Scale Beats Efficiency When You Compare Energy Conversion Technology

NREL's 2023 comparison of thermochemical waste-to-energy pathways models the operating cost of a 50-ton-per-day combustion plant at $111.68/ton and a 3,000-ton-per-day plant at $22.42. Same reactor type, same fuel, nearly five times the operating cost per ton, and the only thing that changed was how much waste arrives at the gate. I've sat through nine-month technology selections where the committee argued about which energy conversion technology squeezes the most kilowatt-hours out of a ton, while the throughput assumption underneath all of it sat unexamined in an appendix.

That's backwards, and it's the most expensive habit in this business. The kilowatt-hour spread between the four main waste energy conversion methods is narrow. The cost spread across scale is wide. So is the gap in how much commercial operating history each path can lend you when a lender asks.

What follows is how I run an energy conversion technology screen on a live project: not technology by technology, but criterion by criterion, because that's the order a lender, a regulator and a host community will come at you in.

Cost per ton moves with tonnage long before it moves with the reactor

The rest of that ladder is worth seeing. According to the same NREL briefing, modeled operating cost sits at $62.82 a ton at 200 TPD and $43.85 at 500 TPD, so most of the improvement gets bought in the first few hundred tons of daily capacity. Capital behaves the same way: the modeled estimate for a 33,000-TPY combustion facility is $35 million to $39 million in 2020 dollars, and a plant ten times that size lands at roughly six times the money rather than ten.

Now put the efficiency numbers beside that. The same briefing gives mass-burn combustion about 544 kWh per ton of waste, conventional gasification 685, plasma gasification 816, and pyrolysis 571. Take the extremes and gasification looks like a 50 percent improvement on burning. At a US wholesale power price that gap is worth a few dollars a ton, which is less than one rung of the scale ladder, and it assumes the gasifier hits design output on your feedstock (which is the whole argument, and not one the table settles).

The assumption that moves on these projects is almost never the efficiency number. It's the tonnage. Waste volumes get pledged by a municipality with a four-year electoral horizon and a diversion target that, if it succeeds, takes tons away from you. A city's recycling plan and your fuel supply agreement are the same document read from opposite ends, and I have been on screens where nobody on the technical side had opened the first one.

Each path refuses a different part of the stream

Thermal vs biological conversion stops being an abstraction here. Mass burn takes mixed household waste roughly as the truck delivers it, then pulls metal and glass out of the ash afterwards. But gasification and pyrolysis don't. Both want inerts removed up front and the feed shredded, sized and often dried, because syngas and bio-oil quality tracks feedstock composition in a way that raising steam simply doesn't. Anaerobic digestion is fussier in the other direction: it wants the putrescible fraction and rejects the rest, which buys you a depackaging and screening line before you have converted anything at all.

Read that middle column as a capital cost, because that is what it is. Every material a path refuses is a machine you buy, house, power and maintain, plus a residue stream you now own and have to find a home for. A Vecoplan or Komptech primary shredder on an RDF line isn't the expensive part of the plant. Its wear parts, over a twenty-year operating life, can be. Anyone choosing waste conversion technology on a nameplate efficiency figure is pricing the reactor and ignoring the building it needs around it. Our deeper look at clean thermal processing beyond incineration works through the reactor side of that in more detail than I will here.

On a Caribbean island RDF facility in 2022 we came in nine percent over budget and still ahead of schedule, and the reason was unglamorous. We pre-positioned shredder spares and a spare gearbox on site before commissioning week. On an island, a wear part is a six-week lead time and a customs file, not a phone call. That line looked like padding in the capital review, and it bought back three weeks of schedule the first time a length of rebar went through the rotor.

Only one of the four has a settled federal air standard

Permitting risk isn't distributed evenly across these options, and it's the criterion developers most consistently price at zero. A large US municipal waste combustor is governed by a performance standard that has applied, with revisions, to units built after September 20, 1994 under 40 CFR Part 60 Subpart Eb. Dozens of plants have been permitted against it. So your consultant knows what the stack testing costs and roughly how long the state agency takes to turn a complete application around.

Gasification and pyrolysis sit in less settled ground. The question of when a thermal unit counts as an incinerator, and therefore which standard binds, has moved more than once in the past decade, and a permit writer facing a first-of-kind application in their state will take the conservative reading every time. Digestion mostly escapes the combustion rules and lands instead in state solid-waste permitting, odour review and nutrient management, which is a shorter clock on paper and a longer one in practice if the host community has opinions about truck movements. Eligibility for state renewable incentives splits along similar lines, and whether renewable energy from waste counts as clean power can differ between two states sharing a border.

NREL is blunt about the commercial gap: combustion has "40+ years" of US operating history, while gasification and pyrolysis have "limited commercial experience in the United States." A permit writer reads that sentence the way a credit committee does. None of which makes the newer paths wrong. It makes their schedule and their contingency different, and a feasibility model that gives all four the same fourteen-month permitting line is telling you a story. The permitting contingency I carry on a screen for pyrolysis systems is roughly double what I carry on a grate, and I've never regretted it.

The output decides who can sign your offtake

Four paths, four products, four unrelated sets of counterparties. Combustion gives you steam and power, and sometimes a district heat customer, which is the dullest and most bankable output on the list. Gasification gives you syngas, burned on site for power or upgraded into fuels and chemicals, and that upgrading step is where the engineering gets interesting and the balance sheet gets thin. Pyrolysis gives you bio-oil, char and off-gases, each with its own buyer and its own specification to hit. Digestion gives you biogas, which either runs a genset or gets cleaned to pipeline quality.

The further your product sits from electricity on a meter, the more specialised the buyer and the fewer of them exist in your region. That's the trade nobody prices properly at feasibility, and I'd rather kill a project at the offtake stage than discover the gap at financial close.

Which brings me to the sharpest cautionary tale in this comparison. The only commercial-scale mixed-waste gasification facility NREL could point to in the United States was Fulcrum BioEnergy's Sierra BioFuels plant outside Reno, Nevada, built to take 175,000 TPY of waste into jet fuel. It started up in December 2022, hit equipment problems that delayed full-scale operation, shut down operations in May 2024, and filed for Chapter 11 in Delaware that September owing creditors close to $456 million, as Waste Dive reported. Actually, that case tests my own rule rather than confirming it. I tell clients most waste-to-energy projects die at offtake; Fulcrum died upstream of its offtake, on equipment and cash burn. Either way the warning is the same, because when the reference fleet for your chosen path is one plant deep, you are the reference fleet.

Biological conversion has its own version. EPA's AgSTAR database logs 87 manure-based digesters shut down between 2000 and 2022, against roughly four hundred still running, and the 3.29 million MWh-equivalent the programme recorded for 2023 came from the survivors. But nobody publishes a brochure about the 87. Landfill gas capture, meanwhile, is the baseline all four paths have to beat: EPA's landfill methane programme counted 542 operational energy projects and 444 landfills it rates as good candidates as of September 2024, and those sites already hold the gas, the land and the permit.

Matching a path to a project

So how do I sequence a screen? Tonnage, then feedstock, then offtake, and technology last. In practice that collapses into a short set of situations.

Above roughly 500 TPD of mixed municipal waste with a power or steam buyer in hand, mass burn on a proven grate wins, and this is the one case where the forty-year track record behind a Martin-style grate is worth more than somebody's better efficiency figure. A clean, consistent single-polymer or tyre stream with a chemical buyer at the other end points to pyrolysis, which is where all four of the US commercial plants in NREL's count sit. And not one of them runs mixed municipal waste. A high-moisture organic fraction with a gas or fertiliser outlet points to digestion, because sending wet food waste into a thermal unit means paying to boil water.

Two cases get decided before the technology does. If a closing landfill already has gas under collection, upgrading what's there usually beats building anything new on cost per ton diverted, and the landfill diversion credit is identical either way. And a first-of-kind gasification project only makes sense with a strategic sponsor who can absorb a two-year commissioning tail, plus a site visit to an operating plant of the same configuration somewhere in the world (if you can't find one to visit, that is itself the answer).

Scale decides which of these is even a conversation, feedstock decides which candidates survive the screen, and the offtake decides whether the survivor gets financed. Financing is why the diversion case and the energy case belong in one model rather than in two documents written by two teams who meet at the steering committee.

Where does this framework stop holding? It rests on two American conditions. The first is a gate fee you can contract for and enforce. Much of my regional market-entry work is in places where the tipping fee is collected politically rather than contractually, and once the kilowatt-hour becomes your only dependable revenue the ranking above inverts, because the kWh-per-ton column moves to the top of the screen where it never belonged in the US case. The second is a creditworthy power purchaser. Yet strip that out and none of the four paths finance on their energy product alone, whatever the efficiency table promises. If you're picking an energy conversion technology for a project in that kind of market, build the revenue side first and come back to this comparison afterwards, because the ranking you get will be a different one. The zero-waste-to-landfill solutions I get asked about in those markets run into that wall far more often than they run into a technical limit.

Sources & Notes

• Every cost figure, kWh-per-ton value, plant count and the operating-history language quoted above comes from Milbrandt, Kraft and Avery, Comparison of Select Thermochemical Conversion Options for Municipal Solid Waste to Energy (NREL/BR-6A20-86461, July 2023). The briefing's costs are in 2020 dollars, so escalate before they go anywhere near a model.
• One scope note on that briefing: it covers combustion, gasification and pyrolysis only. NREL handles digestion and landfill gas in a separate document, which is why the digester and landfill numbers here come from EPA instead.
• Fulcrum's May 2024 shutdown, the Chapter 11 filing and the creditor total are as reported by Waste Dive.
• Digester counts, shutdowns and 2023 generation: EPA AgSTAR data and trends. The agency notes its database isn't exhaustive and that shutdown years are sometimes unknown, so treat 87 as a floor. Landfill project and candidate-site counts are from EPA's Landfill Methane Outreach Program, current to September 2024.
• If you need national combustion and generation totals, EIA's waste-to-energy explainer is the usual reference. Be aware its plant count and NREL's disagree, because they count different populations: sellers of electricity versus commercial facilities including steam-only sites.

Disclosure: I'm a project developer at Renewable Waste Energy and I take fees from clients running exactly the screen described here. The Caribbean RDF facility goes unnamed because the commissioning terms aren't mine to publish.