Avalo bets machine learning can breed climate-resilient crops without editing genes

Durham, North Carolina startup that uses machine learning to speed crop breeding, without gene editing or traditional breeding.

If ML can find the genetics behind complex traits, breeders can pick winning seedlings by genotype instead of waiting to see traits in the field.

$3M seed round, followed by another $3M.

The process builds on chief science officer Mariano Alvarez's post-doctoral studies at Duke. TechCrunch had done a deep dive two years earlier, when Avalo had only a $3M seed round; it has since raised another $3M.

Avalo does not edit genes or breed varieties the traditional way. It uses ML models to identify the genetic basis of complex traits such as heat tolerance. Because it genotypes all seedlings, CEO Brendan Collins said, it knows which will be winners without waiting for the trait in the field. Seedlings can be bred in growth chambers and greenhouses under accelerated conditions, giving four development cycles a year for most row crops instead of one.

Avalo says it proved the process by creating a fast-maturing broccoli for vertical-farming startup Iron Ox, though the effort collapsed when the vertical-farming bubble popped in 2023 and the broccoli cannot yet be tried. Avalo was still working with greenhouses to market it, and shifted focus to a latex-producing dandelion, trait licensing in soy and corn, and a new drought-tolerant cotton project. It announced plans to raise a $10M Series A that fall.

Creating climate-resilient crops is time-consuming and costly, and field testing of crosses usually happens once a year.

Genotyping seedlings lets Avalo identify winners and losers early without waiting to see a trait expressed in the field.

Accelerated growth in chambers and greenhouses gives about four development cycles per year for most row crops.

Collins argues corn gained a 300% yield increase since the 1950s because of heavy investment, and wants to give other crops that level of resource at a fraction of the cost.

A technique proven for one customer can be stranded when that customer's market collapses, so spread proof points across several crops and buyers.

As of 19 September 2023, Avalo had raised about $6M in total and intended to raise a $10M Series A that autumn. Its Iron Ox broccoli could not be tried because the customer's effort collapsed, and it was working on a latex-producing dandelion, soy and corn trait licensing, and drought-tolerant cotton. TechCrunch named Keygene and Benson Hill as direct competitors.