters · Local moratoriums are becoming a real limit on data center growth, and reusing sites that already have grid connections is one concrete way some projects avoid adding new demand.
Know this first
Context from the editor, not the article · 25 sec
1 · Social license to operate
Industry shorthand for ongoing acceptance by local communities, separate from legal permits. Without it, projects face moratoriums, zoning denials and lawsuits even when they meet every regulation.
2 · The data center backlash
AI-driven demand has brought proposals for campuses using a gigawatt or more, about as much as a large city. Worries about electricity bills, water and noise have led some towns and counties to pause approvals.
3 · Why grid connection matters
New large loads often wait years for utility interconnection and transmission upgrades. A site that already has power service and draws it can be put to new use much faster, without adding demand.
4/5Volts (David Roberts)· David Roberts· Oct 14
Group getting supporters to local clean energy hearings says it wins 79%
Greenlight America tracks the county hearings where wind, solar and battery projects are approved or blocked, and gets local supporters to show up. It says it now wins 79% of its campaigns, up from 60% in its pilot.
Why it matters · If cheap clean power is now mostly held up by local permission, not by cost or technology, then turning up at hearings is a lever that can be measured and scaled.
Know this first
Context from the editor, not the article · 30 sec
1 · Local permitting decides projects
In the US, most large solar, wind and battery projects need approval from a county or town. That usually means a public hearing where residents can speak. Local rules such as setbacks or outright bans can rule out whole areas.
2 · The 2025 tax credit cutback
In 2025, Republicans passed a budget law that phased out federal tax credits for wind and solar. To still qualify, developers had to start construction by a deadline. That pushed many projects ahead all at once.
3 · Greenlight America's model
Greenlight America is a nonprofit funded by philanthropy, founded by Matt Traldi, a former labor organizer and a co-founder of Indivisible. It doesn't send its own staff to testify. It finds local groups and residents and helps them advocate.
5/5Construction Physics· Brian Potter· Oct 17
China's biggest solar makers cut a fifth of their staff last year
Brian Potter's weekly links roundup reports that headcount at China's five largest solar manufacturers fell 21% last year, after a 31% drop the year before. Europe, like the US, also lacks top-grade transformer steel.
Why it matters · Cheap solar panels and scarce transformers are two of the main forces shaping how fast grids can add clean power. Both point to supply-chain strain, not demand.
Know this first
Context from the editor, not the article · 25 sec
1 · China's solar glut
China makes most of the world's polysilicon, wafers, cells and modules. Factory capacity has grown far faster than global demand, so prices have collapsed and many manufacturers are selling at a loss.
2 · Why transformers need special steel
Power transformers wrap copper windings around a core made of grain-oriented electrical steel (GOES). This steel is processed so that magnetic losses fall sharply along one direction. Better grades waste less energy as heat.
3 · The transformer shortage
Rising electricity demand, grid replacement and new data centers have stretched transformer lead times to years in many markets. The constraint runs back through the supply chain to inputs like GOES.
In September 2026, two Republican and two Democratic senators introduced the Bipartisan American Affordability and Jobs Act. It would speed up, and sometimes require, long-distance power lines and other decarbonization infrastructure. It would also make some natural gas pipelines easier to build.
Image: Heatmap
The Senate is weighing a bipartisan permitting bill that makes long-distance power lines easier to build, and in some cases requires them. In return, some natural gas pipelines also get easier. On Heatmap's Shift Key podcast, Robinson Meyer asked Peter Erickson whether climate advocates should take that trade. Erickson is a geologist and researcher affiliated with the 1. Few researchers are cited more often in court fights over fossil fuel infrastructure.
From Keystone to second thoughts
Erickson made his name with a 2014 paper on 2, a proposed pipeline from Alberta's oil sands to Gulf Coast refineries. The State Department estimated it would add 1–27 million tons of CO2. That figure only counted how much dirtier oil-sands crude is than the oil it would replace. Erickson pointed out that the pipeline would also add supply to the world market. More supply lowers the price, and a lower price raises consumption. His range was 0–110 million tons. Canada later built the 3 pipeline because its oil was stuck without enough pipeline capacity. Erickson takes that as a sign the real effect was mid-range or higher.
He hasn't changed the math since then. What has changed, he says, is where the exaggeration comes from: "the exaggerated claims are no longer just on the fossil fuel side."
The wealth-transfer problem
Limiting supply cuts emissions only by raising prices. That higher price reaches every producer in the world. In his 2026 paper in Environmental Research Letters, Erickson finds that each $70 barrel kept undeveloped moves $60–$350 from consumers to producers. That is more than the blocked company loses. The result flips only if oil demand falls very fast, and Erickson thinks demand is on a plateau, not yet falling.
He also disputes the slogan that 1.5°C allows "no new oil and gas fields." Of the 97 scenarios the 4 (the UN's climate science panel) assessed that keep warming to 1.5°C, fewer than half have demand falling fast enough to make new fields unnecessary. The 5's net-zero scenario finds that no new fields would be needed, but it explicitly does not recommend blocking them. On the 6, Erickson consulted for a major group that called stopping it its top climate priority. His own analysis found the pipeline's emissions "very, very small."
Grow the replacement
Erickson's 2015 7 paper looked at how fossil systems stay entrenched through cheap operation, institutions and political power. He says it should have stressed a different conclusion: the same unit of policy help benefits wind and solar two to three times more than gas power. So, he argues, permitting reform matters right now for keeping renewables growing. He hasn't taken a position on the bill itself and is waiting for quantitative analysis of its emissions effects.
The Senate is considering the Bipartisan American Affordability and Jobs Act. Two Republican and two Democratic senators introduced it in September. It would make long-distance power lines and other decarbonization infrastructure easier to build, and in some cases require them. It would also ease some natural gas pipelines. For climate advocates, the question is whether more fossil fuel infrastructure is ever worth accepting in exchange.
Heatmap's Robinson Meyer put that question to Peter Erickson on the Shift Key podcast. Erickson trained as a geologist and is an affiliated researcher at the 1. He has spent years as an expert witness against fossil fuel projects in court, and Meyer thinks he may be the most-cited climate researcher in such cases. His recent work argues that some claims made in those venues have run ahead of the evidence. Erickson says the math is the same as before; the problem is that "the exaggerated claims are no longer just on the fossil fuel side."
2: bringing in market effects
Keystone XL was a proposed pipeline from Alberta's oil sands, a carbon-intensive source of crude, to Gulf Coast refineries. President Obama said he would approve it only if it didn't significantly worsen climate change. The State Department estimated 1–27 million tons of added CO2. Erickson's 2014 paper in Nature Climate Change showed that this range only captured how much dirtier oil-sands crude is than the oil it would replace.
He argued it missed the main point of the pipeline. Canada couldn't get all its oil to market because it lacked pipeline capacity, and Keystone would remove that limit. Each added barrel lowers the world price slightly. People then drive a bit more, and carmakers build slightly less efficient cars. Part of the new barrel is newly consumed, and the rest displaces higher-cost production elsewhere. How it splits depends on 3, which measure how strongly producers and consumers respond to price.
The transcript is muddled on the exact split. Erickson describes about 0.4 barrels newly consumed per new barrel. Meyer recalls the 2014 paper using 0.6. Erickson's explanation for the drop is that the 2014 market was tight, on a steep part of the cost curve, so supply changes moved prices a lot. Electric vehicles have since shifted demand down. That moves the market onto a flatter part of the supply curve, where supply changes matter less. He notes this hasn't yet been rigorously published.
His Keystone range was 0–110 million tons. The zero end assumes all the oil reaches market some other way. The top end assumes none of it would have. Neither happened: Canada later built the 4 pipeline, and its regulator warns capacity may get tight again. Erickson's guess is mid-range or higher.
Lock-in, California, and subsidies
A 2015 paper defined 5. Fossil systems persist because they are cheap to run once built, and because they come with trade groups, worker training pipelines and lobbying power. Erickson stresses that lock-in is not binary. Coal plants closed early once gas undercut them, and he says new gas plants built for data centers needn't run for 30–40 years either.
His 2018 Nature Climate Change paper argued that California could wind down oil production as it cut oil demand. Without that, falling local demand lowers world prices and makes oil cheaper for everyone else. The paper prompted a California Air Resources Board workshop where economist Severin Borenstein rebutted it. Borenstein agreed with the math but said the resulting price rise transfers wealth from consumers worldwide to incumbent producers.
A 2021 paper looked at tens of thousands of discovered but undeveloped U.S. oil and gas fields under 16 subsidies and regulatory exemptions. Well over 90% of the subsidy value went straight to profits. Almost no new oil projects this decade depended on the subsidies, and about a quarter of gas projects did. Erickson's illustration: a field that earns a 15% return without subsidies earns 22% with them, and gets drilled either way. Cutting subsidies would hit profits, he and Meyer agree, but not emissions.
The 2026 paper: who gets paid
This summer Erickson rebuilt Borenstein's model from scratch. The result, published in Environmental Research Letters as "Rethinking the Case Against New Oil and Gas Fields," confirmed Borenstein's point. Limiting supply cuts emissions only through higher prices, and that price rise reaches every producer. Erickson finds that each $70 barrel left undeveloped transfers between $60 and $350 from consumers to producers. That is more than the blocked company loses.
Meyer compares it to a small version of what happened when Iran closed the Strait of Hormuz in early 2026: fuel prices jumped and producer profits rose. Erickson also notes that cartels like OPEC exist to restrict supply for exactly this reason.
There is an exception. If demand were falling very fast, the price effect of removing a barrel would be so small that the sign flips. Critics point to EV sales in China and Brazil. Erickson replies that peak oil demand has been predicted for years, and demand looks more like a plateau.
He also rejects the claim that 1.5°C rules out new fields. Of the 6's 97 scenarios that hold warming to 1.5°C, fewer than half have oil and gas demand falling fast enough to make new fields unnecessary, and even fewer of the 2°C scenarios do. The 7's net-zero pathway finds new fields wouldn't be needed if demand fell that fast. But it explicitly says blocking them is not a strategy it recommends. Turning that if-then into a mandate, Erickson says, is "plainly wrong." He adds that if the carbon budget is too small for any new project, then every ton of emissions is equally "incompatible," so the slogan doesn't help choose between policies.
Politics and the bill
Erickson argues that activists misread politicians like Joe Manchin. Manchin once fired a gun at cap-and-trade in a campaign ad, then delivered the deciding vote for the Inflation Reduction Act. Activists attacked him over the 8, a gas line he championed. One major group called stopping that pipeline its top climate priority. Erickson was the group's consultant, and he found the pipeline's emissions "very, very small." Meyer adds that if advocates had accepted Manchin's permitting deal, today's bill wouldn't be needed.
Erickson now reframes his lock-in work. The best way to break lock-in is to grow what can beat it. His 2015 ratio found that a unit of policy help benefits wind or solar two to three times more than gas power. Part of that comes from moving up the cost curve. The rest comes from building the trade groups, workforces and community ties that incumbents already have. He cites a 2026 European study of past energy transitions, which found that how far an S-curve of adoption climbs before flattening depends on policy, including grid integration and permitting.
He hasn't endorsed the bill. He is waiting for quantitative estimates of whether its emissions upside clearly beats the downside, and the article doesn't report any such numbers yet.
Remember
U.S. fossil fuel subsidies mostly add to profits: almost no new oil projects depend on them, and about a quarter of gas projects do.
The IEA says new fields wouldn't be needed on a 1.5°C path, but it doesn't recommend blocking them.
Erickson's 2015 ratio: the same unit of policy help benefits wind or solar two to three times more than gas.
In New York's only governor's debate, Democratic incumbent Kathy Hochul and Republican challenger Bruce Blakeman, the Nassau County executive, offered sharply different explanations for high electricity bills and different plans to fix them. Hochul leads by about 12 points in recent polls, with under three weeks until the election.
The problem
Average residential bills in New York rose 10 percent from 2024 to 2025, faster than the national average. The state has the eighth-highest residential electricity prices in the US, and demand is climbing.
Blakeman's claim vs. the state's numbers
Blakeman blamed clean energy programs, calling them a "slush fund" for "science projects" like offshore wind farms. He also said solar panels are "ruining agriculture."
The state's figures don't support that as the main cause. The 1, which regulates utilities, reported in 2025 that measures required by the 2019 2 made up 5 to 9.5 percent of a typical 2024 residential bill. That law requires 70% renewable power by 2030.
Most of the bill pays for buying and delivering electricity. Prices rise with demand and with the state's heavy reliance on gas plants. Gas plants often set the wholesale price, so bills follow the gas market. Utilities are also spending more on grid expansion to connect new generators and customers.
Two supply plans
Hochul pitched an "all of the above" approach: renewables plus more nuclear. She would not say whether she would reopen 3, a nuclear plant near New York City that closed in 2021. Critics of nuclear cite cost, health concerns and the storage of radioactive 4.
Blakeman's main proposal is to bring back 5, which uses high-pressure fluid to release gas from shale rock. The state banned it more than a decade ago over health concerns. The article notes that wind and solar paired with batteries can typically be built more cheaply than new fossil plants.
Data centers
Data centers, 6 and manufacturing are driving demand. The state's grid operator has a growing backlog of data center connection requests. Hochul cited her yearlong moratorium on new data centers and a pledge that new ones must "bring their own power." In other states, data center demand has led to plans for new gas plants.
She also pointed to rebate checks of up to $200 for 8.2 million households. She would not say whether they would be repeated.
What's missing
The article doesn't break down the rest of the bill, so it's unclear how much of the 10 percent rise came from gas prices versus grid spending. Neither candidate gave costs for their supply plans.
New York Governor Kathy Hochul, a Democrat, and her Republican challenger Bruce Blakeman met in the state's only gubernatorial debate of the race. Electricity costs were one of their main points of conflict. Blakeman has been Nassau County executive since 2022. With less than three weeks to Election Day, recent polls put Hochul about 12 percentage points ahead.
Why bills are the issue
Average residential electricity bills in New York rose 10 percent from 2024 to 2025, faster than the national average. The state has the eighth-highest residential electricity prices in the country, and demand is growing. That made affordability a central question for both candidates.
Blakeman's charge, and what the state data says
Blakeman blamed high utility bills partly on the state's clean energy programs. He called them a "slush fund" for "science projects," such as offshore wind farms, and said solar panels were "ruining agriculture."
The state's own figures put that charge in proportion. In a 2025 report, the 1, which regulates New York's utilities and approves their rates, looked at costs from measures required by the 2019 2. That law sets targets of 70% renewable electricity by 2030 and a zero-emission grid by 2040. The PSC found those measures made up between 5 and 9.5 percent of a typical monthly residential electric bill in 2024.
The article does not give a dollar bill size. As an illustration only: on a $150 monthly bill, that range works out to roughly $7.50 to $14. Climate programs are a real cost, but they are not most of the bill.
According to the article, most of the bill pays for buying electricity and delivering it. Rate increases are often linked to two things:
Rising demand, which pushes prices up.
Heavy reliance on gas-fired generation. In wholesale markets the last plant needed to meet demand usually sets the price for all generators. That plant is often a gas plant, so power prices follow the natural gas market.
Utilities are also spending more to expand the grid so it can connect new generators and new customers. That spending is recovered through delivery charges.
Two plans for more supply
Hochul described an "all of the above" strategy: more renewable energy plus an expansion of nuclear power. She did not say whether she would try to reopen 3, the nuclear plant north of New York City that closed in 2021. Nuclear is still contested in the state. Critics raise concerns about cost, public health, and the long-term storage of 4, which is extremely radioactive.
Blakeman offered few specifics beyond attacking renewables. His main proposal is to bring back 5, which means injecting fluid at high pressure into shale rock to release natural gas. New York banned fracking statewide more than a decade ago over public health concerns. Lifting the ban would mean reversing state law and regulation, which a governor cannot do alone.
The article adds that wind and solar paired with battery storage can typically be built more cheaply than new fossil fuel plants. That could help meet rising demand while keeping the state on track for its climate goals. The article gives no cost figures for New York specifically.
Data centers and demand
Demand is rising because of data centers, 6 (switching heating and appliances from gas to electricity), and manufacturing. The 7, which runs the state's grid, faces a growing backlog of applications to connect proposed data centers.
Hochul pointed to her executive order imposing a yearlong statewide moratorium on data centers. She also pledged to make new facilities "bring their own power," meaning they would supply their own generation rather than draw on the shared grid. The article notes that in other states, data center demand has led to plans for new natural gas plants. "Bring your own power" does not by itself determine what kind of generation gets built. Environmental advocates have asked for more public input on the proposed policy.
Short-term relief
Hochul said she was "laser-focused on affordability." She cited a rebate that will send checks of up to $200 to 8.2 million households, which is at most about $1.6 billion in total. She did not answer whether the rebates would become annual. A one-time check offsets bills but does not change what drives them.
Politics around the edges
Blakeman has aligned himself with President Trump. Hochul tied him to Trump administration policies, including tariffs and the Iran war, which she said have raised prices. Blakeman replied that he and Trump have "significant policy differences." Hochul became governor in 2021 when Andrew Cuomo resigned, and in 2022 she defeated Lee Zeldin, now the EPA administrator.
What the article leaves open
The report is a debate recap, not an analysis of costs. It does not break down the remaining 90-plus percent of the bill between supply and delivery, so it is unclear how much of the 10 percent rise came from gas prices, grid spending, or demand. It does not say what Hochul's nuclear expansion would cost or when it would arrive. It also does not estimate how much fracking would actually lower prices, given that New York's gas price is largely set by regional and national markets.
Remember
State regulators say Climate Act measures made up 5–9.5% of a typical 2024 New York residential bill.
Gas plants often set New York's wholesale power price, so bills follow natural gas prices.
Hochul's $200 rebates go to 8.2 million households, but she wouldn't say if they will repeat.
Harmol Samra, CEO of data center developer Host Digital, told Heatmap's Jael Holzman how his company gets local approval while much of the industry runs into 1. It's one developer's pitch. Most of it is his argument, not independent data.
Small, and already plugged in
Host Digital builds data centers of 20 to 100 megawatts, which it calls "2." It targets sites that are "3," meaning they already have grid power, and that can be delivering within 10 to 14 months. Proposed AI campuses elsewhere run to multiple gigawatts. Samra says his buildings look like the warehouse-style facilities communities have seen for 20 years, not mega-campuses dropped into a town of 20,000.
Using an existing grid connection helps twice over. It avoids the multi-year wait that new large loads often face for 4. And it adds no new demand, which takes away opponents' strongest argument about power bills.
Tulsa
The company's first site was a bitcoin mine in Tulsa, Oklahoma, which it planned to convert. After Host Digital bought it, the city passed a data center moratorium. The company told officials the site had drawn 55 megawatts for years and the building already existed, so the project added no new load. The city council exempted it 9-0. According to Samra, officials said the moratorium was meant for mega-campuses, not projects like his. He admits the limit: there are no gigawatt-scale buildings to retrofit.
Benefits instead of tax breaks
Samra says Host Digital asks for no tax breaks. Instead it funds local programs before a site starts paying property tax, such as a high school AI training program and paid AI usage credits so students can use commercial models. No dollar figures are given.
He also supports 5. In his illustrative example, a town of 10,000 households paying $3,000 each in property tax pays $3 million a year, and a data center could cover that. He lists other commitments developers can make: paying for their own grid upgrades, and building a new renewable plant alongside each data center. The interview doesn't say which of these Host Digital has actually done.
His diagnosis
Samra argues the backlash is really about AI. Data centers are its physical form, and stopping a building feels like the one lever communities have. He claims the factual concerns (electricity costs, water, noise) "have been addressed." He offers no evidence, and rate impacts in particular are still contested. His own Tulsa win depended on showing the project added no new load.
The takeaway: reusing powered industrial sites at a familiar scale is a real way through local opposition. But it's a niche, and it can't supply most of the AI compute now being planned.
Heatmap's Jael Holzman interviewed Harmol Samra, CEO of data center developer Host Digital, about why the industry keeps losing local fights over new facilities. Samra's answer is partly a business model and partly a diagnosis. It's worth reading as one developer's pitch, not as independent evidence. Heatmap ran it as a lightly edited Q&A.
The strategy: "1" and already energized
Host Digital builds data centers of 20 to 100 megawatts. The company calls them "right-sized." It targets sites that are "2," meaning they already have grid power, and that can deliver within 10 to 14 months. For scale, AI campuses now being proposed elsewhere reach multiple gigawatts, so 1,000 megawatts or more each. Samra describes his buildings as the kind communities "have been seeing for the last 20 years," similar to an industrial warehouse in a business park. He contrasts them with "multi-gigawatt campuses in a community that has 20,000 people."
The energized-site focus matters for the grid as well as for public relations. In many regions, new large loads wait years for 3 and transmission upgrades. A building that already draws power can skip much of that queue. It also adds no new demand to the local system, and that undercuts the main argument opponents make about electricity bills.
The Tulsa test case
Host Digital's first site was in Tulsa, in northeast Oklahoma: a bitcoin mining facility it planned to convert into a data center. After the company bought the site, while it was still arranging leasing and financing, the city enacted a 4 on data centers.
The company met with the planning office and city council members. Its argument was that the site had been drawing 55 megawatts for several years, that the building already existed, and that the conversion added no new load. The council voted 9-0 to exempt the project. According to Samra, officials told him "the moratorium was not put in place to capture developments like you." It was aimed at mega-campuses.
Samra says retrofits make a project "more bite-sized to digest," because the main fears can be addressed before anyone raises them. He also concedes the limit: "there are no gigawatt-scale things you can retrofit." The approach works for mid-sized compute. It does not scale to the frontier training clusters behind most of the projected demand growth.
Community benefits without tax breaks
Samra says Host Digital asks for no tax breaks or incentives anywhere it builds, and sets up community benefits programs instead. His argument is about timing: a data center pays property tax only once it is operating, maybe a year out, so the company funds something in the meantime. His examples are an AI training program at a high school and paying for AI "5" (units of model usage, which are billed per token) so students can use commercial models. The interview gives no dollar figures for these programs.
He also likes the idea of 6, and gives an illustrative calculation. Take a town of 10,000 households, each paying about $3,000 a year in property tax, for $3 million in total. If a data center's tax payments could cover that whole amount, residents would see a direct cash benefit. Other "dollar-denominated" commitments he lists are paying for the developer's own infrastructure upgrades and building a new renewable plant alongside each data center to "bring energy back to the grid." He says some of these ideas "have resonated with communities." The interview doesn't say whether Host Digital itself has made any of these commitments. Nor does it explain the mechanics, for example whether payments to a town would be passed through as lower rates for residents.
His wider principle is that the industry has to be "net positive" to a community, through jobs, revenue or infrastructure. He says other energy sectors learned this when they had to deal with their own climate and local impacts.
His diagnosis: the fight is really about AI
Samra argues the backlash comes mainly from unease about AI itself. Data centers are "a physical embodiment" of it. In his framing, communities can't tell a company like Anthropic how to change its models, but they believe that blocking a building might slow AI "until we all figure out where it goes."
He also claims the factual objections have been "addressed": electricity costs, water use, and noise, which he says can be managed by putting cooling equipment outside behind sound barriers. Treat this as advocacy. The interview offers no data to support it. Effects on electricity rates in particular are still debated in utility rate cases and regional grid planning, and the answer depends heavily on who pays for new generation and transmission. His own Tulsa example cuts against the claim: the "no new load" argument won the exemption precisely because new load is a concern that hasn't gone away.
What's left open
The interview shows a workable niche. Reusing powered industrial sites, especially stranded bitcoin mines, at a scale towns already know can get through moratoriums written with gigawatt campuses in mind. It doesn't show that this model can meet more than a small share of AI compute demand. It also gives no costs for the benefits programs and no evidence for the claim that grid, water and noise impacts are solved. Some questions it doesn't answer: how many energized retrofit sites exist, what happens to the grid capacity a bitcoin miner gives up, and whether communities will tell "right-sized" projects apart from mega-campuses once moratoriums spread.
Remember
Tulsa exempted Host Digital from its moratorium 9-0 because the converted bitcoin mine already drew 55 megawatts.
Retrofitting powered sites avoids new grid load and interconnection queues, but Samra admits it can't reach gigawatt scale.
Samra's claim that grid, water and noise concerns are 'addressed' comes with no supporting data in the interview.
Matt Traldi runs 1, a nonprofit that does one thing. It finds the local hearings where wind, solar and battery projects get approved or blocked, and gets supporters to show up. At Climate Week NYC, he told David Roberts of the Volts podcast that the group has grown fast since their January 2025 interview.
The problem
Traldi argues that the main obstacle to building clean power is no longer technology or cost. It is local permitting. A handful of opponents attend a county hearing. Supporters exist but stay home, so projects "lose by forfeit." There is no central list of these meetings across more than 3,100 counties, and the process differs by state.
How it works
Greenlight's 2 combines two sources: tips from developers about permit filings, and web scrapers that pull local agendas and public notices. Algorithms and a research team narrow thousands of alerts down to decisions where local voices could matter.
Greenlight then never shows up under its own name. It recruits local partners, almost 300 so far, from faith groups and unions to the conservative 3. Traldi's view is that residents persuade and outsiders don't. The messages stick to local stakes like tax revenue and property rights, not climate. Because the mood in the room sways officials, the group brings supporters to hearings even when it expects no opposition. Its internal motto is "there is no bag": no vote is ever "in the bag."
The numbers
These are the organization's own figures:
67 gigawatts advanced in total, which Greenlight says avoids more than 1 billion tons of lifetime emissions.
Win rate: 60% in the pilot, 75% last year, 79% this year. Traldi guesses the industry baseline is about a coin flip.
Pace: once four wins in a quarter; last month, eight in one day.
Size: nearly 90 staff, and almost $20 million raised this year.
Traldi admits there is no clean counterfactual, meaning no way to know what would have happened without Greenlight. He says early comparisons with historical decisions look "significantly" better, but gives no baseline numbers.
Why now
The 2025 budget law (the 4) is phasing out wind and solar tax credits. Traldi cites estimates from 5, an energy research firm, that 216–240 GW of solar and about 20 GW of wind locked in credit eligibility by starting construction. A good year used to be about 50 GW. Those projects must be running within four years, so permits need to clear by mid-2028. That means a wave of hearings.
Headwinds
Traldi, a progressive, says Democratic-leaning places can be among the hardest, citing New York's local battery-storage bans. Anger about data centers is also spilling onto solar and batteries. Some opponents stoke it on purpose. Worse, in his view, would-be supporters stay home because they wrongly believe that blocking a solar farm keeps a data center away.
On stage at Climate Week NYC, David Roberts of the Volts podcast checked back in with Matt Traldi, founder and CEO of 1. Roberts first interviewed him in January 2025. Since then the nonprofit has grown fast. Traldi now puts the group's cumulative total at 67 gigawatts of clean power advanced and says it has nearly 90 staff. He says it has raised almost $20 million this year. Roberts's show notes give somewhat different figures: 20.9 GW in 2025, about 55.7 GW since the last interview, and growth from 20 to 37 staff. Treat all of these as the organization's own self-reported counts.
The problem: losing by forfeit
Traldi's case starts from something he kept hearing in the industry. Over the next decade, most US emissions cuts will come from the power sector. The main obstacle there is not technology or cost, since solar and wind are the cheapest and fastest power to build. The obstacle is local permitting. A few opponents show up to a county hearing and make noise. Supporters exist in every community but rarely attend, so projects "lose by forfeit."
This is hard to organize around because there is no central list of these decisions. The US has more than 3,100 counties. Counties are the most common decision-makers, but in some states it is municipalities, and every state's process is different. Hearings are often held at times like Tuesday at 9:30 a.m., precisely because almost nobody goes.
How the model works
Greenlight runs what it calls an 2 with two inputs. The first is developers. Greenlight is not funded by developers, but some share when they plan to file for permits. The second is a data operation: web scrapers pull agendas and public notices from local government sites. Algorithms then filter thousands of alerts down to the ones where local voices could tip the outcome, and a research team checks those.
Once a decision is flagged, Greenlight never appears under its own name. It looks for local partners. It has worked with almost 300, including Faith in Place (a faith-based group in Illinois), 3 (a conservative group), labor unions and Poder Latinx. Traldi's core belief is that almost anyone is an effective advocate in their own community, and almost no outsider is.
The messages focus on local benefits, not climate. In Traldi's framing, commissioners aren't deciding whether climate change is real. They're deciding whether this project is good for the county. So supporters stress tax revenue for public services, and property rights, meaning a landowner's freedom to lease land for a solar farm. The pitch to volunteers is about impact. Greenlight's figure is that a mid-size utility-scale solar project equals about 4 million trees in emissions terms. If five speakers can swing a council, Traldi says, each spends an afternoon and "plants" 800,000 trees.
He also says just being in the room matters a lot. If the first five speakers back a project, it feels like it's sailing through. If they oppose it, voting yes feels like political suicide, whatever officials thought beforehand. Early losses taught the group an internal motto: "there is no bag." Projects developers thought were "in the bag" were lost when opposition appeared late. Now Greenlight brings supporters to hearings even when it expects no fight, and they stay quiet if none comes.
The deadline sprint
The 2025 budget law, called the 4 (OBBBA), phased out wind and solar tax credits. Traldi describes the timeline this way. Projects had to start construction by this July, and under IRS rules that can mean buying equipment off-site. They then have four years to be placed in service. Most projects started in 2025, so they must be running by the end of 2029. Building on site and connecting to the grid takes 12–18 months at minimum. Working backwards, permits must be done by early-to-mid 2028.
Traldi cites 5, an energy research firm, as estimating that 216–240 GW of solar and about 20 GW of wind have been "6," meaning they locked in credit eligibility by starting construction. A good year used to be about 50 GW, so that backlog is several years' worth of normal building. The result is a flood of hearings. Greenlight once had quarters with four wins. Last month it had eight in one day. On fears of a crash after the deadline, Traldi says industry contacts expect a dip of at most 20–30% that won't last, because power demand is rising.
Measuring impact
Greenlight's headline numbers count every campaign it worked on that advanced: 67 GW, more than 1 billion tons of lifetime emissions avoided, and 3.5 times New York City's power needs. Traldi admits advocacy has no clean counterfactual, meaning no way to know what would have happened without it. As evidence he points to cases won by one flipped vote, and hearings where supporters outnumbered organized opponents. The group is also scraping records of past local decisions to build state baselines. Early results put its win rate "significantly above" those baselines, but he gives no baseline figures. Its own win rate was 60% in the pilot, 75% last year and 79% so far this year. He guesses the industry average is roughly a coin flip.
Where it's hard
Traldi, a self-described lifelong progressive, says some of the toughest places are Democratic-leaning ("blue") ones. New York has many local bans on battery storage, though state regulators just approved the 7 transmission line. He also cites a hard fight in Santa Fe County. He calls the idea that "solar on parking lots" can avoid trade-offs over land a fantasy. Most restrictive local ordinances, he says, are either outright bans or "bans on everything but name," for example through extreme 8 (minimum distances from homes or property lines).
Data centers are a growing problem. Greenlight doesn't work for or against them. But anger about them is spilling over onto clean energy. Some opponents draw the link on purpose, and Roberts names Steve Bannon. The bigger harm, in Traldi's view, is that natural supporters are switching sides or staying home. One example is the rumor that blocking a local solar farm will keep a data center out. Unlike a hearing vote, he says, this problem can only be chipped away at, not won.
Greenlight does flag some projects as "red flags," such as one that would wipe out a species' only habitat. It also tries to broker deals between developers and local critics. In one case it introduced a developer planning solar on reclaimed mine land to the local watershed association. Traldi says that turned likely opposition into support.
What's left open
Its own estimate of what's needed is about 600 large projects a year, almost two a day. Greenlight's aim is to work on every local clean energy decision in the country, and Traldi says the field is nowhere near that. Without the baseline numbers, there is no independent way to tell how many of those 67 GW Greenlight actually tipped.
Remember
Greenlight's win rate rose from 60% to 79%, but it hasn't published baselines, so its own causal impact is unproven.
The tax-credit phaseout means the number of local hearings will peak before mid-2028, when permits must clear.
Traldi argues anger about data centers mostly hurts clean energy by keeping natural supporters home, not by creating new opponents.
New York projects bunch at exactly 99 units, just below the 485-x tax break's tougher threshold.1 / 2
Brian Potter's weekly Construction Physics reading list, a roundup of links on buildings, infrastructure and industry, includes two items that bear directly on the energy transition. One is about solar panels being too cheap to make money on. The other is about the steel that transformers need being too hard to make. About two-thirds of the list is paywalled. A promised item on grid-scale battery reliability falls in that part, so it isn't covered here.
China's solar shakeout shows up in headcount
Citing Reuters, Potter reports that employment at China's five biggest solar manufacturers fell 21% last year, after falling 31% the year before. Compounded, that leaves these firms with roughly 55% of the workforce they had two years ago. Potter calls the competition "ruthless."
For context: Chinese factories can produce far more panels than the world installs, so prices have fallen below many producers' costs. Buyers get cheap modules, but the industry is consolidating, and the job cuts are the clearest sign so far of how deep that consolidation runs.
Europe can't make the best transformer steel either
Power transformers are built around a core of 1, steel processed so that magnetic energy losses fall sharply in one direction. Higher grades mean more efficient transformers. An IEEE Spectrum piece quoted in the list says only two EU companies make GOES: Thyssenkrupp and Stalprodukt. Like US producers, neither can make the highest grades.
"The global leaders… by and large are in [South] Korea and Japan," an expert named Alimento told the magazine. The methods are "a heavily protected trade secret." With transformer lead times already long, both the US and EU depend on imports for the best core steel.
AI is pulling on other supply chains
2, the Dutch company that builds the lithography machines used to make advanced chips, says it can't meet customer demand for 2027 and 2028.
Memory shortages caused by AI data center construction have pushed smartphone shipments down 20% year over year, according to Counterpoint.
GlobalFoundries will run two new chip 3 (chip factories) in New York, funded by $21 billion from Japan. The US government is taking a 4% equity stake.
Also noted
New York's "4" tax break requires a higher affordable-unit share and higher construction wages once a project exceeds 99 units. As a result, a very large number of projects are exactly 99 units. Construction 3D-printing firm Icon raised $200 million at a $1.4 billion valuation, below its reported 2022 valuation of nearly $2 billion. Potter says the round still makes him somewhat more optimistic about the company.
This week's Construction Physics reading list from Brian Potter is a roundup of links on buildings, infrastructure and industrial technology, and it carries several data points that matter for the energy transition. The headline numbers come from China's solar industry and Europe's transformer supply chain. Around them are signs of how AI demand is straining other parts of industry. Roughly two-thirds of the list is paywalled. That includes a promised item on grid-scale battery reliability, which isn't covered here because its contents aren't in the free text.
China's solar makers are shrinking fast
Citing Reuters, Potter reports that employment at China's five largest solar manufacturers fell 21% last year and 31% the year before. These cuts compound. Taking an illustrative starting headcount of 100, a 31% cut leaves 69, and a further 21% cut leaves about 55. So within two years these companies have shed roughly 45% of their workforce.
The background, which the item assumes, is overcapacity. China makes the large majority of the world's polysilicon, wafers, cells and modules. Over the past few years manufacturers built factories much faster than global installations grew. Module prices fell steeply, often below production cost, and the industry has been in a price war. Potter's single word for it is "ruthless."
This cuts two ways for anyone tracking the transition. Developers buying panels benefit: hardware is a shrinking share of what a solar project costs. But an industry losing money and staff at this rate is consolidating, and the list doesn't say which firms will survive or whether prices will rebound once capacity leaves the market. Headcount also isn't output. Factories can cut jobs through automation or by closing older lines while still shipping large volumes, and the item doesn't separate those causes.
The transformer steel bottleneck isn't just American
Transformers step voltage up and down across the grid, and they've become one of its tightest constraints. Rising demand, aging equipment and data center build-outs have pushed delivery times out to years in many markets. One upstream input is 1. It's a silicon steel rolled and annealed (heat-treated) so its crystal grains line up, which lets magnetic fields pass through easily in one direction with little energy lost as heat. The transformer core is built from stacked GOES sheets, and the steel's grade largely sets how efficient the transformer is.
An IEEE Spectrum article quoted in the list makes a point usually made about the US: very few companies make GOES. In the EU there are only two, Thyssenkrupp and Stalprodukt, and like US producers they can't make the highest-quality grades. "The global leaders—the companies that make the world's best GOES—by and large are in [South] Korea and Japan," said an expert identified as Alimento. US and EU producers can't match that quality "because it is a heavily protected trade secret."
The implication is that tariffs or subsidies alone can't quickly build domestic supply of top-grade core steel in either region. The know-how is concentrated in a few Asian firms. The excerpt doesn't say how large the quality gap is in efficiency terms, or how much of Europe's transformer output depends on imported steel.
AI demand is straining chips and memory
Several items show where AI data center construction is pushing on supply chains that grid planners also watch:
2, the Dutch firm that makes the lithography machines used to pattern advanced chips, says it can't meet customer demand for 2027 and 2028 and is looking for ways to add capacity.
Memory shortages caused by AI build-out have cut smartphone shipments 20% year over year, according to Counterpoint.
Potter also notes that he's quoted in a Dispatch article about data center construction.
Housing and construction notes
The rest of the free section is about building:
What's left open
This is a link roundup, not an analysis, so most items come without the detail needed to judge them. The battery reliability item is behind the paywall. The solar job cuts lack a breakdown by cause. And the GOES item asserts a quality gap without measuring it. Read the list as a set of pointers worth following, not as conclusions.
Remember
Compounded, the two years of cuts leave China's top five solar makers with about 55% of their former staff.
Only Thyssenkrupp and Stalprodukt make transformer steel in the EU. The best grades come from Korea and Japan.
The free text skips a paywalled item on grid-scale battery reliability.
GlobalFoundries, the chipmaker spun off from AMD in 2009, will run two new 3 (chip factories) planned for New York, funded by $21 billion from Japan. The US government is also taking a 4% equity stake in the company.
A suspicious discontinuity. New York City's "4" tax break goes to new projects that set aside some units as affordable. Above 99 units, both the required affordable share and the mandated construction wages rise. As a result, a very large number of projects are exactly 99 units. It's a textbook case of a threshold rule shaping what gets built.
Robotic prefab. DR Horton, the largest US homebuilder, signed an agreement covering more than 1,250 homes with Promise Robotics, a Canadian startup. Promise says it uses AI to turn existing home designs into production plans and robot instructions, then coordinates factory manufacturing with on-site installation. Potter says one possible future is that AI plus robotics "changes the logic of prefab construction and makes it more attractive," and he's watching whether deals like this bear that out.
Icon's 5. Construction 3D-printing startup Icon raised a $200 million 6 at a $1.4 billion valuation, down from a reported valuation "approaching $2 billion" in 2022. Potter had assumed the company was struggling. He says the raise still nudges his estimate of its chances upward.
Immigrant labor. Harvard's Joint Center for Housing Studies found that from 2013 to 2024, the number of foreign-born construction laborers and carpenters rose sharply while native-born workers in those jobs declined. For electricians and plumbers, which are licensed trades, most of the growth came from native-born workers. That matters for electrification, which depends heavily on electricians.
