After 50 dispatches from the new infrastructure frontier, the pattern is unmistakable: the future gets built by whoever can survive a banker’s questions, and the best raw material is the stuff we pay to throw away.
The silence that costs a billion dollars
Picture a glass conference room on the thirty-first floor. It’s a diligence meeting after all the engineering and chemistry analysis has been done. The founders have done everything right. The technology works. The first stage ran eleven months without a hiccup. The deck is gorgeous.
Then a lender’s associate, young and unhurried, asks four questions.
Who is contractually obligated to buy what you make, and for how many years? Who guarantees the feedstock, and at what price? When does the grid let the project plug in? And what happens to this loan if these machines perform at sixty-five percent of its promised output?
The silence that follows is the most expensive sound in the world of energy.
This is the 50th dispatch in this series, and the technologies and ecosystems covered over this time could not be more different: hydrogen, water, fuel chemistry, dairy. But the endings rhyme. Projects rarely die because the science fails. They die in that room under simpler terms, as the ones that live walked in already holding the answers.
The grid is not coming to save you
The comfortable story goes like this: demand is surging, so the commercial grid needs to expand to meet it, and every clean-energy project will simply plug in. The numbers tell a more tangled story.
Start with demand. Earlier this year, the U.S. Energy Information Administration (EIA) forecast that electricity use will grow 1% in 2026 and 3% in 2027. That would be the strongest four-year stretch since 2000 and the first four straight years of growth since 2007, driven largely by data centres.1
Now look at the waiting room on the supply side. At the end of 2025, roughly 8,200 projects were stranded in U.S. interconnection queues, representing 1,312 gigawatts of generation and about 749 gigawatts of storage. That is a staggering amount of ambition. It is also a staggering amount of waiting. For projects that came online in 2025, in the regions with available data, the median trip from first request to actual commercial operation took more than five years.
And most never arrive. Of the capacity that entered queues between 2000 and 2020, only 13% had reached commercial operation by the end of 2025. Three-quarters had been withdrawn.2
A five-year line is not a plan. It’s a guess about someone else’s priority list.
There’s a quiet opening in these numbers. Berkeley Lab’s count represents generation seeking to connect to the transmission grid, not distribution-connected or behind-the-meter projects. The argument of this series, and it is the key argument, not the lab’s, is that infrastructure built where the waste is generated, producing what it sells on site, does not need a seat in the longest line in American energy.
Technology doesn’t get financed. Cash flows do.
The four questions in that glass conference room are not white noise. The Department of Energy’s Loan Programs Office, which exists to lend to projects that private markets will not touch, publishes its own version. Read the order. The financial questions come first, and the technical ones come second.
The first financial question is blunt: is there a reasonable strategy for selling what gets produced? The office says the nature of project financing is that the certainty of cash flows is optimal, and it ideally wants signed offtake contracts or committed letters of intent covering a significant portion of the project. Only then does it ask for 1,000 to 2,000 hours of successful operation and at least one, preferably two, maintenance cycles. It also expects 30% to 50% of the total capital to be committed equity, and says a loan covering 50% to 70% of project value is typical.3
Now watch what happens to a sector that skipped the first question. The International Energy Agency’s 2026 hydrogen review, as summarised by GreentechLead, shows the global pipeline shrinking by 10 million tonnes a year to 27 million, with more than 100 gigawatts of announced electrolyser capacity at risk by 2030 unless developers reach investment decisions by the end of 2027.4 The reporting is rather explicit about what gets cancelled: projects that lack customers, financing and competitive electricity. One example can be found in Australia’s flagship renewable hub, which shrank from a 26-gigawatt vision to a 1-gigawatt first phase. Woodside discontinued its Oklahoma sited project, due to inadequate economics and insufficient customer demand.
Notice what is missing from that list of causes: the chemistry.
Now notice who survived. Orica’s 50-megawatt Hunter Valley hub achieved its final investment decision (FID) because the hydrogen goes to the company’s own ammonia plant, a buyer standing next door. Over 85% of low-emissions hydrogen investment is flowing to refining, industry and hydrogen-based fuels, applications with established demand. The sector is not dying. Investment nearly doubled to roughly $7 billion in 2025, and IEA expects it to approach $10 billion in 2026.5 It is simply getting pickier, and the picky money follows the contract.
Waste is the new crude
Speaking rhetorically, crude oil is something you pay for. Waste is something other people pay you to take. That inversion is the whole idea.
Consider the gate. The Environmental Research & Education Foundation surveyed 342 landfills for 2023 and found the average U.S. tipping fee was $56.80 a ton, reaching $84.44 in the Northeast.6 Any feedstock that arrives with a fee attached is a different kind of raw material than one with a price tag.
Now consider the pressure behind it. The EPA found that 62.5 million tons of food waste were landfilled in 2020, more than double the 1990 figure and over 24% of everything landfilled. Wasted food accounts for 58% of the fugitive methane at municipal landfills. Between 1990 and 2020, total landfill methane fell 43%, while methane from landfilled food waste rose 295%.7 And the official numbers may flatter the problem. A Harvard-led satellite study of 2019 found landfill methane 51% higher than EPA estimates, and gas collection systems capturing closer to 50% than the 75% the agency assumes by default.8
Policy is starting to treat this as a mandate, not a suggestion. California’s SB 1383 measures against a 2014 baseline of about 23 million tons of organic waste disposed, and sets a 2025 target of a 75% cut. CalRecycle projected the state would need to source-reduce, reuse or recover about 27 of 32 million tons in 2025, and the rules require that diverted material go to uses that reduce greenhouse gases, not just anywhere but the landfill.9
The proof that waste can be a business, not a burden, is older than any of this. The North American Renderers Association says its 32 member companies run 178 plants and, by its own count, upcycle more than 54 billion pounds of food loss and waste a year into pet food and feed, renewable fuels and fertilizer.10 That is a trade group’s figure, and we have not independently verified it. Even so, the same letter carries a warning worth heeding: subsidising new technologies simply move a waste stream from one recycler to another, which is not progress by this single act.
So waste is not free money. It is contaminated, seasonal and contested. But it does answer the second key question in that glass conference room: who guarantees the feedstock and at what price, in the most unusual way in energy systems. Sometimes the answer is the generator, and now the generator is paying its creator.
Walking in with the answers
Return to the glass room. Here is what a project that survives sounds like, question by question.
“Who is obligated to buy, and for how many years?” A signed contract, not a letter of enthusiasm. The DOE lender wants signed offtake contracts or committed letters of intent (LOIs) for a significant portion of output. The hydrogen projects with the head- winds and momentum have a customer next door.
“Who guarantees your feedstock, and at what price?” The best answer is a waste generator with a legal reason to be rid of it and a gate fee it already pays.
“When does the grid let you plug in?” The best answer is that the business case has no reasons to stand in the longest line in the first place. With a median of more than five years from request to operation, a project that makes its product on site, where the waste originates, stands outside that grid queue. It still has its own permits and connections to manage. But it is not waiting on someone else’s calendar.
“What happens at sixty-five percent?” The best answer should be an independent engineer’s verdict and enough run hours to bore a lender: 1,000 to 2,000 hours at full operation, one or two maintenance cycles, and a third-party engineer, paid for by the applicant, hired to hunt for technical risks that could threaten repayment.
None of this is glamorous. And that is the point. Fifty articles in, the lesson is not that great technology wins. It’s that great technology which does the boring paperwork and validation first gets built.
The next fifty questions
Four questions will decide whether any of this scales — and to what level.
Does the demand hold? The EIA’s growth forecast leans heavily on data centres. This is a forecast, not a receipt. If that build-out slows, the urgency behind the grid access argument changes with it.
Do the queue reforms work? Berkeley Lab says FERC Order 2023 and other reforms are being implemented, but it is too early to measure their full impact. If the line gets shorter, the advantage of standing outside it narrows. That should be stated now.
Does the hydrogen shakeout end on a deadline? The IEA’s review, as reported, says more than 100 gigawatts of announced electrolyser capacity should miss 2030 unless developers reach investment decisions by the end of 2027. That target date will sort the contracted from the imagined.
Do we start measuring waste honestly? Satellites are finding landfill methane above official estimates. Better measurement raises the price of doing nothing, and whether mandates like California’s get enforced will show whether a gate fee is a signal or just noise.
The grid isn’t saving us. Neither is the next breakthrough. What will is the routine and improved: a signed contract, a feedstock volume that somebody is paying to be rid of, and a lender’s associate who runs out of questions.
Fifty down. Tell us which questions you want answered in the next fifty.
The hydrogen transition waits for no one.
References and sources
U.S. Energy Information Administration, EIA forecasts strongest four-year growth in U.S. electricity demand since 2000, fueled by data centers, press release, January 13, 2026.
Lawrence Berkeley National Laboratory, Queued Up: 2026 Edition, published May 2026, data through the end of 2025. Covers transmission-level generation requests only.
U.S. Department of Energy Loan Programs Office, Questions All Applicants Should Ask Before Applying to LPO, page last updated March 17, 2025.
GreentechLead, Green Hydrogen Investment May Reach $10 Billion in 2026, but Over 100 GW of Projects Face Deadline, August 28, 2026.
Ibid.
Environmental Research & Education Foundation, Analyzing Municipal Solid Waste Landfill Tipping Fees, May 31, 2024 (2023 data, 342 facilities).
Waste Dive, EPA reports quantify methane emissions from organic waste, retire food recovery hierarchy, October 20, 2023. It summarizes EPA’s landfilling report.
Harvard SEAS, EPA underestimates methane emissions from landfills, urban areas, May 1, 2024, describing Nesser et al., Atmospheric Chemistry and Physics
CalRecycle, SB 1383 Final Statement of Reasons, October 2020.
North American Renderers Association, comment letter to the EPA, February 2, 2024. Industry-reported figures, not independently verified.


