That's nothing, and in addition to that, it's a huge waste of seeds (these companies are buying up all the available seed for certain species of trees, which means net fewer trees planted). Manually planting already germinated baby trees is more efficient both in terms of hours and labor spent and has a higher germination rate.

200 trees is about how many an experienced treeplanter will plant before thier first coffee break.

Yes but do they use AI and machine learning? Do they hover in place and look cool? They don’t even have a cool name like “drone.” I guess you probably also can’t mass produce and automate experienced tree planters too. While they’re swilling coffee they could be launching hundreds or even thousands of AI powered drones shooting fancy seed pods everywhere. You can’t stop progress John Henry.

Depending on the seed that might be enough. The same way just adding another machine to a cloud might be more economic than employing 2 engineers for a year in order to make a system more efficient.

A coffee bean weights 132.5 milligrams. 40k of those is ~5.3 kgs of beans. Sure it would be inefficient to employ all of those beans if only 200 of them end up being viable. But if those "wasted" 5 kg beans saves labour, then it might make economical sense. And if the alternative is running some sort of heavier planter machine, it might end up producing less emissions.

For the record I have not done real numbers here and my gut instinct is that the drone thing is a gimmick. I'm just trying to keep an open mind.

I don't think emissions should be a concern at all for a reforestation project, unless orders of magnitude more waste is emitted compared to that absorbed by a tree over, say, 10-20 years. Forests pay long term dividends beyond carbon capture.