People are reasonably confused as to why this is being viewed as a potential value add for firefighting. This is generally within my area of expertise as a former wildland guy and dev. I personally think this could be a game changer within reason in 8-12 years. There are two situations where this could be useful, either on an active fire or as part of a prophylactic approach.
On an active fire, traditional approaches on the ground involve figuring out where the fire is going, and squeezing it on the sides until it becomes manageable at the point. Backburns are sometimes used (under very specific conditions) to create areas of black (black meaning burned). Black is safe - removes flashy fuels at the front, and can corral the fire's ability to run. However, you do not want to be downwind of a fire, so backburns are used sparingly, and rarely in front of really dangerous fires. If you lay a backburn right with drip torches, you can actually encourage your fire to run backwards (back in backburn) into the active fire by burning the area closest to you, then burning closer and closer to the active fire, even against the wind. This robot could potentially allow us to set deeper backburns, so more effective counterfires. I am skeptical of my own argument because of price point (not sure how you can thermal shield to avoid sacrificing a bot, though to be fair, backburns don't generally get that hot immediately). That's the active side.
On the passive side, we have incredibly unhealthy forests on the west coast. We have changed our approaches but are toughing through the tail end of the consequences not allowing the natural cycle to cleanse ladder fuels from forests. We also live closer to the urban wildland interface nowadays. On the climate change side, we are getting more lightning strikes, which are the primary cause of western wildfires, and you can make an argument that wind patterns are changing. Bark beetles, surprisingly enough, are probably not causative (I will avoid this rabbithole but it's intuitively quite offputting), though they are spreading northward to novel biomes by the day. To get to the point, we have a ton of biomass that is dangerous to have in forests, and it's several orders of magnitude larger than our current work capacity within the USFS/BLM, even just waving off Alaska as a lost cause. Planned burns are increasingly socially accepted, and robots like these may offer an increased level of fire control, multiply work capacity, and minimize person risk. This is a much stickier argument (a lot of coordination, training, predicting demand increase for planned burns) but I'll wait to see.
I'd say bear case for using these things on fire is survivability, but also fuel load capacity, air possibly being a better resource vector (e.g. napalm bomblets), and departmental adoption, not to even address the standard swathe of hard problems (movement planning, battery capacity, dist syst for hell machines) that we all recognize here. I'm not optimistic, but I'm extremely curious to see how things go.