German agritech company Computomics has raised €6.3 million in Series B funding to expand an AI-assisted crop-breeding platform designed to predict how plant varieties will perform under different environmental conditions before breeders spend years testing them in the field.

The round was led by Convent Capital’s Agri Food Fund, which invested €5 million, with existing investors High-Tech Gründerfonds, MBG Baden-Württemberg, and Amathaon Capital also participating. Computomics says the capital will be used to scale commercial delivery of its climate-smart breeding technology across more breeding programmes.

The company’s ×SeedScore platform combines genomic data with environmental information including temperature, rainfall, soil conditions, and field measurements. Its machine-learning models are intended to help breeders identify which crop candidates are most likely to perform consistently across different locations and under hotter or drier conditions.

That approach is being pitched into a sector where breeding cycles are poorly matched to the speed of climate change. New commercial crop varieties can take years to develop, while heat, drought, and shifting growing conditions are already altering the performance assumptions on which breeders and growers rely.

Breeding cycles collide with faster climate change

Traditional plant breeding depends heavily on repeated field trials because performance varies with location, weather, soil, management, disease pressure, and interactions between those factors. The process is scientifically robust but slow, and each additional growing season adds cost before a breeder knows whether a candidate deserves further investment.

Computomics is trying to move more of that selection work earlier by predicting how genotypes are likely to behave in environments they may not yet have encountered in trials. The company does not remove field testing from the process; instead, its models are intended to narrow the number of candidates that need the most expensive stages of physical evaluation.

The practical value of that approach depends on whether predictions remain reliable outside the datasets used to train them. Agricultural data can be uneven across regions and seasons, while extreme weather can create conditions that are poorly represented in historical records. A model that performs well across familiar environments may still struggle when temperature, rainfall, or soil moisture moves beyond previous ranges.

Breeders also have to make decisions across several traits at once. Yield, disease resistance, quality, maturity, and resilience can trade off against each other, which means a useful platform has to support selection across commercial priorities rather than simply identify the plant that performs best on one metric.

AI has to earn trust in the field

Computomics’ commercial challenge is therefore less about convincing breeders that data matters than about proving that its recommendations improve outcomes often enough to justify changing established workflows. Crop breeding already uses statistics, genomics, and computational tools extensively, so AI has to add measurable predictive value rather than merely repackage analysis under a new label.

The company’s customer base spans field crops, forages, vegetables, and specialty crops, giving it exposure to breeding programmes with very different economics and data volumes. Larger seed companies may have extensive historical datasets, while smaller breeders can have less information and fewer technical staff, making implementation requirements vary sharply between customers.

Climate pressure gives the technology a stronger commercial backdrop. European growers have faced repeated heat and drought episodes, and yield forecasts for several spring and summer crops have come under pressure, increasing the cost of varieties that perform well only under historical conditions.

Funding will allow Computomics to push further into commercial breeding programmes, but adoption will ultimately depend on whether the platform can reduce wasted seasons, improve selection rates, or shorten time to market without adding another layer of software that breeders have to maintain. Those outcomes will be visible only over multiple growing cycles.

AI can compress parts of the decision process, although biology still has the final say. The strongest evidence for Computomics will come when varieties selected with its predictions continue to perform across the uneven weather, soil, and management conditions that models are supposed to anticipate.