Wildfire AI in Northern Ontario: what Canada is building and where the gaps are

Canada is developing AI-powered wildfire tools — from aircraft-based prediction systems to a satellite constellation — but most are in research or deployed in BC and Alberta. Here is what exists now, what is coming, and what it means for Northwestern Ontario operators and communities.

AI for wildfire detection and response is advancing in Canada, and the 2026 season made the stakes concrete for Northwestern Ontario. Between January and late July 2026, 637 fires burned 737,892 hectares across the Northwest — including the Thunder Bay 36 fire in the Wabakimi Provincial Park area, which reached 318,813 hectares and is among the largest fires in Ontario's recorded history. Nationally, 3,915 fires had burned approximately 3 million hectares by July 21. The technology to better detect and respond to those fires is being built. The most capable tools — a dedicated satellite constellation, AI-powered aerial prediction models, and sensor networks that detect fires in minutes — are either still in research, undergoing a program rebuild, or deployed primarily in BC and Alberta. Northwestern Ontario is in the coverage gap.

What 2026 showed about fire detection in the Northwest

At its peak, Northern Ontario had 191 active wildfires. The Thunder Bay 36 fire alone covered more than 318,000 hectares in Wabakimi Provincial Park. Evacuations displaced people from dozens of First Nations communities across the Northwest, and the Namaygoosisagagun First Nation community was destroyed. Smoke from Ontario fires pushed south, and air quality in Toronto reached levels among the worst of any major city globally in mid-July 2026. Ontario's response involved more than 155 fire crews and more than 80 aircraft. The scale of the 2026 season — significantly above Canada's long-term annual average of about 8,000 fires and 2.1 million hectares — has sharpened urgency around detection and prediction tools that are not yet in place for Northern Ontario.

Three layers of AI in Canadian wildfire response

Canadian wildfire AI works at three layers. National monitoring: Natural Resources Canada's Canadian Wildland Fire Information System (CWFIS) provides daily national wildfire maps and reports using data from existing satellite sources, supplemented by the Canadian Forest Fire Danger Rating System (CFFDRS) — a set of scientific models for assessing fire danger and estimating fire behaviour. This layer is operational today. Aerial AI prediction: TankerVision, a research project from the University of Toronto and Natural Resources Canada, uses AI to analyze infrared footage from aircraft and predict where a fire will spread. This layer is in research and early validation. Satellite fire monitoring: WildFireSat, a joint federal initiative, aims to provide daily tracking of all active fires across Canada from a dedicated satellite constellation. This layer is in development, facing a program rebuild, with a target launch of 2029.

TankerVision: aircraft-based AI fire prediction, heading toward Ontario

TankerVision is led by Professor Steven Waslander at the University of Toronto Institute for Aerospace Studies (UTIAS), in collaboration with Natural Resources Canada, Simon Fraser University, BC Wildfire, Alberta Wildfire, Canadian Forestry Services, Conair Aerial Firefighting, and monitoring service Voxelis. The system installs high-resolution colour and infrared cameras in bird-dog aircraft — the small planes that monitor fires from the air. As the aircraft flies a fire, AI segments footage to identify which pixels correspond to fire, smoke, or ground, then maps that data onto 3D terrain models using satellite imagery. The combination of aerial observation and terrain mapping feeds a predictive model for where the fire will spread, combined with weather and wind data, supporting better decisions about evacuation timing and firefighting resource placement. In 2024, the team logged data from 50 fires, described as the highest single-season record for the project at that time. The team has stated plans to expand partnerships to Ontario. A full-system research paper and public dataset are expected in fall 2026.

WildFireSat: Canada's satellite fire mission is being rebuilt

WildFireSat is designed to be the first government-owned satellite mission specifically for monitoring all active wildland fires across Canada on a daily basis. Funded in Budget 2022 as a joint initiative between Natural Resources Canada, the Canadian Space Agency, and Environment and Climate Change Canada, it targets a 2029 launch. In February 2025, the Canadian Space Agency awarded Spire Global Canada a $72-million (Canadian) contract to design and build the satellites. That contract was terminated by the CSA "for convenience" in 2026. CSA issued a Request for Information to reengage the satellite industry and has stated the mission will remain on budget and on schedule — but no revised confirmed timeline has been announced. In the interim, the federal government established a $6-million WildFireSat Remote Sensing Capacity program to help provinces and Indigenous agencies better integrate existing satellite data into fire management operations. Northern Ontario agencies considering this program should contact Natural Resources Canada directly about eligibility and intake.

What this means for a Northwestern Ontario operator or administrator

For forestry operators and remote-operations businesses in Northwestern Ontario: the fastest-detection AI systems — SenseNet's integrated AI camera, gas sensor, and satellite networks, which the company reports can detect fires in as little as three minutes — are currently deployed in 10 communities across British Columbia and Alberta. No deployment in NWO has been publicly announced. The CWFIS fire maps from Natural Resources Canada are the most accessible monitoring tool available now and are free and publicly accessible for business continuity planning. For municipalities, First Nations administrations, and provincial fire management agencies: the $6-million WildFireSat Remote Sensing Capacity program is worth confirming directly with Natural Resources Canada. For a Northern Ontario technology company developing wildfire detection or prediction tools: NRC IRAP and NOHFC's Invest North — Innovation program both support technology R&D with non-repayable contributions, and wildfire detection is the kind of genuine technological development those programs are built for. Confirm current eligibility and intake status directly with each program.

Frequently asked questions

• Can a Northwestern Ontario business access AI wildfire detection tools now? The most accessible tool is the Canadian Wildland Fire Information System (CWFIS) from Natural Resources Canada — daily national fire maps, danger ratings, and reports from existing satellite data, at no cost. Fast-detection sensor networks like SenseNet are in BC and Alberta communities; no NWO deployment has been publicly announced. Contact vendors directly about availability outside their current footprint.
• Is there funding to develop wildfire detection technology for a Northern Ontario company? NRC IRAP provides non-repayable contributions for SMEs doing genuine technology R&D, including in detection and sensing systems. NOHFC's Invest North — Innovation program supports R&D and commercialization of new technologies in Northern Ontario, up to $2 million at up to 50% of eligible costs. Both require genuine technological uncertainty — confirm eligibility with each program directly before applying.
• When will WildFireSat provide coverage for Northern Ontario? The target launch is 2029, but the program is in active rebuild after CSA terminated its Spire Global Canada contract for convenience in 2026. A confirmed revised timeline should be checked directly with the Canadian Space Agency or Natural Resources Canada before planning any decision around WildFireSat's availability.

Sources: 2026 Canadian wildfire season statistics — NW Ontario 637 fires, 737,892 hectares; Thunder Bay 36 at 318,813 hectares in Wabakimi Provincial Park, one of Ontario's largest on record; national total 3,915 fires, approximately 3 million hectares as of July 21; Namaygoosisagagun First Nation; mid-July Toronto air quality: en.wikipedia.org/wiki/2026_Canadian_wildfires | CBC News — Northern Ontario fire peak at 191: cbc.ca/news/canada/thunder-bay/northwestern-ontario-new-fires-9.7275405 | CBC News — Ontario response, 155+ crews and 80+ aircraft: cbc.ca/news/canada/thunder-bay/forest-fire-update-nwo-9.7273766 | Natural Resources Canada — CWFIS, CFFDRS, annual fire averages (about 8,000 fires, 2.1 million hectares): natural-resources.canada.ca/forests-forestry/wildland-fires | NRCan WildFireSat overview (2029 target, joint NRCan/CSA/ECCC, Budget 2022): natural-resources.canada.ca/forests-forestry/wildland-fires/wildfiresat-canadian-operational-mission | SpaceQ — WildFireSat 2026 update: Spire $72M CAD contract (awarded Feb 2025) terminated for convenience; program reset; $6M Remote Sensing Capacity program; $316.7M federal aerial aircraft investment: spaceq.ca/canada-updates-2026-wildfire-season-wildfiresat-constellation-to-help-when-online/ | TankerVision — U of T Engineering News (Prof. Waslander, UTIAS; NRCan and SFU collaboration; aircraft infrared AI; Ontario expansion plans; 50 fires in 2024; full-system paper fall 2026): news.engineering.utoronto.ca/new-ai-prediction-model-could-transform-how-canada-fights-wildfires/ | MobileSyrup — U of T TankerVision (July 2026): mobilesyrup.com/2026/07/16/university-of-toronto-researchers-new-ai-model-help-canada-fight-wildfires/ | SenseNet — AI cameras and gas sensors, 3-minute detection, Vernon BC pilot, 10 BC/AB communities, Government of Canada validation: foresightcac.com/article/ai-powered-wildfire-detection-how-sensenet-is-protecting-canadas-forests-and-communities | NRC IRAP: nrc.canada.ca/en/support-technology-innovation/about-nrc-industrial-research-assistance-program | NOHFC Invest North Innovation: nohfc.ca/private-programs/invest-north-innovation/