Two markets wearing one name
The phrase carbon credit hides two fundamentally different products. An avoidance credit claims that an emission did not happen - a forest was not cut, a plant was not built - relative to a counterfactual that no one can observe. A removal credit claims that a tonne of carbon already in the atmosphere was taken out and stored. These are not two grades of the same thing; they are different assets with different physics and different failure modes. Treating them as fungible is the mistake that has distorted the market for a decade.
The problem with avoidance is baseline integrity. The credit is only as good as the counterfactual it is measured against, and counterfactuals are guesses that project developers have every incentive to inflate. Repeated scrutiny of avoidance portfolios has found that a large share represent far less real abatement than claimed, and in some cases close to none. Once buyers understood that, confidence eroded, and price followed confidence down. That is why we do not buy avoidance - not on principle alone, but because the asset does not reliably deliver the thing it is sold as.
What durable removal actually requires
Removal has to clear a higher bar, and the good projects do. Two properties matter. Additionality is easier to establish, because the carbon is physically captured and can be measured rather than argued from a hypothetical. Durability - how long the carbon stays out of the atmosphere - is the harder and more important axis. A tree can be logged or burned in a decade; the question for any removal is whether storage is measured in years, decades, or centuries.
This is why verified agricultural sequestration and biochar are the categories we focus on. Biochar in particular converts biomass into a stable, carbon-rich solid that resists decomposition and can lock carbon away on a horizon of centuries when properly produced and applied, with a measurable mass balance rather than an inferred one. Verified soil and agricultural sequestration adds co-benefits on land productivity that give the underlying activity a reason to exist beyond the credit. Durability plus measurability plus a real underlying use - that is the profile we are willing to underwrite.
The premium, and why it is rational
Quality costs money, and it should. Where low-integrity avoidance credits have traded down to a few dollars a tonne and struggle to find serious buyers, durable removal from verified agricultural and biochar sources clears at a substantial premium - illustratively on the order of forty to one hundred and twenty dollars per tonne, with the range driven by durability, verification standard, co-benefits, and delivery risk. These figures are indicative ranges, not quotes, and they move with the market. The direction of travel is what matters: as buyers and regulators tighten on integrity, demand concentrates into the credits that survive scrutiny, and the premium for genuine removal widens rather than compresses.
For an allocator, the spread between a collapsing avoidance market and a firming removal market is not noise - it is the trade. The thesis is not that carbon prices rise in general; it is that quality separates from quantity, and that owning the supply side of durable, verifiable removal is a position on that separation.
Underwriting the risks honestly
Removal is not risk-free, and pretending otherwise would repeat the avoidance market's error. The live risks are delivery - a project has to actually produce and apply the biochar or achieve the measured sequestration - verification and standard integrity, permanence monitoring over the claimed horizon, and future policy and definitional changes to what qualifies. Each is a diligence item that can be structured around: staged payment on verified delivery, exposure to recognized standards, and conservative durability assumptions.
The allocator's takeaway is that in a market defined by low average quality, the return lives at the top of the quality distribution, and getting there requires paying up and underwriting the specifics rather than buying the index. Removal, structured and verified properly, is a real asset with a physical basis. Avoidance, too often, is a story about something that did not happen.
