Physical AI needs more than capable models. It needs data from the environments where those models are expected to operate. That’s what makes @vangrid_io interesting to me. Instead of relying only on existing datasets, Vangrid turns smartphones into distributed capture devices for collecting 3D spatial data from real world locations. The useful part is the full pipeline: Real place → smartphone capture → verified spatial data → Physical AI Captures can also carry onchain provenance, giving the data a traceable origin rather than treating every dataset as an anonymous input. As robots move into more real-world environments, having fresh and location specific spatial data could become just as important as improving the models themselves. That’s the infrastructure layer Vangrid is building around.