In 2023, U.S. CO₂ suppliers reported sending 4.59 million metric tons of carbon dioxide into food and beverage manufacturing. Total supply that year was 42.90 million metric tons, and food and beverage ranked as the second largest end use after enhanced oil and gas recovery. Those figures come from EPA's Greenhouse Gas Reporting Program.
They exist because of one short section of federal regulation, 40 CFR Part 98 Subpart PP, Suppliers of Carbon Dioxide. In September 2025, EPA proposed to remove it. On October 14, 2026, the agency pushed the deadline for 2025 reports out to March 10, 2027 to give itself time to decide.
What Subpart PP Collects
Subpart PP covers facilities that capture a CO₂ stream to supply it for commercial use, CO₂ production wells, direct air capture units, and bulk importers and exporters. Producers face no size threshold. EPA's supplier table lists "all producers of CO₂." Importers and exporters report once their bulk volumes reach 25,000 metric tons of CO₂ equivalent a year.
That is a much lower bar than the one 45Q sets. A facility that does not generate electricity has to capture at least 12,500 metric tons a year to claim the credit. An RNG plant putting 1,000 MMBtu a day into the pipeline vents roughly 38 tons of CO₂ a day and clears that floor by about eleven percent. A dairy digester at 100 MMBtu misses it by a factor of nine. A capture unit on that digester selling its CO₂ would still file under Subpart PP.
The report goes well past a single tonnage. Each supplier states the annual and quarterly mass through every flow meter, the quarterly CO₂ concentration of the stream, where each meter sits relative to capture, dehydration and compression, the standard used to calibrate it, and how many days ran on substitute data.
Where the supplier knows the destination, it reports tonnage to fourteen end-use categories, and food and beverage is first on the list. Each capture unit also reports what percentage of its stream was biomass-based that year.
Taken together, those lines form the only federal dataset on how much CO₂ the country supplies, where it comes from, and roughly where it goes.
Where the Rule Stands
EPA signed the reconsideration proposal on September 12, 2025 and published it September 16. It would end reporting for nearly every source category after reporting year 2024, keeping only most of the oil and gas sector, whose reporting would resume in 2034. It groups the four carbon capture subparts together (PP for suppliers, RR for geologic sequestration, UU for injection, VV for storage through enhanced oil recovery) and proposes to remove all four.
The agency's reasoning is legal. It acknowledges that Treasury and the IRS reference GHGRP data in the 45Q regulations, and that Treasury has asked whether Subpart PP data could substantiate capture amounts. It then says that use falls outside the Clean Air Act authority it relies on to collect the data.
The proposal also said nobody would need to file for calendar year 2025. EPA received more than 50,000 comments. A February 27, 2026 rule moved the 2025 deadline from March 31 to October 30, 2026. The October 14 interim final rule moves it again, to March 10, 2027, and changes nothing else. Until a final rule is published, Subpart PP stays in force.
What the Count Has Shown
Food and beverage tonnage reported under Subpart PP ran between 4.13 and 5.00 million metric tons a year from 2016 through 2023. EPA withheld 2020 because the aggregate did not meet its confidentiality criteria. The 2023 figure of 4.59 was down from 5.00 the year before.
The same table shows how the national supply is built. Of the 42.90 million metric tons supplied in 2023, 26.80 million came from natural CO₂ sources and 16.10 million was captured at industrial plants. About 62 percent of U.S. supply came out of the ground.
Two limits apply. End-use tonnage is reported only "if known," so a capture plant selling to a distributor may never learn the final use, and the food and beverage figure is best read as a floor. The dataset also covers volume and origin only. Nothing in it describes purity.
What Changes for a Beverage Buyer
If EPA finalizes the proposal as written, 2024 becomes the last year of national CO₂ supply data. Anyone tracking whether food and beverage demand is growing, or how much supply now comes from capture instead of natural wells, loses the only public series that answers either question.
The biomass-based percentage goes with it. That line was the one place a federal form asked a capture plant how much of its CO₂ came from biological carbon. Without it, a biogenic claim rests entirely on the supplier's own evidence, which in practice means a radiocarbon result under ASTM D6866. Fossil carbon has lost its carbon-14, so the isotope ratio separates a digester stream from a fossil one regardless of what a contract says.
Utilization claims under 45Q are less exposed. Under IRS Notice 2024-60, a utilization credit is sized by a lifecycle analysis report reviewed by the IRS and DOE, and that process does not run through Subpart PP. Storage claims are the ones tied to GHGRP sequestration plans and annual data, and that link to Subpart RR is what carbon capture groups pressed in their comments on the proposal.
The Record Moves to the Supplier's File
Subpart PP asked capture plants to name their meters, their calibration standards, and the concentration of what they sold. A buyer qualifying a CO₂ source can ask for the same things, and should, whatever EPA decides by March.
- Where is the custody flow meter, and what standard is it calibrated to?
- What was the stream's CO₂ concentration, quarter by quarter?
- What share of the carbon is biogenic, and which lab measured it?
- Which certificate of analysis goes with each load, and against which specification?
The last question was never part of any federal form. Beverage grade CO₂ is certified against the International Society of Beverage Technologists guidelines: 99.9 percent minimum purity and roughly twenty individual impurity limits, with benzene held to 20 parts per billion. That is the spec CleanCycleCarbon certifies to, and every load carries its own certificate.
A federal count of tons can disappear with one final rule. The certificate on each load stays, because the bottler asks for it at the dock, and it has always been the document that decides whether CO₂ goes into a drink.
