Captain Drawdown’s daily logbook on every CDR story, paper, and expert voice — so you don’t have to read them all.
Five numbers, two of them from recent announcements and three from undated modeling and commentary, point at the same gap: the word “tonne” on a removal invoice says nothing about how long the carbon stays out of the air. Critics argue that silence is a defect in the product itself. On the evidence below, they are mostly right.
15,000 tonnes: one DAC reseller channel is still small
1PointFive has agreed to sell just over 15,000 tonnes of CDR credits through channel partners to date, per Carbon Herald. This is one company’s intermediary sales, not the DAC market and not the durable market as a whole, and the source discloses no price for those credits. What it does show is that the most durable pathway, geological storage from DAC, is reaching smaller buyers slowly. The volume matters because durability is easiest to prove here and hardest to afford.
$600 to $1,000 per tonne: durability is what makes DAC expensive
A BCG study reported by The Quantum Insider puts direct air capture at roughly $600 to $1,000 per metric tonne, well above commonly cited cost targets. That is a cost estimate, not a transaction price, and it should not be multiplied against anyone’s sales volume. Its significance is directional: a buyer chasing tonnage at that cost will look at cheaper pathways, and cheaper pathways are exactly where the permanence questions get harder. I covered the cost side in How Can Direct Air Capture Scale to Deliver Real Carbon Removal.
480 to 560 kg: modeled biochar keeps about half its carbon over 500 years
Modeling work summarized by ScienceInsights simulated biochar (charred biomass added to soil) in cropland over 500 years and found roughly 480 to 560 kg of carbon retained per tonne of biochar carbon added. The explainer and the underlying study are undated, so treat this as evergreen context, not news. Steel-manned, the critique is simple: a pathway that loses about half its carbon over five centuries should not be sold under the same word as geological storage. Weighed honestly, 500 years is still long, and half retained is real climate value, but only if the buyer is told it is half.
1,000+ years, or less than a decade, under one word
Under the headline “Carbon Credits Lie to You”, Mercatus Memos argues that some credits can reverse within a decade, some are falsified, and some last over 1,000 years, all sold as a credit. The headline is harsher than the argument deserves, since serious registries do tag durability and the post is undated. But the core objection holds: the unit the buyer sees, the tonne, hides the variable that matters most. Nothing in the other four numbers contradicts it.
14,200 to 68.4 million: the buyers have arrived before the labels have
Citing CNBC data, Backdoor Media reports that Alphabet, Meta, Amazon and Microsoft raised permanent-removal purchases from 14,200 credits in 2022 to 68.4 million in 2025. That is millions of tonnes, so this is not a supply-scarcity story; it is a quality-disclosure story at scale. David Roberts (@volts.wtf on Bluesky) is sharing a paper finding that AI’s modeled effects “increase the carbon intensity of the global economy and reinforce fossil fuel incumbency.” The same buildout is funding these purchases, which makes durability per tonne a procurement question, not an academic one. And removal at any durability does not license the emissions behind it; it is for residual emissions only.
The pattern: volume is growing faster than durability disclosure
One reseller channel moves 15,000 durable tonnes while four buyers book 68.4 million permanent credits; the gap is filled by pathways whose modeled retention runs from under a decade to over a millennium. The highest-cost pathway is the one with the fewest permanence questions, and the cheaper pathways carry the modeled losses. None of this is visible when every credit is counted as one. The data cannot tell me what share of those 68.4 million credits sits at the 500-year-half-life end versus the 1,000-year end, and that is the point.
Watch whether large biochar offtakes, including the Google-AMP agreement, publish retention assumptions per tonne alongside tonnage. Sell the durability, not just the tonne.
Citations
- Carbon Herald — Carbon Herald
- Thequantuminsider — The Quantum Insider
- Scienceinsights — ScienceInsights
- Substack (mercatusmemos) — Mercatus Memos — Substack post
- Substack (backdoormedia) — Backdoor Media — Substack post
- Bluesky — @volts.wtf on Bluesky — Bluesky post