Editor's note
Five slots this week. Far more stories.
The shortlist had ten candidates before I added three of my own, and cutting it to five was harder than usual. Several of the stories that stayed out could easily have run, and some may come back as they develop. If you saw something this week that deserved a slot, reply and tell me. The next shortlist starts on Sunday.
My own site has widened too. woodcomp.world now covers pellets alongside pulp mills and sawmills. If pellets are part of how you buy, sell or plan wood, have a look and tell me what is missing.
Axel
Kodama raises $7M to scale remote-run skidders
Kodama, last covered in #042 when Weyerhaeuser piloted its system, has raised a $7M seed round co-led by Konvoy Ventures and Breakwater Ventures. It makes no machines. Its cameras, lidar and software go on skidders from the big equipment makers: the skidder drives the routine runs to the landing itself, and a remote operator, who can sit at home, steps in when the work gets hard. Each remote shift also trains the autonomy further, and the operator no longer has to live near the cut.
Teknosavo says its software cuts pulp mills' wood use
Finnish Teknosavo says pulp mills running its system use 3 to 5 percent less wood for the same output. Its AI-based software ties automated wood measurement at the mill gate to live readings from the yard, debarking, chipping and chip monitoring, and adjusts each step to what is actually arriving rather than to averages. Wood is a mill's largest cost, which makes even a few percent worth chasing.
Södra asks outsiders to solve browsing damage
Södra is looking outside forestry for an answer to browsing damage and has opened an innovation challenge for new ways to protect seedlings and young stands. Repellents already on the market are ruled out. It wants new materials, sensors with AI and image analysis, light or sound, biological methods, or products from other sectors that could be adapted. The strongest ideas get tested and developed with Södra. Applications close on 30 November.
Arbonaut splits Boxholms Skogar's stands into micro-stands
Finnish Arbonaut has split the stands on about 4,000 hectares of Boxholms Skogar's estate in Östergötland into smaller, more uniform micro-stands. Remote-sensing layers sort each stand first by moisture class, which governs when it can be harvested, then by its tree layer, and finally flag impediments, afforested farmland and likely nature values. The old register followed the trees; the new boundaries follow the soil too, which the companies say should speed up operational planning.
Chalmers reads conservation value from canopy height alone
Researchers at Chalmers trained a deep-learning model to sort forest in southern Sweden into high and low conservation value using only a 1-metre canopy height layer from the national laser scan. Accuracy jumps at 2 metres or finer, where single trees become visible. The test is easier than it sounds: high value meant stands already registered as key habitats or Natura 2000, low value meant young or recently cut forest, and the ordinary forest in between was never scored. A first screen from one nationwide dataset is still worth watching for.
From the Forestry Tech Landscape
The Forestry Tech Landscape maps 100+ companies and startups building technology for forests. Three to know this week:
NEW · Mazex (Planting & silviculture). Transport drones for steep forestry sites, carrying seedlings, tree shelters and supplies, sold through Sumitomo Forestry and forestry cooperatives in Japan.
Kodama Systems (Harvesting & machinery). Teleoperated and autonomous forestry machines for thinning fire-prone forests, combined with biomass burial for carbon removal.
Flash Forest (Planting & silviculture). Drone swarms and seed pods for planting trees at scale after wildfires and deforestation.