Living Models Raises $7M to Build Foundation Models for Biology, Starting with Plants
Climate change is outpacing conventional plant breeding. A Paris–Berkeley AI startup believes the answer is the same technology that transformed natural language processing — applied not to text, but to DNA.
BERKELEY & PARIS — March 11, 2026 — As extreme weather events accelerate faster than the eight-year cycle required to develop a new crop variety, Living Models today announced $7 million in seed funding to bring foundation model AI to plant biology. Its first model family, BOTANIC, is dedicated to plant genomics and already matches leading academic work from Cornell and InstaDeep (backed by BioNTech, $100M+ raised) on key benchmarks — trained on a fraction of the compute those models required.
The Platform
Living Models trains transformer neural networks on biological sequences rather than human text. The core architecture learns representations of genomic data that can be fine-tuned for specific downstream applications — in the same way that a single language model can be adapted to translation, summarisation, or code generation.
Plant biology is the first vertical, but the underlying architecture is designed to generalise across biological domains wherever sequence-driven discovery is bottlenecked by slow, empirical methods — pharma, microbiome research, and synthetic biology among them.
Living Models is building the AI infrastructure layer for the life sciences: the tools that let biologists do in weeks what used to take years, and that demonstrate transformer models can master any information system, not just human language.
“Foundation models proved their value in natural language processing by learning deep patterns from large text corpora. We are applying the same transformer architecture to biological sequences — DNA, RNA, multi-omics measurements. Agriculture is our first vertical because the data is abundant and the commercial need is acute, but the models we are building are designed to transfer across biology.” — Cyril Véran, CEO and co-founder
Why Agriculture First
The global seed industry is dominated by a handful of incumbents — Bayer CropScience, Corteva, Syngenta, BASF, and Limagrain — which collectively spend approximately $8 billion per year on breeding research using methods largely unchanged since the 1960s. Development timelines average eight years from initial crosses to commercial variety release. Pathogen resistance genes are typically overcome within 3–5 years of deployment, and current yield improvement rates of roughly 1% annually are insufficient to meet projected 2050 food demand.
The consequences are already measurable. The 2024 European heatwave alone reduced French wheat yields by an estimated 15%, with agricultural losses exceeding €2 billion. Climate volatility is accelerating faster than conventional breeding can respond.
Plant biology offers three properties that make it an ideal first domain: genomic data is abundant and largely unrestricted, the commercial need is quantifiable, and the feedback loop between computational prediction and real-world validation is well established. BOTANIC, trained across 43 plant species representing over 60 billion base pairs, performs trait predictions computationally — reducing the cycle from initial cross to candidate variety from eight-plus years to two to three, while maintaining field validation for phenotypic confirmation. They don’t replace scientists. They turn researchers into discoverers.
In practice, this means a plant breeder working on drought tolerance no longer needs to grow thousands of field plots to identify which genetic variants confer heat resistance. BOTANIC scans the genome, ranks the candidates by predicted phenotypic impact, and surfaces the top targets for physical validation — compressing what was a multi-year empirical process into weeks of directed
Experimentation.
“Biology is an information problem at every scale, from a single cell to an entire ecosystem,” said Leonard Strouk, CTO and co-founder. “The genomic data exists across many domains — what’s been missing is a model architecture capable of learning from it at scale. We start with plants because the data is rich and the breeding cycle is a clear bottleneck, but the same approach applies wherever sequence data meets slow, empirical discovery.”
“Plants and humans share the same code — DNA, RNA, proteins, phenotype. The difference is we can actually run experiments on plants at scale, fast, without ethical committees or decade-long trials,” said Bertrand Gakière, CSO
Technical Results
BOTANIC (up to 1 billion parameters) achieves competitive performance across 22 standard benchmark tasks, initially trained on just 8 NVIDIA H100 GPUs. The $7M seed round secures a dedicated 120-GPU NVIDIA B200 cluster — likely the world’s largest compute cluster dedicated to plant biology — which the company expects to translate directly into larger models, higher predictive accuracy, and expansion beyond plants.
A research partnership with the University of Florida — one of the world’s leading plant breeding programs — is applying BOTANIC directly to breeding applications, providing the real-world feedback loop between computational prediction and field-validated outcomes that the industry requires.
Founding Team
Most biology AI startups are either biologists who learned to code, or engineers who learned some biology. The Living Models founding team is the rare case where both sides are genuine: a CEO who spent years in the field identifying the exact bottleneck, a CTO with a biochemistry degree from ENS Ulm and a generative AI track record, and a CSO who has already discovered drought-resistant varieties through conventional research — and joined to compress that process from years to months.
Cyril Véran (CEO) built a prior computer vision startup for plant biology — he could see plants, but couldn’t read them. That gap led him to foundation models.
Leonard Strouk (CTO) holds a biochemistry degree from ENS Ulm, conducted research at UC Berkeley and NYU, and previously co-founded a generative AI company.
Bertrand Gakière (VP Biology at Paris-Saclay University), a researcher at the Institute of Plant Sciences Paris-Saclay, joined Living Models to industrialise what he had spent years doing by hand.
The technical team includes PhDs from Huawei Noah’s Ark Lab, Owkin (the Paris-based AI drug discovery company, $300M+ raised), Datadog, Mila, and École Normale Supérieure.
Investors and Funding
The $7 million seed round spans three continents: UC Berkeley’s institutional investment arm, Asterion (Europe), Artesian / GrainCorp (Australia), Galion.exe, Juniper (Silicon Valley), Pascual, Kima Ventures, and Station F.
About Living Models
Living Models develops foundation models for biological sequences — DNA, RNA, multi-omics data. Based in Paris, France and Berkeley, California.
Research paper: https://www.biorxiv.org/content/10.64898/2026.02.23.706817v1
Models: https://huggingface.co/living-models
Asterion Ventures
Based in Paris and Amsterdam, Asterion Ventures is a pre-seed/seed impact-focused VC reshaping how venture capital operates. In just four years, it has mobilized nearly €60M into 30 early-stage startups — four already scaling into Series A/B with top European funds, and one successful exit. Its edge: scale powered by community – a network of 900+ entrepreneurs and executives investing as business angels and backing founders with real experience, networks, and time. Beyond financing, Asterion has built dedicated programs for exited founders and entrepreneurial families. One of the very few European VCs intentionally designed for long-term partnership, Asterion stands for patient, committed, durable support. https://www.
GrainCorp Ventures / Artesian
GrainCorp is one of Australia’s largest integrated agribusinesses, connecting regional producers with global markets for more than 100 years. “The investment reflects our strategy of backing scalable innovation that has the potential to deliver practical value for growers and customers.”
Galion.exe
Seed VC, Europe & US, AI infrastructure focus. “We are entering a decade where the pressure on global agriculture will be unlike anything in modern history. Living Models is building the intelligence layer that changes this equation fundamentally: a full-stack, multi-modal foundation model platform that finally brings the full power of modern AI to plant biology.” — Willy Braun, Founding Partner
Juniper
Early-stage VC, biology & computation focus. junipervc.com “The founding team has assembled one of the largest plant multi-omics datasets in existence. The timing is right, their data moat is exceptional, and the business model fits how the seed industry actually works.” — Jenny Kan
Kima Ventures
Xavier Niel’s investment arm. 1,500+ startups since 2010. Station F, Paris.
Station F
STATION F is the world’s largest start-up campus, founded by Xavier Niel in 2017 and based in Paris. The 50,000m² campus supports over 1,000 start-ups a year. The start-up community already boasts a number of success stories, including the two unicorns Hugging Face and Alan, as well as Yuka, Pasqal, Greenly and many others. In 2025, STATION F companies have collectively raised more than €1.5bn. More info: https://stationf.co/
