The 45ZCF-GREET Model Was Updated. The Field-Evidence Question Wasn’t.

On June 12, the Department of Energy released an updated 45ZCF-GREET model for the §45Z Clean Fuel Production Credit. For the renewable fuels industry, this matters. Ethanol producers, renewable diesel and SAF producers, tax advisors, credit buyers, and insurers have been waiting for a calculation framework that reflects post-2025 statutory changes and gives the market a more stable basis for pricing, diligence, transfer, and claims preparation.

The trade coverage has rightly called the release a step toward certainty. That is true. But it is worth being precise about what kind of certainty arrived.

The update gives the market a cleaner calculation engine. It does not, by itself, resolve the upstream feedstock-evidence problem. It does not independently verify what happened on a field. And it does not create a working user-facing path for no-till, reduced tillage, cover crops, crop rotation, or nutrient management claims.

GREET answers the emissions-rate question. It does not answer the field-evidence question.

What the update actually delivered

The new model reflects several changes the industry had been waiting on: indirect land-use change excluded from post-2025 carbon intensity results, North American feedstock eligibility restrictions, retired pathways for generic animal manure in favor of species-specific treatment, updates across RNG, corn stover, and wet-mill corn ethanol pathways, and, critically important, results presented separately for fuel produced in 2025 versus fuel produced after December 31, 2025.

These changes are not minor. They change how the market thinks about eligibility, pathway selection, credit value, and risk. They give fuel producers a firmer basis for decisions they had been deferring.

But the model release is one component of the §45Z architecture, not the whole of it. The industry is still waiting on final USDA feedstock guidance and final Treasury and IRS regulations. Which means the most consequential upstream question is still open: when a feedstock’s carbon intensity depends on agricultural practice claims, what evidence will be required to support those claims?

Three layers, and they should not be collapsed

A review of the updated workbook and DOE guidance points to a three-layer structure.

The first layer is 45ZCF-GREET: the emissions-rate calculator. It takes facility and pathway data and produces lifecycle emissions results for covered fuel pathways.

The second layer is USDA FD-CIC: the feedstock carbon-intensity calculator. This is where farm-practice treatment belongs, once USDA’s guidance is finalized and operationalized.

The third layer is evidence: the layer that determines whether the upstream practice facts behind a feedstock CI claim are independently supportable.

A GREET result can be clean while the field evidence behind it is untested. An FD-CIC output can return a feedstock CI value while the market still needs to know whether the practice inputs behind that value are defensible. A transfer file can carry a number while the buyer or insurer still asks: what supports the underlying field-year claim?

That is the unresolved issue, and it does not sit inside the calculator.

What the workbook itself shows

Our review of the updated model surfaces a useful contrast.

The facility-side calculation path is present and operationally represented. The model exposes pathway and process inputs, including energy use, production volumes, electricity treatment, RNG/CMM, and CCS selections that can produce a lifecycle carbon-intensity result when the workbook is run in the proper Excel/GREET dependency environment.

The upstream agricultural layer does not compute. The model references USDA FD-CIC, but not as an active, user-facing farm-practice workflow. There is no working input path inside GREET for no-till, reduced tillage, cover crops, crop rotation, or nutrient management. That treatment is deferred to USDA’s feedstock calculator which, as of this release, has not been finalized.

That contrast is the point. On the fuel side, 45ZCF-GREET provides the operational calculation path: facility and pathway inputs can be entered, and the model can produce lifecycle carbon-intensity results when run in the proper Excel/GREET environment. On the upstream agriculture side, the workbook shows something different: FD-CIC / CSA hooks, but not an active farm-practice calculator. The model contains a template question asking whether to model feedstock CI with USDA-FD-CIC, but the current setting is “No,” the dashboard target is blank, and there is no working input path inside GREET for no-till, reduced tillage, cover crops, crop rotation, or nutrient management. In practical terms, GREET appears prepared to accommodate a separate USDA feedstock-CI layer, but the climate-smart-agriculture contribution is not computed inside 45ZCF-GREET in this release. The added validation scaffolding points toward more formal submission-level controls, but it does not yet define the field-evidence standard. That is exactly why disciplined, independent evidence files will matter.

ILUC is a good example of why precision matters

The coverage describes the update as eliminating ILUC. The more precise statement: ILUC is excluded from applicable post-2025 results, while the model still retains ILUC structures for 2025 treatment and reporting.

That distinction matters because §45Z is now year-sensitive. A fuel produced in 2025 and a fuel produced after December 31, 2025 are not the same risk object. A buyer or insurer needs to know which production-year rule set applies. A producer needs to know which output is relevant. A diligence file needs to preserve the model version, the production year, the input set, and the result table.

The question is no longer just what is the CI score. It is: which model version, which production year, which inputs, which assumptions, which output, and which supporting evidence.

The policy direction is toward specificity

The movement away from generic animal-manure treatment toward species-specific pathways is a signal worth reading. For manure, the model now wants specificity. For imported feedstocks, it wants origin specificity. For post-2025 fuels, it wants year-specific treatment.

The same logic should apply upstream. If a feedstock CI claim depends on conservation tillage, cover crops, or crop rotation, the market should not lean on generic practice claims detached from field-year evidence. It should ask what practice occurred, where, when, and whether the claim can be independently supported.

The guidance offers one more signal in the same direction. For RNG, gas energy attribute certificates (effectively book-and-claim) are not available for §45Z use absent a future Treasury determination. The RNG book-and-claim decision does not foreclose every book-and-claim concept. But it shows Treasury will say no when an attribute-accounting mechanism drifts too far from physical or evidentiary substantiation. That rationale matters for the agricultural side, where book-and-claim has been proposed as one way to monetize climate-smart agriculture attributes. If CSA attributes are separated from physical feedstock flows, field-year evidence, and auditable substantiation, Treasury may ask the same question it is already asking elsewhere: what keeps the attribute connected to the underlying fact?

Why this matters: to farmers first

This is not an anti-farmer argument. It is a pro-honest-farmer argument.

The farmer who has run no-till for years should not be pooled with an unverifiable claim. The producer who actually planted a cover crop should not be averaged into a generic risk bucket. The operation whose crop rotation is visible in a field-year evidence stack should be able to distinguish that evidence from an assertion.

If the market cannot separate strong evidence from weak evidence, real practice adoption is undervalued. Independent observation helps fix that. It does not replace farmer records; it strengthens them, by adding a second signal that is not generated by the claimant and not dependent on paperwork alone.

And to buyers and insurers

For buyers and insurers, the issue is defensibility. A clean GREET output proves a calculation was performed. It does not prove every upstream input is independently supportable.

Plant-side data is strong: gas is metered, electricity is billed, volumes are recorded. Upstream practice evidence is a different problem: spatial, seasonal, field-specific. That scenario is where independent field-year evidence earns its place.

Where KVASIR sits

KVASIR is not a replacement for 45ZCF-GREET. It is not a replacement for USDA FD-CIC. It is not a tax opinion, a CI verifier, or a credit broker.

KVASIR is an upstream evidence layer. Its purpose is to produce independent, reproducible field-year evidence for observable practice claims (conservation tillage, cover crops, crop rotation) before those claims enter a feedstock CI workflow, a buyer diligence file, an insurance review, or an audit record. For observable practices, independent satellite-based evidence can separate what is supportable, what is not, and what remains unresolved.

That last category is the one most often mishandled. Unresolved acres should not be converted into favorable treatment because the market wants a larger number. They should be disclosed, carried separately, and treated conservatively. That is how evidence stays useful.

The June 2026 update is good news. It gives the market a better engine and aligns outputs with post-2025 law. It also makes the remaining gap easier to see.

The model arrived. The field-evidence standard did not.

Measure first. Settle second. Monetize third.