Captain Drawdown’s daily logbook on every CDR story, paper, and expert voice — so you don’t have to read them all.
CarbonPlan published a review this week of Lithos Carbon’s first commercial enhanced weathering credit issuance, and the findings should stop every durable CDR buyer mid-contract. CarbonPlan’s analysis says Lithos’s gross removal numbers are surprisingly high and that the team could not reproduce them from the public record. This is the same group whose 2021 forest offset work forced a reckoning in the voluntary market. They are now aiming that same lens at enhanced rock weathering (ERW), the pathway most of the industry has been selling as DAC’s cheaper, near-term successor. The timing matters: ERW developers are the loudest voices at SF Climate Week right now, pitching ERW as the scalable answer after Microsoft’s high-profile pause on some purchases. A credibility challenge to the flagship issuance lands at the worst possible moment.
The mechanism, and where it breaks
ERW works by spreading crushed silicate rock (often basalt) on farmland. The rock reacts with CO2 dissolved in soil water, forms bicarbonate, and eventually that carbon ends up in the ocean where it stays for tens of thousands of years. The pitch is gigaton potential at low cost. The problem is that you cannot see the reaction happen. Developers have to infer it, usually by measuring cations in soil samples and running a model.
CarbonPlan’s post says the implied removal rate per tonne of basalt in Lithos’s credits is high relative to what the published ERW literature would predict, and that key modeling inputs sit behind closed doors at the registry. A parallel peer-reviewed letter in ACS ES&T Letters argues that the most intuitive alternative method, measuring soil CO2 flux directly, does not work for ERW because biological fluxes swamp the weathering signal. So the measurement method in use is hard to audit, and the obvious backup method is confounded. That is a pathway-wide MRV (measurement, reporting, verification) gap, not a Lithos-specific one. See our earlier ERW primer and our prior note on the Climeworks/Lithos credits partnership for why this quantification question is load-bearing for the entire cost argument.
The market angle
ERW credits have been clearing in the $200 to $350 per tonne range, well below DAC but at a premium to nature-based credits, specifically because buyers trust the durability story. The premium only holds if the gross-removal number holds. If a reviewer can shave 30 to 50 percent off an issuance because the model inputs cannot be defended in public, the economics for developers compress fast and the comparison to DAC on a verified-tonne basis narrows. Buyers paying the durability premium are effectively paying for the MRV stack. If the stack is not audit-ready, they are overpaying.
This is not academic. Mombak just issued its first Isometric-verified ERW credits in Brazil, and Terradot, Eion, and others have issuances queued for the next two quarters. Every one of them will now be read against the Lithos questions.
Policy and registry context
Durable CDR registries, Isometric, Puro, Frontier-adjacent protocols, have positioned themselves as the adults in the room after the VCM’s forest-offset failures. The pitch has been: rigorous protocols, third-party verification, durable storage only. CarbonPlan’s critique is structural. It argues the registries still let too much of the critical quantification math happen behind proprietary walls. There is no U.S. federal MRV standard for ERW. The EU’s CRCF (Carbon Removal Certification Framework) is still drafting ERW methodology. So registries are the de facto regulator, and their transparency rules are the ceiling.
What the analysts are saying
“Lithos’s first enhanced weathering credits claim a surprisingly high amount of gross carbon removal that we think deserves scrutiny. We tried to figure out what could make these high numbers possible, but we were limited by the available data and left with more questions than answers.” - CarbonPlan (@carbonplan.org)
“If this carbon removal outcome is real, we need public evidence to understand why. If it’s not, then this is a case of over-issuance that should be corrected. Either way, there’s a lesson here that’s impossible to learn with the relevant data behind closed doors.” - CarbonPlan (@carbonplan.org)
That is unusually direct from a group that normally hedges. And it rhymes with a broader point Mark Gongloff made this week about BECCS (bioenergy with carbon capture and storage) that applies just as well here: “Technology that definitely generates carbon plus technology not yet proven to remove carbon at scale does not equal negative emissions” - Mark Gongloff (@markgongloff.bsky.social). Mining, crushing, and spreading basalt has a real carbon cost. Gross removal has to be both real and large enough to clear that cost with margin.
The counter-argument
Lithos and its supporters will say, reasonably, that ERW science is moving fast, that peer-reviewed models support the approach, and that commercial confidentiality is normal in any industrial process. They will argue that CarbonPlan is holding a first-of-kind issuance to a standard no early pathway can meet, and that the right answer is iterative disclosure, not retroactive revision. That is a fair defense. The problem is it was also the defense the forest-offset developers used in 2021.
Verdict
This is the week ERW stopped being the easy answer. The pathway is not discredited. The commercial thesis, cheap, durable, gigaton-capable, is still the best one in CDR if the measurement holds. But the burden of proof has shifted. Buyers should ask for primary field data, not registry summaries. Developers should treat opacity as a commercial liability, not a moat. Registries should publish the inputs, or explain in writing why they will not. The next Mombak, Terradot, and Eion issuances will be read with regression code open. That is healthy. It is also the exact moment the VCM failed to rise to a decade ago. Durable CDR gets one chance to do it differently.