Editor's note
This week opens with a call-out.
Moritz, a researcher at NIBIO in Oslo, is writing a review of agentic AI in forest management, meaning software that plans and carries out a series of steps on its own rather than answering a single question. He interviewed me for it, and afterwards asked who else he should talk to. That is a question this list can answer better than I can.
If you are building an agent for forestry work, testing one, or have written anything about one, published or not, he would like to hear from you. Reply to this email and I will put you in touch.
I will forward whatever comes in, and when the review is out I will tell you what it found.
Axel
Metsä lets AI quote young stand work across Finland
The price is set before the job starts, not looked up in a rate table, and it stays put once the work is done. The application reads satellite imagery alongside working-time records from sites already cleared, then prices pre-commercial thinning site by site. We covered the launch in January; September is when it reaches every region. That puts it in front of about 90,000 member owners, who hold roughly half of the country's private forest.
Finland audits subsidised forest work from orbit
Metsä prices young stand work before it starts; the Finnish Forest Centre now looks at it afterwards, comparing vegetation in satellite images taken before and after each job the state has paid towards. Since spring 2025 that has turned up about 1,500 hectares where the grant was not justified, and around 200,000 euros has been recovered or held back. The Centre counts 450 hectares that missed the conditions, and 64 where nobody had done the work at all.
BEV_r and Virkeslogistik put Swedish timber haulage into a simulator
BEV_r's plan is to try each change on a digital twin of the fleet before a truck moves, with loads, driving and rest rules, delivery windows and energy use modelled together in its INSIGHT platform. The partners name route planning, fewer empty kilometres and electrification as targets. Forest transport is a hard case, says product chief Tony Edmundsson, because the potential has been hard to see and put a number on. It will be interesting to see what the first results show.
RIEGL's new drone lidar hears five returns from every shot
A laser pulse fired into a canopy hits leaves, branches and, with luck, the ground, and each surface sends back its own signal. The miniVUX-4UAV, shown at INTERGEO in Munich last week, logs up to five of them per pulse, which is what matters under closed crowns, where the terrain is only reached by what slips through. RIEGL puts its accuracy at 12 millimetres and its weight at 1.55 kilograms. The company has not said when it ships.
American Forests opens its forest risk tool to every kind of US landowner
Private, state, Tribal, federal and NGO land now sit in one open-source system that layers more than 40 datasets, by the group's own count, to show where wildfire, drought, water stress and pests threaten a stand. An assessment that took months or years now takes minutes, American Forests says. The science underneath comes from the USDA Forest Service, and the project was first described by USDA's Climate Hubs in 2024.