Eleven of the scientists who built nickel agromining — van der Ent, Aarts, Morel, Chaney, Baker, Simonnot, Laubie, Pollard, Tang, Qiu and Echevarria — have published a state-of-the-art critical review in Environmental Science & Technology (60(9):6852–6861, published online 2026-02-24) that is less a victory lap than a referee’s report on their own field. Its most important sentence is not about plants at all. Working from the DTPA-extractable nickel pool of tropical Cambisols (<400 mg kg⁻¹), they budget the topsoil itself: 3,200 tonnes of soil in the 0–20 cm layer hold about 1,280 kg of Ni-DTPA per hectare, of which roughly 40% is estimated to be ultimately accessible to a yearly harvested crop. That arithmetic lands on ~500 kg Ni ha⁻¹ yr⁻¹ — and leads directly to their warning: any yield claim substantially above about 500 kg per hectare per year “should be regarded as very doubtful”, because it exceeds what any known metal crop, on any known natural ultramafic soil, can physically deliver.
What the field record actually shows
The best field demonstration to date sits just under that ceiling: an improved cultivar of Odontarrhena chalcidica grown in Oregon, USA, reached 2.0 wt% Ni in whole harvestable biomass at 20 t biomass ha⁻¹ — 400 kg Ni ha⁻¹. Demonstrated European yields are far lower: 105 kg Ni ha⁻¹ and 106.3 kg Ni ha⁻¹. The commercial threshold the authors state for a crop to be attractive at all is >1 wt% Ni in total harvestable dry biomass and
10 t dry biomass ha⁻¹ yr⁻¹, i.e. ~100 kg Ni ha⁻¹ yr⁻¹. Keep the distinction in mind, because it is where agromining hype usually inflates: the spectacular leaf maxima — 22,000 µg g⁻¹ in O. chalcidica, 11,600 in Berkheya coddii, 25,060 in Phyllanthus rufuschaneyi on this site’s own records — are foliar extremes from wild plants. Whole-harvestable biomass, diluted by stems and wood, is necessarily lower, and the review explicitly calls it “particularly poor practice” to multiply a highest recorded foliar concentration by an imagined biomass yield. Anyone selling you a Ni agromining yield derived from a leaf record is doing exactly that.
Three companies, no numbers
The review counts three companies currently implementing Ni agromining: MetalPlant in Albania (crop: O. chalcidica), Botanickel in Greece and Malaysia (O. chalcidica and P. rufuschaneyi), and Genomines in South Africa (B. coddii) — matching one crop to each climate zone where those species perform. The Sabah operation is the most instructive case: the Malaysian state holds over 3,500 km² of ultramafic soils, a 2010–2022 string of PhD theses and openly published ecology underpinned commercialization, which began in 2023 with the creation of Botanickel Sabah and a joint laboratory (“TANi Lab”) with Universiti Teknologi MARA, and a Nagoya-Protocol commercial access license with benefit sharing granted through the Sabah Biodiversity Council in 2025. Yet all three companies, the authors note, publish neither cultivated area nor production figures. “Commercial scale” today is a licensing and planting achievement, not an audited tonnage — and the authors estimate, “in our estimation”, that getting an agromining venture established takes 8–10 years and well over US$10M.
The downstream mechanism: bio-ore is the product, not biomass
A second under-appreciated point: the metal-enriched biomass itself is nearly worthless. Combustion concentrates it into a “bio-ore” ash that typically carries >20 wt% Ni, and only electric-arc smelters can accept such material — stainless-steel feed is emerging as the main off-take route, with hydrometallurgy able to refine battery-grade nickel sulfate (99.98% purity) as the higher-value alternative, at the cost of strong-acid chemistry whose footprint can partly be credited back by recovering K, Mg and P as fertilizer. Combustion is not trivial either: herbaceous leaf-rich biomass sinters in boilers designed for wood, and high chlorine content drives acid-fume, dioxin-formation and corrosion risks that demand flue-gas treatment. The review’s blunt economic point: a startup that plans only to grow hyperaccumulators has built half a supply chain. The conversion end — smelter agreements or hydrometallurgical capacity — is the part that fails.
The genuine risks the review flags to its own field
The paper spends its third section auditing agromining’s own claims. Genetic modification of non-accumulators for Ni agromining it calls a “complete fallacy” — Ni hyperaccumulation is a multigenic trait of hyperaccumulator lineages whose molecular basis is still unresolved even in the flagship crops, and engineering it into food crops would risk introgressing the trait into cultivated varieties. Invasiveness is documented history, not hypothetic risk: O. chalcidica and O. corsica introduced to Oregon in the 1990s escaped and caused local invasions — reason enough that the US DOE’s ARPA-E program now tasks its agromining grantees with developing “non-spreading” variants. And the conservation irony is sharpened here: of the more than 500 known Ni hyperaccumulators, fewer than 20 have the traits a metal crop needs, handheld XRF screening of herbaria has doubled the known inventory in five years, the authors estimate thousands remain undiscovered — and the renewable-energy mining boom is destroying the metallophyte habitats those candidates live in faster than they can be found. Sabah Parks’ ex situ “Hyperaccumulator Garden” is cited as the kind of germplasm insurance the whole field depends on.
A reviewer’s caveat, disclosed in the open
This is a Critical Review, not a primary experiment: every number above rests on the references it cites, and its own abstract concedes no independent economic modeling exists “for commercially sensitive and proprietary reasons”. Its authors’ conflicts are printed in the paper itself — eight of the eleven are scientific advisors to Botanickel or employed by Botanickel/Econick, the very companies whose “finally becoming a reality” this review heralds. That does not make its skepticism worthless: the ~500 kg ceiling, the 105–106 kg European record, the invasiveness history and the anti-hype argument are the field’s own leaders policing it, and the hedges (“estimated”, “in our estimation”, “should be regarded as very doubtful”) are doing honest work. But treat “commercial scale” as a direction, not an accomplishment — the next few years of published, field-scale, whole-chain tonnage are what will prove or break the claim.
Source: van der Ent, Aarts, Morel, Chaney, Baker, Simonnot, Laubie, Pollard, Tang, Qiu & Echevarria (2026), Environmental Science & Technology 60(9):6852–6861, DOI 10.1021/acs.est.5c08259 (published online 2026-02-24; received 2025-06-21, accepted 2026-02-06). Full text read from the legal open-access copy in the Wageningen University & Research repository (Article 25fa). Every load-bearing number — the 1,280 kg Ni-DTPA ha⁻¹ soil budget and 40% accessibility estimate, the ~500 kg ha⁻¹ yr⁻¹ ceiling, the 400 kg Oregon and 105/106.3 kg European demonstrations, the >1 wt% / >10 t thresholds, the >20 wt% bio-ore and 99.98% nickel sulfate figures, and the 3500 km², 2023 and 2025 Sabah milestones — was verified verbatim by three independent reviewers on three different model vendors, each retrieving the repository PDF and the Crossref/OpenAlex metadata independently; no numeric discrepancies were found, and the reviewers’ wording flags (leaf-maximum versus whole-biomass concentrations, the authors’ Botanickel/Econick conflict of interest, and the estimated nature of the 40% and 8–10-year figures) are reflected in the text above. See /methodology/ for how we source and check analysis pieces.