Jellyfish, sea lice and harmful algal blooms are among the hidden hazards threatening salmon farms that FP Aqua’s technology is designed to detect © Bob Brewer, Unsplash
The biggest threats on salmon farms are hard to see and cause major economic losses. FP Aqua* is building an around-the-clock microscope system that lures zooplankton with light, samples algae every five minutes and flags up jellyfish in the water column, so farmers can act before problems hit their pens.
Founded in December 2023 in Denmark by Frederik Taarnhøj, FP Aqua is a technology spin-off from Evolito (formerly Fauna Photonics A/S), a Danish technology company that develops and provides automated, real-time insect monitoring systems for biodiversity and environmental monitoring.
“Frederik always had this idea that there was something in aquaculture that we should go for,” recalls Jesper Lemmich, research development lead at FP Aqua.
The FP Aqua team had identified an opportunity in aquaculture’s growth and rapid digitalisation, and ventured into developing a water-monitoring sensor for fish farmers.
“The FP Aqua acquired the core technology from Fauna Photonics A/S. We brought a patent, some early prototypes and a few team members with us. It was a good start,” says Taarnhøj.
Spotting trouble before it strikes the farm
Atlantic salmon farms face a range of biological threats. Sea lice remain the most persistent, reducing fish welfare and making them more susceptible to disease. Jellyfish can cause sudden mass mortalities – in 2024, a barbed-wire jellyfish (Apolemia uvaria) bloom killed nearly 200,000 salmon in Scotland. Harmful algal blooms (HABs) also pose a major risk, with past incidents in Norway leading to widespread stock losses. These events create serious financial and fish welfare challenges.
FP Aqua’s system is being developed to detect these problems earlier, giving producers time to deploy preventative measures.
Rather than relying on a single method, the platform runs continuous monitoring tailored to each threat. The “live” module lures positively phototactic (light attracted) zooplankton with light, photographs them and uses AI to classify what is present – for example, detecting free-swimming infectious stages of sea lice (copepodids, measuring around 0.7mm) before they reach or as they leave the pens. For the phytoplankton module, a micro-sampling microscope draws in water every five minutes, records images of cells as small as 10 µm, then flushes and repeats – building a near-real-time picture of bloom species and dynamics.
FP Aqua completed proof-of-concept testing with commercial farms in December 2024. From there, they agreed to a new concept design that salmon farmers have nicknamed a “Swiss army knife” of sensors.
“It was clear from our interactions with salmon farmers that they want one sensor. They don't want 10 different sensors that you have to employ differently, because that makes life very difficult for them,” explains Lemmich.
The collaboration will continue, in Scotland, where trials will focus on jellyfish – an increasingly frequent issue in local waters.
Finding the right place in the water
According to Lemmich, the sensors are robust and built for the rough seas that exposed Atlantic salmon farms face. The aim is to deploy at least three units across a site so different observations can be made. Currently, the team places sensors near feed barges, close to the cages, but other locations could also have their advantages.
“The best would be to have a sensor placed a little bit away from the pens in the direction of the current so that you can see what's coming in. If we can see algae or jellyfish coming into the fjord, say a kilometre from the pens, then we can make a warning and the farmers can start doing different preventive measures like raising bubble curtains,” explains Lemmich.
© FP Aqua
Another key position would be downstream in the current, allowing farms to monitor what leaves the site and build a fuller picture of the processes at play.
“With our technology, you can also start to understand how long a threat is likely to be present at a site. Currents bring threats into the site, but currents can also take threats out of the site. And how does this all happen? Nobody really knows. But with our technology you can start to get that knowledge,” adds Taarnhøj.
Current HAB monitoring technologies often rely on satellite imagery or manual sampling at the farm. FP Aqua say that the advantage of their system, aside from continuous monitoring and early warning, is the ability to detect organisms at different depths and build knowledge. There’s also scope to learn how far down the water column these hazards operate and, as Taarnhøj notes, to pair that knowledge with preventative tactics such as submerging sea cages – optimising depth to target the layers least at risk.
This dataset can also train farm crews and improve decisions. As Taarnhøj puts it: “We've observed some quite difficult biological challenges that have given some of the salmon producers some unforeseen financial costs and these can be devastating, depending on how hard they hit. If we can make some of this a little bit more predictable by offering data and knowledge on it, I think, over time, the salmon producers will be able to find ways to better circumvent this unpredictability and be able to manage some of these mortality events.”
Built for the sea
Designing a microscope to operate for long time periods at depths of 40 metres, with a design capacity down to 200 metres, meant addressing both saltwater corrosion and pressure in offshore environments.
“The underwater microscope’s housing is made from marine-grade aluminium with a hard-anodised finish to resist saltwater corrosion, and further protected by sacrificial zinc anodes to prevent galvanic wear. Imaging is performed through high-strength quartz glass windows, chosen for their clarity and durability under pressure,” details Knud Poulsen, FP Aqua’s technical lead.
To keep the system reliable in rough seas, marine-grade gaskets prevent water ingress and the electronics are shock-mounted and thermally managed to withstand wave action, pressure changes and temperature swings.
“One of the key lessons was the importance of sealing and pressure management,” says Poulsen. “We also learned that biofouling could quickly degrade image quality, so we integrated UV-C illumination to keep the optical windows clean and extend deployment time.”
© FP Aqua
What’s next?
FP Aqua has chosen to initially focus on the Atlantic salmon sector, but they think there is also significant potential for their technology in other applications. From a regulatory perspective, it could one day support continuous monitoring of sea lice escapes from around the farms, providing documentation for both farmers and authorities.
“In the images we analyse daily for AI training, we not only detect harmful species of interest but also a wide variety of other taxa, including crustaceans, echinoderms, polychaetes, and marine snails, that may provide valuable insights into ocean health and biodiversity,” says Zandra Maria Skandrup Sigvardt, lead biologist at FP Aqua.
The team has also had promising discussions about how their platform could be adapted for shrimp farming.
“We could certainly see our basic live sensor being used in the shrimp industry. It would have to be cheaper, but also won’t have to live in a rough sea so we could design it in a different way. That is definitely something we will be looking into because this market is a lot bigger and it's developing very fast,” says Lemmich.
He stresses, however, that the company’s immediate priority remains salmon.
A key technical challenge is minimising false positives – instances where the system incorrectly flags something as a problem. For example, under a certain angle or lighting, the AI might detect a copepod and mistake it for a sea lice or be inconsistent between two similar-looking species of gelatinous transparent jellyfish. To tackle this, the company is committed to ongoing AI training, continually feeding the model new images to improve accuracy and aim for close to zero false positives.
Like most early-stage ventures in deep tech, financing also remains a hurdle. But for Taarnhøj, it’s simply part of the startup journey, and one that motivates rather than deters.
What really matters now, he stresses, is proving the value of their systems on salmon farms.
“Priority number one is to finally finish developing the solution and show it has some operational value – not just with test results and field trial results, but actually having a good one-to-one conversation with the site personnel and them agreeing that it is really helping them manage their operations and hopefully also giving them some economic benefit,” he concludes.
*FP Aqua is one of Hatch Blue’s portfolio companies, but The Fish Site retains editorial independence.
