Imagine a world where crops are engineered to thrive in drought, resist diseases, and pack more nutrition, all thanks to AI. Sounds like science fiction? It's not. Biographica, a London-based startup, has just secured a whopping $9.5 million to make this a reality. But here's where it gets controversial: can AI truly revolutionize agriculture, or are we placing too much faith in technology to solve complex biological challenges?
Biographica (https://graphica.bio/) is on a mission to transform crop gene-editing using artificial intelligence and machine learning. Founded in 2022 by Cecily Price (CEO) and Dominic Hall, PhD (CTO), the company has developed a cutting-edge platform that identifies high-value gene targets for editing, promising to accelerate the development of improved crop traits. Their recent £7 million ($9.5 million) seed round, led by Faber VC (https://www.faber.vc/), with participation from SuperSeed (https://www.superseed.com/), Cardumen Capital (https://www.cardumencapital.com/), The Helm (https://thehelm.co/), Chalfen Ventures (https://www.chalfenventures.com/), and Entrepreneurs First (https://www.joinef.com/), will fuel their expansion into new crop traits and deepen partnerships across the seed industry.
The company has already partnered with BASF’s vegetable seeds business, Nunhems (https://www.nunhems.com/us/en), though details of their collaboration remain under wraps. “We’ve seen AI reshape pharma, turning trial-and-error pipelines into learnable biological systems. We’re bringing that same discipline to crops,” explains Price.
But this is the part most people miss: despite advancements like CRISPR, which allows precise gene editing, the bottleneck lies in identifying which genes control critical traits like drought tolerance or disease resistance. Traditional methods, such as literature mining and high-throughput testing, are inefficient, with success rates below 1%. This forces seed companies to test thousands of edits, often overlooking promising genes. Biographica aims to change this by using AI to pinpoint the most impactful genes and provide clear editing instructions.
In pilot projects with leading seed companies, Biographica identified proven gene targets 12 times faster than conventional methods. Even more impressively, they uncovered novel targets that traditional approaches missed, potentially bringing new, high-value traits to market. The company is now integrating its platform with rapid experimental validation, creating a “lab-in-the-loop” model that continuously improves through feedback—a concept borrowed from drug discovery.
“With climate change intensifying pressure on agriculture, improving crop genetics is our most powerful tool to sustainably boost yields and build resilience,” notes Sofia Santos, partner at Faber VC.
Here’s where it gets technical—but stick with me. Many firms still rely on Genome-Wide Association Studies (GWAS) and Quantitative Trait Loci (QTL) mapping to link DNA variants to traits. However, these methods only prove correlation, not causation, and fail to prioritize genetic targets. Dominic Hall explains, “GWAS struggles to move beyond ‘this variant is associated with this trait’ to understand why they’re linked.”
Biographica’s platform takes a different approach. Using knowledge graphs and machine learning, it predicts which genes are most influential, how they interact, and which are ideal for editing with minimal side effects. At its core are foundation models trained on multi-modal genomic datasets, capturing both gene-gene and gene-trait interactions. These models predict genes likely to causally impact traits, design edits, and refine predictions based on experimental results—a cycle of continuous improvement.
Over time, this approach promises higher hit rates and faster results. Partners can use Biographica’s insights as a starting point for in-house trait development or adopt experimentally validated traits ready for trials.
Commercial validation has been key to their success. Partnerships with industry giants like BASF and Cibus (focusing on disease resistance in rapeseed/canola) have bolstered investor confidence. “Having contracts with some of the largest seed companies globally definitely helped,” says Price. “These projects also served as technical validation, which is critical in a space where true validation takes years.”
What sets Biographica apart is its crop- and trait-agnostic platform. While many companies tailor AI solutions to specific crops, Biographica’s models are pre-trained on vast public datasets and fine-tuned with their own lab-generated data. “We designed our platform to be independent of customer data from the start,” Price explains. “This approach has allowed us to validate our speed and efficacy while uncovering novel genes partners hadn’t considered.”
And this is the part that sparks debate: as AI becomes more integrated into agriculture, questions arise about data ownership, ethical implications, and the potential for unintended consequences. Is AI the silver bullet for food security, or are we overlooking the complexities of biological systems?
Biographica’s work is already progressing novel targets through downstream R&D stages with multiple partners. With applications across vegetables, tomatoes, chickpeas, oilseeds, and major cereal crops, their impact could be far-reaching.
What do you think? Is AI the future of crop design, or are we moving too fast without fully understanding the risks? Share your thoughts in the comments—let’s start a conversation that could shape the future of agriculture.
