SoilFlux
Vision

Enabling regenerative agriculture at scale,
without compromising food security.

Crops are fed by a mix of chemistry and biology, and the right mix differs from field to field. Finding it requires knowing how each input performs in a given soil and climate. We are building the intelligence layer that provides that knowledge, starting with the trial programmes that bring biologicals to market.

The right balance for every field.

Chemistry and biology both feed the world's crops. The practical question is which input to use, at what rate, and on which ground. A model of soil and climate that covers any location and improves with every trial turns that question into a decision.

How input decisions change

TodayWith SoilFlux

One trial per question, per site, per season

One model that carries every result into the next decision

Trial sites chosen by availability and experience

Sites chosen for the conditions the claim needs

Replication padded against failure

Replication sized to the ground

A failed site is a lost season

Each site's result is explained and used to plan the next programme

Evidence that stays in the report

Evidence that shows where the product works

Why the ground is the deciding factor

Biologicals and biostimulants act in living soil, with native microbes, changing moisture and soil chemistry that can vary within a single field. The same product can respond strongly at one site and weakly at the next. We read that context before the season and learn from every season that follows.

Four questions, one model

Every input decision depends on the same four questions. Answering them in one model means each answer improves the others.

  1. 01State

    What is this ground?

    Chemistry, physical structure, biology and climate at any location on land.

  2. 02Dynamics

    How will it respond?

    How water, carbon and nutrients change over the season and across years, under different management and climate.

  3. 03Decision

    What should go on it?

    Which input to use, where, and in what balance, from the site plan of a trial programme to the input mix of a single field.

  4. 04Measurement

    What should we measure?

    Where, when and what to test, so each measurement either supports a claim or improves the model.

The engine is the soil. The first product is the trial programme.

Before a biological reaches farms, it has to pass registration and claim trials across crops, soils and regions. This is where knowing the ground pays off first, through sites that deliver and seasons that are not lost. It is where we start.

  1. 01Plan

    Where do we trial?

    Sites that match the conditions the claim needs, with replication set for each site.

  2. 02Explain

    What drove the result?

    Each site's result read against its soil and climate, so site effects are separated from product effects.

  3. 03Apply

    Where does it hold?

    Which soils and regions the evidence covers, as the basis for the claim.

  4. 04Learn

    What comes next?

    Each season of trial data improves the next programme and the model behind later decisions.

Built on evidence

Each recommendation comes with its evidence: the soil and climate at the site, the comparable trials, and the measurements that would confirm it. That is what registrars, trial partners and management need before acting on it.

Each programme returns several times its fee.

Trial programmes are one of the largest costs of bringing a biological to market, and part of every programme is spent on sites that cannot deliver. SoilFlux is priced per programme and measured against the trial cost it saves.

2–6×

return on the fee per trial programme, from avoided trial costs alone.*

Fewer failed sites

Sites that cannot provide the claim's conditions are removed before the protocol is written, so fewer sites are needed to reach the required number of valid trials.

Replication sized to the ground

Each site gets the number of replicates its variability requires, without extra padding.

Seasons protected

A claim that falls short of valid trials loses a full season. Programmes built on reliable sites keep the claim on schedule, which adds value beyond the trial savings.

Mid-size makers run four to five trial programmes a year, and each new product or claim adds another, so revenue grows with the customer's pipeline.

* Estimate: our per-programme fee compared with the trial cost saved by avoiding failed sites, based on trial costs reported to us by industry practitioners. Not drawn from a published dataset.

Same vision. Two lenses.

If you're bringing a biological to market

Trial programmes built on sites that deliver, with evidence of where your product works.

Trial sites are usually chosen based on experience, availability and a consultant's judgement, and a failed site costs a season. A model of the ground that covers any location turns that choice into a plan: the conditions your claim needs, the sites that meet them, and the replication each site requires.

Chemistry and biology both feed the world's crops, and choosing the right balance requires knowing how each performs in real soil. We start where that knowledge pays off first: the trial programmes that take a biological to market.

If you back intelligence layers in deep verticals

The intelligence layer for the inputs that feed the world's crops.

Input decisions in agriculture still rely on trial and error, one site and one season at a time. SoilFlux builds one model of the ground covering soil, water, carbon and biology, constrained by physics, that improves with every trial and measurement. The data grows with each customer programme, and so does our advantage.

Our first customers are mid-size makers of biologicals and biostimulants running four to five registration and claim trial programmes a year, each a significant cost and each exposed to failed sites. SoilFlux plans those programmes with fewer failed sites and is priced per programme, so each programme returns 2–6×* its fee in avoided trial costs, before counting the season a failed claim would lose. Regulation drives demand: EU FPR claims require valid trials across crop groups and soil classes, and India's FCO requires efficacy data from three agro-ecological zones. We start as a per-programme service and move to self-service software as customers grow. The same model then extends from trial design to the balance of chemistry and biology across the input market.

Request investor materials

In closing

“One model of the ground beneath agriculture that improves with every trial and every season. It shows which input to use, where, and in what balance.”

The SoilFlux team · 2026

Work with usMeet the team