$6 a Gallon for Manure, Near Zero for Landfill Gas: What the New 45Z Rules Actually Pay

Divide the $1.00 per gallon maximum by the 50 kilograms of CO2e per mmBtu that the statute uses as its zero line, and you get the only number that governs the 45Z clean fuel production credit: two cents. Every kilogram of carbon intensity a producer can argue off a pathway is worth two cents per gallon-equivalent. Every kilogram that can't be defended costs the same. On 8 September 2026 the IRS published Notice 2026-53 with the 2026 emissions rate table, and what it actually published was a price list for modeling assumptions.

The mechanics fit in three lines. Credit equals an applicable amount times an emissions factor, and the factor is 50 minus your carbon intensity, over 50. The base amount is 20 cents a gallon, 35 cents for aviation fuel, multiplied by five if you clear prevailing wage and apprenticeship. Land above 50 and you get nothing (it's an eligibility threshold, not the shallow end of a slope).

So the feedstock question most operators are asking is the wrong one. This credit doesn't pay for converting waste into fuel. It pays for the gap between what your methane would have done and what it did, and that gap is set by regulation rather than by chemistry. Two organic streams with near-identical gas composition can price an order of magnitude apart under the same table.

Manure Is the Only Feedstock Allowed Below Zero

The 2025 tax act banned negative emissions rates outright, then wrote one exception: transportation fuel whose primary feedstock is animal manure. Notice 2026-53 puts dairy and swine rates into the table and confirms the exception survives into 2026 production.

An emissions factor can therefore exceed 1.0, which is the part nobody outside the dairy sector has fully absorbed. A dairy pathway landing near minus 250 kg CO2e per mmBtu, the kind of range dairy projects have carried under California's low-carbon fuel program, produces a factor above six and a credit north of $6 per gallon-equivalent at the full rate [modeled estimate, LCFS-range CI applied to the statutory formula]. There is no cap on the upside, only on the CI you can defend. For scale: at the full rate a merely zero-CI fuel is worth roughly $8.60/MMBtu in credit alone [author arithmetic on the statutory formula], which is a multiple of what the gas fetches as a commodity. The molecule is the byproduct here; the avoided methane is the product.

Which assets collect is settled already, and not evenly. EPA's AgSTAR program counted 400 operating anaerobic digesters on US livestock farms as of June 2024: 343 dairy, 50 hog, nine beef and eight poultry. Another 73 were under construction or in modification. Compressed gas end-use went from five projects in 2017 to 134. So the two species with rates in the table are the two species that built out first. The eight poultry systems and nine beef systems are waiting on a model that doesn't describe them yet.

Landfill Gas Lost Before the Model Ran

EPA's Landfill Methane Outreach Program counted 542 operational landfill gas energy projects as of September 2024, about a fifth of them making pipeline-quality gas. On volume that's the biggest waste-derived fuel pool in the country. Under this credit it's close to worthless, and gas quality has nothing to do with it.

Any MSW landfill above the regulatory size threshold is already obliged to collect and destroy its gas under 40 CFR Part 60 Subpart Cf. Destroying methane you were legally required to flare isn't an avoided emission. It's compliance. The landfill pathway therefore carries a positive carbon intensity assembled from upgrading parasitic load, tail gas and compression, with no negative baseline term to net against it. Run that through the same formula and the factor collapses toward the floor.

And then there's what the meters find, which is a separate problem. According to a Colorado State METEC field campaign, nine landfill RNG plants across Kansas, Oklahoma and Texas lost 1.8 percent of their methane on average in 2021, and the worst performer in the sample lost 5.7 percent. Plants that vented tail gas directly ran 4 to 6 percent. The ones that didn't stayed under 1 percent, a spread the authors put down mostly to whether the tail gas had anywhere to go. At two cents a kilogram, the distance between a vented and an oxidized tail stream off a Honeywell UOP membrane skid is the difference between a revenue line and a rounding error, and it's a capital item, not a modeling argument.

That measurement burden is going to land on operators who have never had to defend a plant-level loss rate to a tax examiner. Continuous emissions logging, the kind of waste intelligence software that timestamps every vent event rather than reconstructing it from a monthly gas balance, stops being an efficiency project and starts being audit defence.

Food Waste Is the Contested Middle

Source-separated food waste sits between the two, and its carbon intensity runs on a regulatory clock. EPA's national materials accounting put food at roughly 24 percent of everything sent to landfill, the largest single landfilled category. Where that material would otherwise rot in a cell, a digestion pathway earns an avoided-methane term. Where a state organics mandate has already pulled it out, it doesn't, because the counterfactual has become a windrow.

California's organics rules are the clean example of the paradox: the better a state diverts organics, the less a 45Z clean fuel production credit is worth on that state's food waste. Actually, that's too tidy. The erosion only bites where the mandate is genuinely enforced, and enforcement, not the statute, is the variable that decides how fast a food waste pathway's CI drifts upward. Either way, the landfill diversion programs a municipality is proudest of are quietly repricing the fuel credit on the material they collect.

Food waste also carries a contamination discount that manure doesn't. Plastic film and packaging fragments raise the energy cost per unit of gas and show up in the model as process emissions.

What Isn't Priced Yet

USDA's manure carbon intensity rates are still to come, expected later in 2026, with poultry and beef pathways alongside them. So the $6 figure above is arithmetic on a modeled CI, not a published rate. Treasury has signalled the model will be updated; it hasn't said what the numbers will be, and the gap between those two statements is where a year of construction schedule goes.

A carbon-credit issuance audit for an Asian facility, which I ran in 2023, turned on exactly this problem: the pro forma and the registry were working from different vintages of the same methodology. Nobody was wrong, exactly. The gap surfaced as a haircut at review, months after the financing closed. 45Z has that shape. The model version is a deal term, and any term sheet that doesn't name the 45ZCF-GREET release it assumed is leaving a haircut unpriced.

Three other changes in this guidance move money, and each deserves a line. Indirect land use change is excluded from the calculation, worth a great deal to crop feedstocks and nothing whatever to waste streams. Canadian and Mexican feedstock eligibility came back after the 2025 restrictions, which reopens cross-border manure and organics flows that had gone quiet. And the 45Z clean fuel production credit now runs through fuel sold on or before 31 December 2029, four years later than the original sunset.

None of this holds evenly. Below roughly 1,000 milking cows the interconnect and gas conditioning cost per unit swamps the credit, and a small digester that can't reach a pipeline is selling electrons, not fuel. A producer who doesn't clear prevailing wage and apprenticeship is working from the 20-cent base, which turns a $6 pathway into something closer to $1.25 and changes the financing entirely. Projects burning gas on site for power aren't in this credit at all, they're in the clean electricity regime, and we covered which waste projects actually qualify for those separately. The registration, measurement and recordkeeping load is also real: this is claimed by the producer, per facility, per fuel, and the paperwork failure mode is losing a year of credit on a registration technicality rather than on a modeling dispute.

What the Next Three Years Look Like

The window is narrow. USDA rates arrive sometime in 2026, the credit stops applying to fuel sold after 2029, and a farm digester takes eighteen to thirty months from term sheet to gas flow. Call it three production years of priced certainty for an asset class that needs most of that time just to get built.

Capital follows the arithmetic, and the arithmetic says dairy and swine. The 73 systems already under construction will be joined by a wave underwritten on the negative-rate exception, and every one of those financings will hinge on a CI number that hasn't been published yet. That's an uncomfortable place to sign debt documents. But how much of that 2027 pipeline is genuinely bankable before the rates land, as opposed to bankable on a letter from a consultant?

Landfill gas projects will keep getting built on low-carbon fuel standard credits and RIN economics, the way they were before any of this existed, and their sponsors will treat 45Z as a small positive rather than a thesis. Food waste digestion is the one I'd watch, because its value under the 45Z clean fuel production credit decays fastest in exactly the jurisdictions doing most to divert organics. A credit that pays least where the policy works best will not survive contact with the next reauthorization. Somebody is going to spend 2028 arguing that poultry litter should have been in the 2026 table.

Sources & Notes

• The guidance itself: the IRS announcement of Notice 2026-53 and the 2026 emissions rate table, issued 8 September 2026, is the source for the dairy and swine rates, the exclusion of indirect land use change, North American feedstock eligibility, and the statement that poultry and beef manure pathways are expected in the model later this year.
• For the formula and the sunset, I worked from the statute rather than a summary: 26 U.S. Code 45Z gives the 20-cent and 35-cent base amounts, the five-times increase for prevailing wage and apprenticeship compliance, the 50 kg CO2e/mmBtu threshold and the December 2029 termination.
• Digester counts, the dairy/hog/beef/poultry split and the construction pipeline come from EPA AgSTAR's data and trends page, current to June 2024.
• Landfill project counts and the RNG share are from EPA's Landfill Methane Outreach Program. The food waste share of landfilled material is EPA's national materials accounting.
• The measured loss rates are from Colorado State University's METEC campaign at nine landfill RNG plants in the central United States. Worth reading before anyone signs an attestation about plant-level methane loss.
• The $6 per gallon-equivalent figure is my own arithmetic: a minus-250 kg CO2e/mmBtu carbon intensity run through the statutory factor at the full rate. It is a modeled illustration of what the negative-rate exception permits, not a rate anyone has published.