Biochar is the dominant durable carbon removal pathway. It is also riddled with technical problems the industry tends to downplay. This piece names them. Not to kill biochar, but to make the case for it stronger.
Biochar is, at the time of writing, the dominant durable carbon dioxide removal pathway by delivered tonnes. According to CDR.fyi's April 2026 dashboard, biochar accounts for roughly 92 percent of all delivered durable CDR credits to date. That dominance is real, and we have written about why in Biochar Carbon Credits 2026 and Terra Preta and the Permanence Question.
So why publish a skeptic's case for the technology we cover most?
Because the Genesis principle of this publication is "no claim without source, no data without context." A media brand that only publishes the bull case for its preferred technology is a marketing arm, not a publication. That principle was the entire argument of Truth Needs No Embellishment, and it is the same logic that drove the failure analysis in When Green Projects Fail.
Biochar has real problems. Yield variance is enormous. Methane leakage is a documented production hazard. Feedstock sourcing has displacement risks that mirror the early failures of first-generation biofuels. The half-life math has wider error bars than the marketing copy admits. The cost curve is not behaving like solar's. Application rate math does not yet support agronomic deployment at scale. Verification costs eat smallholder margins.
None of these problems kill biochar. They define where it works, where it does not, and what has to improve. The honest version of biochar is more useful to a buyer, a policymaker, or a producer than the brochure version. Below, the seven problems, in order of how often they get glossed.
A pathway worth defending is a pathway worth interrogating. The credibility of every biochar dollar deployed is backed by the willingness to name where the math breaks.
The most quoted biochar statistic is "biochar boosts crop yields by 14 percent." It is technically true. It is also a meta-analysis average across 254 field studies, and it conceals a distribution wide enough to drive a tractor through.
Across the same 254 studies, individual yield responses ranged from negative 28 percent to positive 63 percent. Roughly 30 percent of trials showed no benefit at all. A non-trivial fraction showed a yield penalty. The "+14%" headline is a real number. It is just not the number any individual farmer is going to experience.
What drives the variance? Six factors, all of them well documented in the agronomy literature:
Soil pH. Biochar is alkaline. On acidic soils (the majority of tropical and many temperate cropland soils), it acts as a soil conditioner and the yield response is consistently positive. On already-alkaline soils, adding more alkalinity can push pH out of the optimal range for most crops. The same input that adds 30 percent in Indonesia can reduce yields in parts of Spain.
Soil texture. Biochar improves water and nutrient retention in clay-heavy and degraded soils. On well-structured sandy loams or organic-rich soils, the marginal benefit is small to zero. The soils that benefit most are the ones that need help most.
Application rate. Field trials at 5 tonnes per hectare often show no measurable effect. Trials at 40 tonnes per hectare can show large positive effects, or, with the wrong feedstock, can suppress germination through salt accumulation or nutrient lock-up.
Feedstock quality. High-ash biochar from manure or rice husks can effectively salt the soil. Wood-derived biochar is generally cleaner. The category "biochar" hides enormous variation in input material.
Pyrolysis temperature. High-temperature biochar (above 550 degrees Celsius) is more stable in soil and earns higher carbon credit value. It is also less bioactive and provides less immediate nutrient availability. The carbon-optimal biochar and the agronomy-optimal biochar are not the same product.
Crop type. Legumes, which fix their own nitrogen, respond differently than nitrogen-hungry cereals. Vegetables respond differently than tree crops. There is no universal dose-response curve.
The honest framing is that biochar is conditional, not universal. The "+14% average" headline is true the same way "the average human has one ovary" is true.
Pyrolysis is supposed to happen in an oxygen-limited environment. The whole point is to convert biomass to stable carbon without combustion. In practice, real-world pyrolysis systems leak.
Traditional biochar kilns, the kind still used by smallholder producers and many community-scale operators, release between 1 and 7 percent of feedstock carbon as methane. That sounds small. It is not. Methane has 28 to 86 times the warming potential of carbon dioxide over a 100-year horizon, depending on which IPCC accounting framework you use. A biochar kiln that leaks 5 percent of its carbon as methane can be net-positive on greenhouse gas emissions despite producing stable carbon for the soil.
This is the dirty secret of legacy biochar production. A facility can be honestly making "carbon-negative" biochar while operating as a net climate liability.
Modern industrial pyrolysis systems address this. Operators like Carbofex (Finland), Mash Makes (India and Denmark), and Pyreg (Germany) run sealed reactors with secondary combustion of off-gases. The methane and other volatile organic compounds are flared or used for process heat, which converts them to carbon dioxide before they escape. The leakage rate drops below 0.5 percent, often well below.
The problem is that certification standards do not always require flue gas measurement. The Puro.earth methodology v3.0 (the most rigorous of the major registries) requires lifecycle accounting and applies discount factors for leakage, but other methodologies are looser. Some allow self-reported emissions data without independent verification. A 2023 review by Whitman et al. estimated that between 5 and 15 percent of biochar projects globally may be net-positive on emissions due to undetected leakage.
If you are buying biochar credits, the question to ask is not "how stable is the carbon" but "what is the leakage rate during production, and who verified it?" If the producer cannot answer that with measured flue gas data, the credits are not what they look like.
Biochar requires biomass. The carbon math only works if that biomass would otherwise have decomposed and released its carbon back to the atmosphere. The moment you start sourcing biomass that had a different fate, the math gets harder.
Sustainable feedstock sources are real. Sawmill waste, rice husks, coffee chaff, urban tree pruning, manure from concentrated animal operations, and crop residues from regenerative systems with surplus carbon flow are all legitimate inputs. A biochar facility built next to a sawmill that was previously paying to dispose of bark and offcuts is unambiguously additional. The carbon was going to be released anyway. Now it is not.
Unsustainable feedstock sources are also real, and they are growing. Three categories deserve scrutiny:
Forest harvesting. Some biochar operators have started contracting for "energy wood," which is a polite term for trees harvested specifically to produce biochar feedstock. If the forest was going to be harvested anyway, the carbon math is neutral on the harvest decision. If the forest was harvested specifically because biochar created demand, the project is replacing live carbon storage with stable but lower-density carbon. The accounting depends on baseline counterfactuals that are hard to verify.
Crop residue removal. Most cereal crops leave residues (straw, stalks, husks) that, in healthy farming systems, return to the soil as organic matter. That residue is the carbon and nutrient flow that feeds the next crop. Removing it for biochar production sounds carbon-positive (you are stabilizing carbon that would have decomposed), but over time it can degrade soil organic matter and reduce yields. The IPCC's 2019 special report on land use flagged this as a structural risk for residue-based biochar at scale.
Tropical biomass exports. The "biomass colonialism" critique, raised by groups like the Global Forest Coalition and Biofuelwatch, points out that wealthy buyers in Europe and North America are increasingly sourcing biochar feedstock from tropical countries. The carbon credit value flows to the buyer. The land use impact stays local. This is the same pattern that derailed first-generation palm oil biofuels, and it is starting to repeat.
The certification gap is the core problem. Most biochar standards require carbon stability measurement, but very few require full lifecycle analysis of feedstock sourcing, including baseline counterfactuals and land use change. Puro.earth's methodology requires "sustainably sourced biomass" but the verification burden is light. A producer who claims they are using sawmill waste is generally taken at their word.
The honest framing: biochar is only as carbon-negative as its feedstock chain. A great pyrolysis system fed by ecologically destructive biomass is not a climate solution.
"Five hundred year permanence" is the marketing claim. It is also the figure most major registries use as a baseline assumption for high-temperature biochar. The actual permanence is more variable than that single number suggests.
Biochar permanence is governed by three variables:
Pyrolysis temperature. Biochar produced above 550 degrees Celsius has a high proportion of aromatic carbon, which is biologically resistant and slow to decompose. Biochar produced at lower temperatures has more labile carbon and degrades faster. The 500-year claim only applies to high-temperature product. Producers who run cooler kilns (often the cheaper option) are not making the same molecule.
Soil and climate conditions. Tropical soils with high microbial activity and warm, wet conditions degrade biochar faster than cool, dry temperate soils. The same biochar applied in northern Sweden and southern Brazil will have meaningfully different decay curves. Most permanence models use temperate-zone defaults.
Application method. Biochar incorporated into the soil profile is shielded from UV degradation, oxidation at the surface, and physical erosion. Surface-applied biochar (the cheaper application method) decays faster. Some producers apply biochar to fields and call the job done. The numbers assume incorporation.
The empirical record is split. Lehmann et al.'s 2021 review in Nature Geoscience found field studies showing 80 percent or more of biochar carbon remaining after 100 years, which supports the long-permanence claim. But the same review noted other studies showing 60 percent remaining after just 30 years, which implies a half-life closer to 50 years than 500.
The variance is large enough that the major registries apply discount factors. Puro.earth, for example, uses a permanence factor that already discounts the nominal carbon content to account for expected losses. The credit you buy is not for the carbon physically present in the biochar. It is for the carbon expected to remain after a defined accounting period, with the variance baked into the discount.
This is fine, as long as buyers understand it. The problem is that producers and brokers tend to quote the gross carbon content, not the net credit, and the "500 year permanence" claim gets repeated without the conditions attached. The honest version is: biochar permanence is high relative to other CDR pathways, but the science still has wide error bars, and the conditions matter.
Solar photovoltaics cost 99.6 percent less per watt today than they did in 1976. That decline followed a steady learning curve, often described in terms of Wright's Law: every doubling of cumulative production reduced costs by a fixed percentage. The same pattern, with different slopes, has played out for batteries, LEDs, and wind turbines. Manufactured technologies that benefit from semiconductor-like learning effects ride a steep cost decline.
Biochar is not on that curve. It is on a much shallower one.
The reason for the difference is structural. Solar PV is a manufactured semiconductor product. Each doubling of cumulative production yielded measurable improvements in cell efficiency, manufacturing yield, and supply chain integration. Biochar is a thermal-mechanical product. The reactor still has to heat biomass to several hundred degrees Celsius. The energy input per tonne of biochar is bounded by the heat capacity of biomass and the enthalpy of pyrolysis. There is no equivalent of Moore's Law for boiling sticks.
Industry estimates put biochar's cost decline since 2015 at roughly 30 percent. That is real, and it is meaningful. It is also an order of magnitude shallower than what solar achieved over a comparable early period. Direct air capture, which is also a thermal-mechanical process but benefits from cross-industry learning in chemical engineering, is currently following a steeper learning curve than biochar despite starting from a higher cost base.
The implication is that biochar may plateau. Most industry forecasts assume biochar carbon credit prices will stay in the 100 to 150 dollar per tonne range through 2030 and decline only modestly thereafter. If that holds, biochar will remain a credible CDR pathway, but it will not become cheap the way solar became cheap. Buyers who are betting on a solar-style cost collapse for biochar are betting against the physics of pyrolysis.
Two problems that get less attention than they deserve, both about scale. The first is about whether biochar can reach the cropland it is supposed to improve. The second is about who can afford to participate in the credit market.
Problem 6: The application rate math does not pencil at scale. Field studies converge on an optimal biochar application rate of 10 to 40 tonnes per hectare for soil improvement. A typical Iowa corn farm is around 200 hectares. Optimal application for a single farm at the high end of that range is 8,000 tonnes of biochar.
Total US biochar production capacity in 2025 was approximately 250,000 tonnes per year, according to the US Biochar Initiative. That is enough to fully treat about 30 farms at high-end application rates. The United States has roughly 2 million farms. The math is off by four orders of magnitude.
This is the single biggest reason the biochar industry is currently driven by carbon credit revenue rather than agronomic demand. Selling biochar at 130 to 160 dollars per tonne CO2-equivalent to corporate buyers is more economically viable than selling it at 200 to 400 dollars per tonne to farmers as a soil amendment. The agronomy story is real, but at current production scale it is a niche application, not a global soil management strategy. The carbon credit pathway is keeping the industry alive while production capacity scales up.
Problem 7: Verification costs are a smallholder filter. Lab certification of biochar carbon stability, third-party measurement of production volumes and lifecycle emissions, and registry listing fees together cost roughly 20 to 40 dollars per tonne CO2-equivalent. For producers selling credits at 130 to 160 dollars per tonne, that is 15 to 30 percent of gross revenue going to verification overhead.
Industrial-scale producers can absorb that. They have the volume to amortize fixed certification costs across thousands of tonnes per year. Smallholder and community-scale producers cannot. The verification overhead is roughly the same in absolute terms whether you produce 10,000 tonnes per year or 100, and at small volumes it eats the entire margin.
The result is a quality tier where credible biochar credits flow primarily through large industrial producers in wealthy countries (Carbofex, Pyreg, Mash Makes industrial sites). The community-scale and smallholder biochar projects in Kenya, India, and Indonesia, which are often presented as the development-friendly version of the technology, are largely shut out of the credit market by verification costs they cannot pay. The smallholder narrative is more aspirational than operational at present prices.
This is fixable. Pooled verification, simplified methodologies for smallholder pools, and direct buyer contracts with built-in verification subsidies are all being piloted. But the current state of play is that biochar's credit market is structurally biased toward industrial producers, and the equity story has not caught up with the business model yet.
None of the seven problems above kill biochar. They define where it works.
Biochar is conditional. It works on acidic, degraded, or clay-heavy soils. It does not always work on alkaline or sandy soils. It works when produced above 550 degrees Celsius in sealed reactors with flue gas combustion. It does not work in leaky kilns. It works when feedstock is genuinely waste biomass with no competing use. It does not work when feedstock displaces forests, food crops, or soil nutrient cycles. It works at the carbon-credit price tier with industrial-scale verification. It works less well as a smallholder product at current verification costs.
That is the honest version. It is also the version that holds up to scrutiny. The bull case for biochar that ignores yield variance, methane leakage, feedstock displacement, half-life uncertainty, cost curve realism, application rate math, and verification overhead is not actually a bull case. It is a marketing case. And marketing cases get punctured the moment a critic does the source-checking.
The Gr0ve covers biochar because biochar is the best durable carbon dioxide removal pathway available right now, conditional on production quality, feedstock sourcing, and deployment context. Direct air capture may eventually be cheaper, but it is not there yet. Enhanced rock weathering is promising but earlier-stage. Ocean alkalinity is largely speculative at the scale needed. Bioenergy with carbon capture and storage has the same biomass sourcing problems as biochar, plus the storage problem on top. Biochar wins the comparison on delivered tonnes, on price, on permanence per dollar, and on operational maturity.
It also has the seven problems above. Both things are true at the same time. That is what conditional means.
The Genesis principle that governs this publication is "no claim without source, no data without context." Applied to biochar, that means we cover the meta-analysis average and the variance behind it. We cover the carbon-negative claim and the methane leakage that can invert it. We cover the permanence claim and the soil and temperature conditions it depends on. We cover the cost decline and the pyrolysis physics that bound it.
Pretending biochar has no problems is bad for biochar. It hands ammunition to anyone willing to do the source-checking, and the source-checking is not hard. The honest version, the conditional version, the version with the failure modes named, is the version that survives the next decade of scrutiny. That is the version we are going to keep writing.
The companion pieces in this cluster lay out the upside case in equal detail. Terra Preta and the Permanence Question covers why some biochar carbon has remained stable in soil for over 2,500 years. Biochar Carbon Credits 2026 covers the buyer-side market dynamics. Biochar in the Sea covers an emerging coastal application that may sidestep some of the agronomic problems. Read all four together. Then make your own call.
For background reading on the technology itself, start with What Is Biochar? and the comparative explainers Biochar vs BECCS and Biochar vs Compost. For the broader editorial logic that produced this piece, see When Green Projects Fail, Truth Needs No Embellishment, and What Precision Fermentation's Failures Teach Us.
The honest version of biochar is the version that survives the next decade of scrutiny. The marketing version is the version that gets a journalist's investigation in 2028. We would rather publish the first one.
A weekly data digest covering which green technologies are winning economically, and which ones are not. One email. No spam. Just the data that matters.
Free via Substack. Unsubscribe anytime.
On average, yes. A 2020 meta-analysis published in GCB Bioenergy covering 254 field studies found a mean yield response of approximately +14 percent. But the average hides enormous variance. Across the same studies, individual results ranged from -28 percent to +63 percent, and roughly 30 percent of trials showed no benefit or a negative response. Yield response depends on soil pH (biochar works best on acidic soils, often fails on alkaline), soil texture (helps clay-heavy soils, neutral on sandy), application rate, feedstock quality, pyrolysis temperature, and crop type. The honest framing is that biochar is a conditional input, not a universal one.
Source: Ye et al., GCB Bioenergy, 2020.
Yes, it can. Pyrolysis is supposed to be oxygen-limited, but real-world systems leak. Traditional kilns release 1 to 7 percent of feedstock carbon as methane, which has 28 to 86 times the warming potential of carbon dioxide over 100 years. A leaky biochar plant can be net-positive on emissions despite producing stable carbon. Modern industrial pyrolysis systems from operators like Carbofex and Mash Makes capture flue gases and dramatically reduce this loss, but certification standards do not always require flue gas measurement. Lifecycle estimates suggest 5 to 15 percent of biochar projects globally may be net-positive on greenhouse gas emissions due to leakage.
Sources: Whitman et al., Carbon Management, 2023; Puro.earth methodology v3.0.
It is a defensible upper bound under specific conditions, not a guaranteed outcome. Biochar permanence depends on pyrolysis temperature (above 550 degrees Celsius is generally required for high stability), soil conditions (tropical soils degrade biochar faster than temperate), and application method (incorporated biochar persists longer than surface-applied). Some field studies show 80 percent of biochar carbon remaining after 100 years, which supports the long-permanence claim. Other studies show 60 percent remaining after just 30 years, which does not. Major registries like Puro.earth and Verra apply discount factors to account for this variance, but the underlying science still has wide error bars.
Sources: Lehmann et al., Nature Geoscience, 2021; IPCC AR6 WG3, Chapter 7.