From left: Aquaticode's Christoph Podes (chief operating officer), Stian Rognlid (chief executive), and Adriana Castellano Cardenas (automation manager) in Chile.

10,000-smolts-per-hour gender sorter launched in Chile

Published

A device which uses machine learning to automatically separate salmon smolts by gender and remove unviable fish is being launched commercially in Chile after being assessed over time and at scale by fish farmer Salmones Camanchaca.

SORTpro, created by Norway-based Aquaticode, is designed to categorise up to 10,000 juvenile salmon per hour based on gender and other performance traits.

The artificial intelligence (AI)-powered solution achieved a verified gender classification accuracy exceeding 97%, and didn’t increase mortality rates, Aquaticode said in a press release.

Manual ultrasound operators currently sort around 100 million fish annually in Chile. Gender sorting has proven to promote faster growth in both male and female batches compared to mixed groups. By automating this procedure, Aquaticode said it not only enhances flexibility, efficiency, and accuracy but also extends gender sorting capabilities to regions where high labour costs previously made it unviable.

"Camanchaca, which historically sorts around 15 million fish per year, has given us invaluable support in bringing our solution to the market. We appreciate its openness in sharing its facilities and expertise with us,” said Aquaticode chief executive Stian Rognlid.

Aquaticode's SORTpro can separate 10,000 smolts per hour into male and female cohorts.

Aquaticode’s R&D efforts at Camanchaca, dating back to 2021, originally began with the development of an early proof-of-concept alongside the manual sorting process. Aquaticode said that over the years, Camanchaca’s contributions have led to highly accurate classifications not only in gender but also in early sexual maturation and deformities.

Salmones Camanchaca is Chile’s seventh largest salmon farmer by volume, and expects to harvest between 55,000 and 59,000 tonnes of Atlantic and coho salmon this year

“Innovation is central to our strategy at Camanchaca,” said Carlos Soto, head of R&D. “Artificial intelligence is emerging as one of the most promising megatrends in food production. We are quick to support initiatives that can lead to more efficient and sustainable salmon farming.”

More functions

Addressing the importance of smolt quality in fish welfare and performance, Rognlid expressed ambitions beyond the SORTpro’s current capabilities. “The SORTpro is a platform. In the near to medium term, we will introduce functionality to identify additional fish growth, health, and robustness traits. Each of these will bring significant value to farmers, but we truly believe that the sum is even greater than the parts,” he said.

Founded in 2018 and backed by Nacre Capital, a venture builder focused on AI in life sciences, Aquaticode raised US $6 million Series A funding in 2022. Its advisory board includes former Cargill Aqua Nutrition chief executive Einar Wathne.

“It was easy to get behind Aquaticode’s mission,” said Wathne, now chairman at the NCE Seafood Innovation Cluster in Bergen, Norway. “Making decisions based on individual fish analysis and acting in real-time is the future of salmon farming. It’s about growing better fish as well as growing fish better.”