Start with the accepted product

Someone offers you a waste stream, a gate fee, and a machine that turns carbon into gas. Would you finance the fuel plant? I wouldn't start with the gate fee. I'd start with the product the customer will actually accept.

The chemistry can work. The investment case needs something harder: a connected system that keeps making the right product, from the material you can really obtain, at a cost the buyer can support.

My early verdict is a real bet for a well-defined, demonstrated system with credible contracts. Not yet for a project whose feedstock, cleanup, operating record, and customer are still separate assumptions. That distinction matters more than the reactor's temperature.

The full process belongs inside the boundary

First, what are we buying? Gasification uses heat and a controlled gasifying environment to convert carbon-rich material into a gas containing carbon monoxide and hydrogen. The mixture also contains other gases and contaminants. Its composition is not a universal constant.

Pyrolysis, gasification, and combustion aren't interchangeable labels. Their operating conditions and intended products differ. For this episode, follow the gasification route. We're not assuming that any process advertised as advanced recycling produces the same gas, or faces the same obligations.

The process starts at reception. Sort and prepare the feed, gasify it, remove contaminants, and condition the gas. Only then do you reach the additional synthesis, separation, and finishing steps needed for a particular chemical or transport fuel.

IEA Bioenergy's Task 33 devoted a March 2025 review to gas cleaning. Its central engineering problem is straightforward: downstream uses impose impurity limits. Producing raw gas doesn't prove that you've met them.

A gas that works in a heat application doesn't automatically qualify for a synthesis catalyst. A reference plant can be useful evidence and still be the wrong reference for your product. Ask what it actually produced, at what specification.

The September 2025 Task 33 status report separates applications and project states. That's the right habit. An operating heat plant, a planned fuel plant, and a paused project are not three equivalent demonstrations of a fuel process.

Sources: S1 · DOE: Thermochemical Conversion Processes, S3 · IEA Bioenergy Task 33: Gas Cleaning, March 2025, S4 · IEA Bioenergy Task 33: Status Report, September 2025

Three feedstocks, three evidence files

Now put three delivery trucks at the gate. One carries biomass. One carries prepared municipal waste. One carries shredder residue. Those descriptions are useful beginnings. None is a procurement specification, a carbon certificate, or a guarantee of performance.

Take biomass first. A consistent wood-processing residue and a variable agricultural residue can create very different handling and ash problems. Moisture, particle size, storage, and mineral composition matter. Calling both renewable doesn't make them interchangeable in the feed system.

Water is part of the delivered mass. If the contract prices wet tonnes and the process model assumes dry material, your spreadsheet has quietly changed the product you're buying. Drying also needs equipment, energy, and a destination for the water.

I would want representative sampling across the expected supply conditions, not just a convenient bag from one delivery. The contract should define what happens when the material moves outside the agreed envelope. Who rejects it, who replaces it, and who pays?

Municipal solid waste is the second case. The relevant feed for many projects is a prepared fraction, not everything in a collection truck. Sorting, removal of unsuitable material, size reduction, and sometimes drying belong inside the plant's economic boundary.

A tonne received is not necessarily a tonne fed to the reactor. If preprocessing rejects part of the delivery, you still need to handle that material. The model must show its mass, disposal route, cost, and any recoverable value.

A gate fee can help. But its value depends on the contract: accepted material, volume commitments, interruptions, rejected loads, and who remains responsible for the waste. Don't book an unconditional payment from a conditional service obligation.

And municipal waste can contain both biogenic and fossil-derived material. Carbon that once sat in a plastic product doesn't become biogenic when it passes through a gasifier. The policy companion will examine how that distinction affects fuel classification and eligibility.

The third case is automobile shredder residue, often called ASR. It's the remainder after shredding and material separation. It can contain polymers, rubber, foams, fines, and residual metals. Its composition depends on what entered the recycling operation.

Argonne's 2011 vehicle-only shredding study is useful precisely because it defines the material it tested. It isn't a current certificate for every shredder's output. Nor does it justify assuming that every residue stream has the same contamination problem.

For a shredder-residue proposal, I would ask for representative material analysis and a process-specific acceptance specification. Pay attention to the difficult fractions, including halogens, metals, and ash. The question isn't whether carbon is present. It's what accompanies it.

A high-energy fraction may look attractive. That doesn't settle cleanup, emissions, residue handling, or the carbon classification of the final fuel. Treat a promising heating value as one input to the test, not as the answer.

Sources: S5 · Argonne: All Auto Shredding, July 2011

The integration risk between vendor guarantees

This is where the clean diagram becomes a real plant. Contaminants must be transformed, separated, captured, or discharged through a permitted route. A cleaning system moves material into other streams. It doesn't make your responsibility for that material disappear.

Tar is one example. Depending on the system, it can require removal or reforming before downstream use. That brings equipment, utilities, maintenance, and operating constraints. You need performance evidence for the actual sequence, not a generic promise of clean gas.

Particulates, sulfur compounds, chlorine compounds, and trace species create different problems. The synthesis supplier's limits should match the gas-cleaning guarantee. If those two documents use different boundaries, units, or operating conditions, the integration risk belongs to someone.

Usually, the owner discovers the gap. One vendor guarantees a gasifier outlet. Another assumes a cleaner inlet. The plant can meet each carefully worded promise and still miss the product target. Evaluate the connected system and every interface.

Increasing temperature can change conversion and impurity formation, but it isn't a free repair. Utilities, materials, heat recovery, and operating limits still matter. Compare the complete design. Don't select a reactor by the hottest number in the brochure.

There are also two separate yield questions. How much feed becomes useful gas? And how much of that gas becomes a product that meets the buyer's specification? A strong answer to the first question doesn't automatically answer the second.

Methanol needs its synthesis and purification system. A liquid-fuel route needs its own conversion and finishing steps. Those downstream plants add capital, utilities, losses, and requirements. Selling syngas and selling finished fuel are different business cases.

Sources: S3 · IEA Bioenergy Task 33: Gas Cleaning, March 2025

Availability changes the cost denominator

So the denominator I care about is accepted annual output. Not peak gas flow during a test. Not gross tonnes received. And not the reactor nameplate multiplied by every hour in the calendar.

Here's a deliberately simple stress case, not a claim about any operating plant. Suppose availability falls from 90% to 60%. Keep the running rate, accepted yield, and annual fixed cost unchanged. What happens to fixed cost per unit?

It rises by 50%. 90 divided by 60 is 1.5. The machine hasn't become more expensive to buy. You have fewer accepted units over which to spread the same annual fixed burden.

That isn't a 50% forecast for total fuel cost. Variable costs can change, and repeated stops can bring additional losses or penalties. The calculation isolates one mechanism. A proper project model must then add the others.

If treatment revenue depends on material accepted, downtime can also reduce gate-fee income. So a model that keeps every gate-fee dollar while cutting fuel production may be inconsistent with its own contract. Check both sides together.

Build the cash bridge explicitly. Product receipts and enforceable treatment revenue, less feedstock, preparation, utilities, consumables, maintenance, residues, logistics, and the capital burden. Keep policy value visible as a separate line with its own evidence.

Compare like with like. A gas cost estimate at the factory is not comparable with the cost of delivered finished fuel. A historical study is not a current equipment quote. A price for wet material is not the same as a price for dry material. Put every proposal on the same basis before comparing them.

That's why cheap feed can become expensive feed. What matters is the cost of getting it into a reliable, usable form and through the whole plant. The difficult day belongs in the model, not only the day the demonstration ran well.

Five evidence gates before financing

Now the five gates. First, feedstock. Show representative sampling, an acceptance envelope, the preparation balance, and enforceable supply terms. If the supposed supply advantage disappears when those conditions are applied, redesign the project before buying the reactor.

Second, integrated operation. Show a relevant operating record for the proposed feed and connected process. Include interruptions, maintenance, consumable replacement, and recovery after a stop. A short successful test is useful. It is not a year of production.

Third, product and buyer. Get the specification in writing, demonstrate it, and reconcile it with the offtake contract. The buyer's tolerance, delivery requirements, pricing, and rejection rights are part of the technology's commercial boundary.

Fourth, residues and permissions. Account for every solid, liquid, and gaseous outlet. Alberta's waste-to-energy guidance makes the approval boundary explicit: activities are regulated under the Environmental Protection and Enhancement Act and its regulations. A conversion label isn't a permit.

Fifth, downside economics. Stress availability, usable yield, preparation, residual disposal, product price, and policy support together where they are connected. A good feed contract cannot compensate for an unproven product specification. Don't average away a failed gate.

If you want a score, grade the evidence from one to five. One is an assertion. Three is partial, relevant evidence. Five is demonstrated performance supported by enforceable terms. Those are diligence labels, not measured ratings for entire feedstock categories.

Sources: S6 · Alberta: Waste-to-energy facilities

Advance, redesign, wait or stop

My biomass verdict is conditional. A consistent, well-characterized stream can simplify the task. That doesn't make every biomass plant bankable. Supply, moisture, ash behavior, product demand, and the integrated operating record still have to clear the gates.

For prepared municipal waste, the opportunity includes a waste-service business as well as a fuel business. Underwrite both. Sorting losses, treatment obligations, plant interruptions, and product acceptance can't be left outside the model merely because the feed arrives with a fee.

For shredder residue, I would require stream-specific proof before assuming the route is investable. Compare gasification with the actual available recovery and disposal alternatives. Neither a generic rejection nor a generic clean-fuel label is a substitute for that work.

Advance when the evidence and contracts fit together. Redesign when a feed, cleanup, product, or site change can resolve the weakness. Wait when decisive evidence is missing. Stop when the credible downside and the available alternatives don't support the investment.

What would change my mind about a weak proposal? A relevant integrated operating record, independently checkable product results, and a contract model that survives a sensible downside. A new rendering of the plant wouldn't do it.

The companion episode asks a different question. Even if the plant can operate, what can its product legally claim, and which revenue can it actually earn? Waste classification, fuel pathways, carbon accounting, and certification can change that answer.

Use the evidence checklist

Open the free interactive checklist and printable worksheet. Record the evidence, its date, the owner of each gap and the fact that would change your decision. No email is required. Entries are not transmitted or saved automatically; print to retain them.

Read the companion: Waste-to-fuel market access and policy revenue.

Scope and source dates

The analysis was prepared from the 26 October 2026 source review; the IRS notice, Canadian consolidation status and China consultation notice were rechecked on 27 October. Currency, amendment, effective date and project eligibility remain separate. Historical technical studies are identified by their dates and do not certify a current plant. This is general analysis, not project-specific engineering, legal, tax or investment advice.