Evidence, as we find it.
Method notes, research and field reports from our lab and our scientists.
LOAM, our soil foundation model
Tim Rajakumar, Quentin Ferry · 1 October 2026
Trained on 67.5 billion bases of soil DNA, LOAM-624M rivals the far larger Evo 1.5 model.
Read
PolicyThe EU Soil Monitoring Law: what it requires, and when
Soilytix · 11 November 2026
In force since 16 December 2025, with national laws due by the end of 2028.
Read
ResearchWhat is blocking land restoration: scale and standards
Soilytix · 11 November 2026
A 30 metre satellite pixel serves national monitoring and is close to useless on a two-acre farm.
Read
NewsAndreas Jacobs joins our Board as Chairman
Dr. Bruno Steinkraus · 20 October 2026
He chaired INSEAD from 2015 to 2024 and Barry Callebaut from 2005 to 2016.
NewsNorman Gentsch joins Soilytix as Head of Agronomy
Cleiton Sena · 20 October 2026
After 16 years at Hannover's Institute of Soil Science, he turns lab measurements into advice.
Field reportFour rooms, four different questions about soil
Soilytix · 8 October 2026
Barcelona, Basel, Valencia and Thessaloniki: four rooms with almost no overlap in who attends.
PolicySoil health finally got its own day at a UN COP
Soilytix · 8 October 2026
Up to 40% of the world's land is degraded, at an estimated cost of USD 878 billion a year.
Field reportThree of five goldmine soil samples had no DNA left to measure
Soilytix · 8 October 2026
Forest soil averaged 28.4 ng/µL of DNA; the three bare mine sites averaged 0.1.
PolicyWe are going to Barcelona to argue that soil is not financeable yet
Soilytix · 8 October 2026
Norrsken backs over 150 start-ups; our case to that room was that soil lacks measurement, not appetite.
Method noteWhy we read soil in four layers
Dr. Matthias Schaks · 8 October 2026
Three layers come from the soil sample; the fourth, landscape, is read from the site.
ResearchLet the ground truth decide where to look
Tim Rajakumar, Quentin Ferry · 1 October 2026
Our largest LOAM model ranked first at sorting genes into 128 enzyme functions, ahead of models 10× larger.
ResearchWhen the same seed gives different answers (G×E)
Julien Winkel · 31 August 2026
Change the field and the ranking of the best variety can flip; soil biology is part of why.
Method noteFrom genus to species: why full ITS beats ITS1
Dr. Matthias Schaks · 20 August 2026
Extended-read sequencing now covers the whole fungal ITS region, about 1,000 base pairs.
Method note"But you're also measuring dead DNA, right?" Yes, and here's why that's not a flaw.
Dr. Matthias Schaks · 13 August 2026
Relic DNA can inflate apparent richness by up to 55%, and it also records the recent past.
Method noteNematode pests: detect them before they cost you
Dr. Matthias Schaks · 16 July 2026
Plant-parasitic nematodes cost an estimated €80–100 billion a year, mostly before symptoms show.
Method noteOomycota are not fungi, and that mistake is costing you data
Dr. Franziska Böwer · 25 June 2026
Clubroot costs oilseed rape up to 60% of its yield, and 16S and ITS sequencing miss it.
Method noteSoil texture is not a background variable. It is a confounder.
Dr. Franziska Böwer · 16 June 2026
Clay and sand give microbes different pores, water and oxygen, so texture skews any comparison.
NewsSoilytix joins the NVIDIA Inception programme
Soilytix · 5 March 2026
Accepted in March 2026 into NVIDIA's programme for start-ups building on AI and GPU computing.
Method noteA triage framework for soil health analysis
Soilytix · 15 October 2025
Test chemistry across 50 fields and it may explain 45; the other five need a deeper, costlier look.
ResearchMicrobes as miners: how MICP is transforming bioremediation
Berenice di Biase · 25 August 2025
Bacteria that grow calcite in soil pores can bind sand and lock lead and arsenic in place.
NewsSoilytix, SLM Partners, Nogam and Migdalo launch a soil microbiome study to enhance nut quality and yield
Soilytix · 4 July 2025
Soil biology across 1,000+ hectares of Portuguese nut orchards, matched to yield and nut quality.