The picture of thousands of salmon moving through a frigid Norwegian pen while underwater cameras silently record their feeding habits, detect lice before they spread, and automatically adjust feed dosage based on an algorithm that has seen more fish than any human crew could ever hope to capture is almost cinematic. It sounds like a demonstration of technology. It isn’t. In partnership with Tidal, a spinout of X, the experimental division of Google’s parent company Alphabet that once discreetly set out to solve the most challenging issues in commercial fish farming, SalMar, the second-largest salmon producer in the world, is currently implementing it throughout its farming sites.
Tidal’s all-in-one camera systems and environmental sensors will be available at several SalMar locations thanks to the partnership, which was announced in late April 2026. Tidal’s platform is constantly receiving data that tracks lice levels, forecasts risk, monitors fish welfare, and assesses the effectiveness of treatments. The autonomous feeding system, which is arguably the deployment’s most immediately useful component, is made to monitor salmon feeding patterns, make real-time delivery adjustments, and reduce the kind of wasteful overfeeding that raises costs and carbon emissions. The ambition here is difficult to overlook, even though it’s still early and no financial terms have been revealed.
It’s worth stopping to consider Tidal’s beginnings. The business was born out of X, the so-called Moonshot Factory, which made daring bets like stratospheric internet balloons and self-driving car technology. Astro Teller, the head of X, stated in public that Tidal is a prime example of the factory’s conviction that solving the most difficult issues first is frequently the quickest route to significant impact. Before Tidal ever got close to a commercial farm, years of development were shaped by that philosophy. It’s possible that patience is precisely what prepared this technology for what SalMar really requires.
The industry’s biggest nightmare is still sea lice, a biological and financial issue that costs producers hundreds of millions of dollars a year and has eluded simple fixes for decades. Although the autonomous in-pen lice mitigation system is still in the validation stage, SalMar and Tidal are currently testing it. It is truly unknown if it operates at scale. However, the willingness to test it in actual industrial settings as opposed to a controlled research setting indicates that both businesses think this is something worth pursuing.
Observing this collaboration take shape gives the impression that the aquaculture sector is at last reaching a point that the larger agriculture sector reached years ago, where automation and data ceased to be accessories and began to function as the main operations. Gustav Witzøe, chairman of SalMar, put it simply: the future of salmon farming could be shaped by automated feeding and more precise site data. It wasn’t oversold by him. That self-control seems intentional and most likely well-deserved.
The collaboration’s wider reach goes even farther. Large language models created in collaboration with Google will be used by the two businesses to support breeding initiatives, track fish growth, and identify individual fish. Even five years ago, that level of biological granularity would have seemed unrealistic for an open-water pen. It’s still unclear how that layer of the platform will operate in the day-to-day chaos of a working farm site, so it remains to be seen if the LLM integration actually fulfills that promise.
It seems certain that the industry is keeping a careful eye on things. The use of AI robotics by a Google-affiliated technology company in one of the biggest salmon operations globally is the kind of development that sends a message beyond Norway. This could be the start of something much bigger. It could also be an extremely costly and complex experiment. It’s both at the moment.
