Carbon removal learns to write a purchase order

The carbon-removal ledger cleared $190 million on September 21 across two transactions, and both arrived with the thing that separates a purchase order from a gesture: names attached to the commitments. Amazon's Reforestation Fund II launched at $150 million with Mombak and BNDES in the party list, while a corporate buying group signed a $40 million deal with every member disclosed — McKinsey, Meta, Microsoft, Bain & Company, Salesforce, Google, Symbiosis, REI. Japan Airlines announced its own agreement with Climeworks Solutions, the direct air capture developer, without a disclosed size, which makes those first two figures a floor rather than a total.

The $40 million contract is the more instructive of the two, because companies buying credits one at a time make a decision that lives beside a marketing budget and can be cut in a weak quarter, while a buying group that has already agreed on scope and counterparty has built a procurement function — and procurement functions re-tender, ask suppliers for delivery schedules, and compare cost per tonne across vendors. Group purchasing also hands a developer something individual buyers cannot: a demand base that money can be raised against, which is how removal stops being a footnote in a sustainability report and starts being a counterparty relationship with a calendar attached.

The $150 million on the Amazon side is the supply trade: Reforestation Fund II puts money into producing removal rather than buying it, and the party list carries the interesting part — Amazon, Mombak and BNDES on one side of the table, corporate capital and development capital funding the half of the market that has to be built before anything can be sold. Nobody purchases tonnes that nobody has paid to grow, and the balance sheets willing to fund trees years ahead of the first delivery are rarely the ones signing purchase orders. The numeral on the fund also implies a predecessor vehicle, whose size the coverage does not give, which leaves the second fund's $150 million impossible to read in proportion to the first.

Japan Airlines extends the buyer base into a different kind of company, and deserves more attention than its missing number suggests: a carrier's appetite for removal tracks an operation and a schedule, and an agreement with a direct air capture developer is the sort of supplier relationship that can migrate into an operating budget once first volumes land. What the announcement lacks is both a price and a volume, so it cannot be measured against anything else in the day's ledger, and the demand it promises remains a claim rather than a line item.

PWD's deal log recorded one more climate vehicle on the same date, a $63 million launch for Pulse Fund, and stopped short of describing the mandate as removal — a distinction worth holding, because a climate fund can hold assets that remove nothing at all, and blending the two categories is how a sector's totals outrun its deliveries. Add the day's three named commitments and you get $253 million, a figure that mixes a nature-based fund, a corporate purchase and a general climate vehicle, which is precisely the arithmetic the category should stop publishing.

None of the three removal disclosures carried a price per tonne, and the ledger can record that $190 million was committed without recording what it bought. Watch for the buying group to return with a second purchase, and for either side of the Japan Airlines agreement to put a figure on the contract. Whichever publishes a price first makes the rest of these numbers comparable, and turns a procurement habit into a market with a clearing level.