Marine Farmers Deploy AI Underwater Drones to Monitor Submerged Crops

South Island marine farmers are deploying AI-powered drones and facial recognition for fish to boost efficiency and reach a $3 billion revenue goal by 2035.

By: AXL Media

Published: Apr 15, 2026, 7:59 AM EDT

Source: RNZ Pacific

Real-Time Analytics from the Ocean Floor

The Christchurch-based firm UWAI Robotics has spent eight years developing a specialized underwater drone, affectionately nicknamed "Crabby," designed specifically for the greenshell mussel industry. Traditionally, assessing the quality of a mussel crop required farmers to physically haul heavy lines out of the water and manually shuck shellfish to inspect meat quality. The new drone utilizes AI and advanced imaging to scan these lines in situ, providing a live video feed to operators on the surface.

This technology distills raw visual data into essential metrics such as mussel count, size, and distribution variability. By identifying whether mussels are clumping or well-distributed, farmers can make immediate, informed decisions about reseeding and harvesting schedules. According to UWAI Robotics co-founder Tim Rensen, the objective is to provide a comprehensive look at the condition of the mussels and the biofouling levels on the lines without disturbing the marine environment.

Enhancing Supply Chain Accuracy and Efficiency

For operational managers like Maegan Blom of Mills Bay Mussels in Havelock, the accuracy of this AI-driven data is transformative. In the fresh live mussel market, providing customers with product that matches specified quality metrics is critical for maintaining brand value. The ability to receive an instant graph showing mussel sizes directly from a drone scan allows for precision in harvesting that years of manual experience could not always guarantee.

While farming in the open ocean involves uncontrollable variables such as water temperature, weather events, and predators like snapper AI allows farmers to control their response to these factors. By identifying patterns and trends in crop growth over time, companies can optimize their seeding density and choose the most effective windows for harvest. This data-centric approach moves the industry away from anecdotal observations toward a more industrial, efficient model of aquaculture.