According to this study a mealworm farm uses more energy per kg of protein produced compared to chicken, but much less energy than any other meat. However, mealworm farms rank lowest in CO₂-equivalent emissions per kg of protein and lowest in land use compared to all meat products, including chicken.
Apparently soy beans produce 6.82 kg of CO₂-equivalent per kg of protein isolate (which is 90% protein, therefore 7,5 kg of CO₂-equivalent per kg of protein), while mealworm farms produce 14 kg of CO₂-equivalent per kg of protein (and around 30 kg for chicken, the next best option). Worse, but less than double.
As for land use, the first study calculates that to produce 1kg of protein from mealworms it is necessary to use 18 square meters of land per year (including the land to grow food for the worms) while according to this other study vegetable proteins need up to 25 square meters of land per year for each kg of protein.
I admit it's not as big a difference in land use as I thought (it's different studies, they might have slightly different metrics) , but I think there are other factors that make it a much more complicated issue: mass use of fertilizers, monocultures, deforestation, soil impoverishment... An advantage of mealworms might be that you can give them a variety of foods that are easier on the soil (the first study mentioned carrots, grains and other stuff) in order for them to produce protein, while protein-heavy plants require rich soil and tend to drain it fast.
I agree, I honestly expected a much starker difference in land use. I also agree that soy beans can be grown responsibly, except of course it's often not the case. The fact that both soybeans and insects are being grown largely as a source of protein for cattle brings us once again back to the main issue: cows!