Can AI deliver new crops in just four years?

Heritable Agriculture, a spinout from Google X, is using artificial intelligence, genomics and environmental data to dramatically reduce crop development timelines, potentially bringing new varieties to market in just four years, according to AgFunderNews.

The company’s approach is based on digital twins, virtual models that simulate how real plants grow in real environments. These models integrate high-resolution data on soil, weather and field conditions, allowing researchers to test crops without waiting for full growing cycles.

CEO Brad Zamft explained that the goal is to create a comprehensive system for crop improvement. “What Heritable provides is a comprehensive solution to improving crops,” he said, highlighting the ability to simulate plants and identify key genes with high precision.

The technology goes beyond traditional breeding methods. By combining AI with genomics, the company can identify which genes control specific traits and how they should be activated depending on environmental conditions such as drought or heat.

This approach is inspired by advances in large language models. Just as AI can understand human language, Heritable applies similar principles to decode the “grammar” of DNA. This allows researchers to pinpoint the exact genetic sequences responsible for desirable traits.

Validation has been a key focus. According to Zamft, the models have been tested in real-world conditions. “None of that matters unless it works in real plants in a real environment,” he said, noting that the company has successfully validated its predictions across multiple species.

The system has already demonstrated the ability to simulate crop performance in new environments with high accuracy. It also supports gene discovery and has been tested through full cycles that include field trials, tissue sampling and gene editing.

Unlike traditional agricultural R&D, which often focuses on major crops like corn and soy, Heritable is targeting a broader range of species. This includes fruits, vegetables and forestry crops, where genetic knowledge is still limited.

The business model is flexible. The company offers software licensing to breeders, as well as trait development partnerships and royalty-based agreements. One of its most notable collaborations is with Red Sun Farms to develop a new strawberry variety.

Zamft made a bold prediction about the impact of this technology. “I think we’re going to get a new strawberry variety on supermarket shelves in four years from start to finish for a few million dollars,” he said.

The broader transformation is driven by three converging trends: the falling cost of DNA and RNA sequencing, the explosion of environmental data from satellites and sensors, and the rise of AI capable of integrating complex datasets.

Together, these advances are enabling a new model of agriculture where crop development is faster, cheaper and more precise. This could have major implications for food systems, especially in regions facing climate challenges.

The company is already working with major industry players such as Syngenta and KWS, as well as receiving support from the Gates Foundation to improve crops for smallholder farmers in Sub-Saharan Africa.

As agricultural innovation becomes more data-driven, technologies like digital twins could redefine how crops are developed, making it possible to respond faster to global food demands and environmental pressures.