The Why
Return on capital employed (pre tax ROCE, NOK) for leading Norwegian salmonids farmers has fallen from ~20% in 2016 to <10% during 2025 and early 2026. Over the decade through to Q3’26, price achievement contributed positively to returns, but returns compressed due to rapidly increasing production cost and increased capital intensity of production.
The high cost inflation in the industry over the past 10+ years has been primarily driven by lower productivity on inputs (higher use of e.g. vessels per harvested kg), higher raw material cost for suppliers (particularly in feed through both FX and commodity price increases), and shifting supplier economics.
As of Q4 2026, a return recovery into 2027 and 2028 is on the cards, with salmon price growth and biological- and FX headwinds reducing impact of feed prices reaching all time highs. Beyond favorable seawater temperatures, this improvement is mostly down to historical investments starting to yield results, including from post smolt, new vaccines against reemerging challenges, preventive- and proactive sea lice treatments.
If the return recovery manifests, it will not be due to farmers fundamentally reshaping decision making processes. The question farmers should ask now is: How do we deliver sustained cost and return improvements? Digitalization and AI should be the core pillar to doing so.
The What
Salmonids farming remains an industry with a high degree of experience-based decision making. Data volume and quality is growing rapidly yet is rarely effectively analyzed to isolate what drives outcomes, or leveraged in day-to-day decision making. Digitalization through AI and agentic will increasingly allow decision making to be faster and more provable.
BCG works shoulder to shoulder every day on AI- and agentic transformations at the leading companies globally, seeing the value it creates. Salmonids farming lag other comparable industries on digitalization, including peers in other protein farming and agricultural industries. Leading companies in the latter sectors are already reshaping yield achievement in processing, feed selection and practices, sales and operations planning, among several other examples to improve profitability and returns.
Farmers should structure their digitalization efforts around five overarching priorities:
- Sales and operations planning (S&OP) that links harvest, processing, logistics and sales
- Biomass planning and site outfitting
- Feeding operations, and the selection process of roe, vaccines and feed recipes
- ROI based treatment and fish health decision making
- Business process automation
Together, these priorities hold a potential above 6 NOK/kg EBIT improvement for salmonids farmers in the coming 3 to 5 years.
The How
The supplier industry features several highly innovative companies effectively supporting farmers with successful initial steps of digitalization. Suppliers are best placed to help farmers where a single data source answers a given question. This includes biomass monitoring, start-stop feeding based on pellet detection, and preventive sea lice management. Here, farmers should buy what they see ROI on.
However, delivering value from digitalization comes down to improved decision making and not improved access to data. In salmonids farming, the highest impact decisions are complex and multidisciplinary. Data need to be drawn from multiple sources and suppliers simultaneously to effectively digitalize or improve these. To realize the value in such processes, farmers need to orchestrate or directly build the decision support or automation solutions themselves.
Historically, the cost for developing industry-bespoke software has been a major bottleneck for niche industries. Now, AI is dramatically reducing the cost and effort of building proprietary software, and prototypes that used to take months now take weeks. Farmers that manage to combine internal builds with buying supplier solutions with proven ROI will develop a competitive advantage. This paradigm favors larger companies with scale, but mid-market operators still have a path to catch up.
Technology is therefore no longer the bottleneck. BCG’s experience point to 70% of the success in AI transformation relating to navigating people, organization and process implications. About 20% comes from data and technology, and 10% from algorithms. To get going today, farmers should:
- Have a clear priority list of key value opportunities, and on what to buy vs. build
- Pick a few big bets focused on influencing the high value decisions along the value chain. Don’t work incrementally through several pilots of smaller use cases
- Ensure internal builds fit with how the relevant process will be run 2-3 years from now to avoid significant technology debt or rapidly obsolete solutions
- Staff each bet with a small cross functional teams. Ensure the relevant P&L owner is in the lead with accountability for value delivery, with IT teams enabling but not owning
- Establish a common context layer leveraging all available data from all suppliers that all internally built use cases can reuse