
Methods
Machine learning and quantitative finance, applied to salmon.
We develop machine learning models for what cannot be observed directly, from biomass to harvest timing, and carry their uncertainty through to the supply forecast.
From sea to market
Tidepoint applies machine learning to physical data from across the value chain, producing real-time supply estimates and forecasts.
01 · Tidepoint Terminal
What is happening now, what it means for supply, and where supply goes next. Get the full picture of the salmon market.
01Now
Harvests, wellboat traffic, slaughterhouse activity, sea lice, and sea temperature across every site, as the data lands.
02Supply
Standing biomass, sea temperature, feed data, and harvest schedules, read against observed export volumes.
03Forecast
Where supply goes next for every weight class, pen, farm, and company, with its interval attached.

Methods
We develop machine learning models for what cannot be observed directly, from biomass to harvest timing, and carry their uncertainty through to the supply forecast.
02 · Estimates
An estimate for every weight class, pen, farm, and company, added up to the week's harvest and carried forward into the weeks ahead. Each one carries its error band and is scored when the week settles.
03 · Value chain
The estimates say where the week lands. Tidepoint Terminal shows how it gets there, stage by stage from pen to price, updated as new data comes in.
01Sea
Standing biomass, sea temperature, and sea lice, site by site.
02Harvest
Wellboat trips from site to slaughterhouse, as they happen.
03Slaughter
Activity at each slaughterhouse, day by day.
04Export
Volumes leaving Norway, as they are reported.
05Price
Settled prices, as they are published.