getting thousands of people to capture places is only one layer of a spatial data network. @vangrid_io also has to turn those submissions into usable 3d scenes. its workflow includes buyer review, reconstruction and quality checks, with processing capacity affecting how quickly work moves through the pipeline. that creates a less glamorous but important scaling problem: more incoming captures can also mean longer queues. for physical AI customers, useful infrastructure means more than broad coverage. it means getting consistent, usable outputs within a timeframe they can plan around. throughput and quality control deserve as much attention as contributor growth.