Can AI disrupt the seed giants? The start-up betting on climate-resilient crops

Jobs and Wozniak. Gates and Allen. Great things can start with two guys playing around on a computer. Brendan Collins and Mariano Alvarez, the founders of Avalo, the young US start-up developing new climate-resistant crops, hope to prove that once again. “The goal is to be the world’s fifth biggest seed company,” says Collins, now Avalo’s CEO. “But for good.” Collins and Alvarez met through a book club at Michigan’s Duke University. Collins was a biologist moving into the start-up world while Alvarez had just started a post-doctorate in computational evolutionary biology. The pair quickly bonded and began meeting up after work to tackle complex computational biology problems. Through Alvarez’s work, they got to know computer scientist Cynthia Rudin – a “god tier statistician,” according to Collins - who’s work understanding black-box AI models was an inspiration for the two biologists. They recognised the potential in Rudin’s form of machine learning to understand gene combinations behind traits such as flavour or drought-resistance which due to their complexity, have long-escaped even the finest plant breeders. “We took [Rudin’s] idea and applied it to population level genetics. To use it as this discovery tool,” Collins explains. “It’s the kernel of what we’ve built and allows us to do advance breeding programs much faster than normal.” Avalo’s idea is effectively what’s known as ‘marker assisted breeding’, a common technique of extracting DNA from a seedling to test for the desired genes. The difference, Collins explains, is that thanks to Rudin’s ideas Avalo can now “take that concept and do it on steroids.” Avalo only launched in 2020 but it is already notching up some notable wins. Its flagship project was with cotton farmers in Texas concerned that plummeting water levels are already triggering major declines in the harvest, Collins explains. The core problem is that many use a generic seed capable of producing huge volumes of cotton but only if there’s plenty of water and fertiliser. Avalo recognised that most wild cotton grows in arid, desert conditions and it has only grown to be water-hungry under human influence. Collins and Alvarez therefore sought to expand the gene pool by collecting 600 cotton varieties from universities and seed banks around the world for an experimental breeding programme back in the US. After growing the seeds and collecting data on their performance, Alvarez and Collins trained new AI model on the genomes, giving them an idea of exactly which genes were actually responsible for drought tolerance. This ability to identify complex genetic combinations already goes beyond the capabilities of many breeding programmes. But the thing that Avalo believes could truly set it apart is the speed and cost of its technology. “Our goal as a company is to reduce the cost of high-end breeding so much that basically any crop could benefit from these programs,” says Collins. That would truly be revolutionary. Avalo still has much to prove but in the most extreme case, neighbouring farmers with slightly different soil conditions could each use unique seed tailored for their exact farm, Collins explains. “Our vision is to always have this core of diversity that we pull out to create regionally distinct varieties,” he says. “And so it’s a system where those crops are actually designed to be there. We’re not trying to turn all these places into an unrealistic Eden.” The potential has caught the eye of Coca Cola who launched a partnership with Avalo last year to breed a new type of sugarcane that needs less nitrogen fertiliser for its farmers in Australia. For Collins, what is particularly exciting about sugarcane is that its genetic complexity has so far put it beyond the realms of nearly all commercial breeding attempts. “It’s just crazy. Since the 1950s, corn has seen a 300% yield increase per hectare. Everywhere except Brazil, sugar cane has seen yield per hectare actually go down,” he says. Bayer, Corteva, Syngenta, and BASF – ‘the big four’ seed firms – are unlike to be too concerned just yet. As Collins admits, Avalo’s sugar cane programme is still too early to report any real-world impacts. But there is also a widespread perception that the seed giants hold little interest in developing new seeds anyway, beyond the narrow criterion of boosting yield. If this involves corn, all the better, says Collins. “Especially within the US, there’s one crop they care about, which is corn. That’s taking up 90% of their R&D.” In the seed giants’ defence, breeding programmes have long been limited by an inability to select for several complex traits. If you wanted drought resistance you would likely have to sacrifice significant yield. It is only thanks to recent advancements and the falling price of genome sequencing that a company like Avalo is possible at all. That trade-off is far from settled and most breeders still say a small drop in yield is unavoidable. That often makes it tricky to persuade farmers to buy these new innovative seeds instead. But Collins argues there is a clear difference between Avalo’s approach and the big seed firms. Avalo works with farmers to understand their budget, the amount of fertiliser they use, and how much irrigation is needed all so they can develop a seed which is both more sustainable and better for profitability, he argues. “We can increase that crop diversity, really focus on farmer profitability and lower overall externalities caused by agriculture as a whole.” That’s a big difference to the big seed firms who have chiefly sought to boost sales by hiking yields, he adds. Combined with the fact that much of their revenue comes from agrichemicals rather than seeds, this means they’re “not that concerned with making a more drought resilient cotton.” It is therefore not unreasonable to conclude – as Avalo has – that this is a market ripe for disruption. But sceptical farmers will not be persuaded easily. Avalo must somehow prove they are worthy of the chance.

Can AI disrupt the seed giants? The start-up betting on climate-resilient crops

Related Topics

Inside HiveTracks: Can bees become the ultimate biosensor for nature?

Switzerland’s agtech paradox: world‑class innovation, limited scale

Why BBLeap believes crop spraying is fundamentally broken