Problem

Living Models is tackling the slow, expensive crop-discovery pipeline. Developing a single crop trait can take more than eight years, while the roughly $60 billion crop-protection market still relies heavily on workflows developed before modern genomics. Climate volatility increases the cost of this delay: seed and breeding companies need to identify useful traits faster, but conventional genomic-selection models are often narrow, crop-specific, and data-hungry.

The principal use case is accelerating crop breeding. BOTANIC is designed to scan a plant genome, rank candidate genes or genomic regions by their predicted phenotypic impact, and direct breeders toward the small number of candidates worth validating physically. In principle, this turns a largely empirical, multi-year search for improved yield stability or other traits into a weeks-long computational prioritization exercise.

Product / Service

Living Models is building foundation models for biology, beginning with BOTANIC, a family of genomic transformer models for plant biology. BOTANIC-0 learns patterns in DNA and can predict gene function, regulatory elements, and trait associations. The company says the model was trained across 43 plant species and more than 60 billion base pairs, with up to one billion parameters, and has matched leading academic genomic models on 22 benchmark tasks while using comparatively modest initial compute.

The commercial model is intended to combine open research with enterprise licensing. Seed companies and research institutions can fine-tune BOTANIC inside secure environments using their own data, while retaining ownership and control of that data; Living Models says it will co-develop integrations that fit existing R&D workflows. The company claims this approach can reduce the crop-development cycle from eight-plus years to roughly two or three by concentrating physical experiments on the most promising candidates.

Market

Living Models competes in biological foundation models, computational genomics, and AI-enabled agricultural research, with an initial focus on plant genomics and crop breeding. Its disclosed technical comparators include leading academic genomic work associated with Cornell and InstaDeep, while the broader category includes companies developing foundation models for DNA, RNA, proteins, and multi-omics data. Living Models is differentiated by specializing its initial model family in plants and by emphasizing deployment against customers’ private breeding data rather than requiring that data to leave their environment.

The company appears to be at an early commercialization stage rather than a mature revenue business. It emerged from stealth in March 2026 with a $7 million seed round, has made BOTANIC open weights and research available, and secured funding for a dedicated 120-GPU NVIDIA B200 cluster. It reports initial commercial traction in North America and Europe and academic partnerships with INRAE, Paris-Saclay Plant Science Institute, CRAG Barcelona, Aarhus University, and the University of Florida, but no revenue or customer-count figures are disclosed in the available research.