A decade after the peak of Pokémon Go’s global popularity, an AI firm has leveraged billions of images captured by millions of players to develop navigation technologies for delivery robots and potentially military drones. The company, Niantic Spatial, spun out from the original Pokémon Go developers in May 2025, using data from users who took short videos of physical neighborhoods and landmarks.

The scans were part of a broader effort to train AI systems that recognize and interpret real-world spaces. Despite being optional, they provide valuable metadata—location, orientation, lighting—that could be crucial for robust geospatial models.

With over 30 billion images, mostly clustered around urban environments, the data set is vast. It captures locations from multiple angles under varying conditions, offering a detailed view of the physical world. This technology could revolutionize how we navigate and interact with our surroundings, but it also prompts questions about mass surveillance and its ethical implications.

Transparency has been key. Niantic Spatial’s privacy policy and public announcements have always acknowledged that these scans would improve their technology platform since 2019. However, the sheer scale of data collection raises concerns about consent and potential misuse.