The article Niantic builds AI cards for robots using Pokémon Go data first appeared in the online magazine BASIC thinking. With our newsletter UPDATE you can start the day well informed every morning.
Millions of people chased virtual Pokémon through real streets for years. What hardly anyone knew: They collected 30 billion images with precise geodata for their app developer Niantic. The company has now sold its gaming division and is building AI cards for autonomous robots under the name Niantic Spatial.
From Pokémon Go to AI cards: The path of Niantic Spatial
- On July 6, 2016, players in Germany were able to collect Pokémon themselves for the first time. All that was necessary: a reasonably powerful smartphone with a camera and the “Pokémon Go” app from the game developer Niantic. The Memories from your own childhood attracted millions of people to the streets of the world for years – the reward was downloads and billions in sales.
- It all started in the Birthplaces of Google in California. Niantic Labs was founded there in 2010 by John Hanke before becoming an independent company in 2015 as part of the Alphabet restructuring. After ten extremely successful years, the surprising spin-off follows: Niantic sells its games division Scopely for $3.5 billion and transforms into a company for digital measurement of the world.
- Under a new name, Niantic Spatial is focusing on bringing together augmented reality (AR) and artificial intelligence (AI) in the form of geodata models – the so-called Large Geospatial Models (LGM). Behind it are models that use AI to connect data and images with exact coordinates in the world. She translate the physical world into a machine-readable language.
Have Pokémon Go players become willing data collectors?
The training basis for Niantic Spatial’s LGM system is not freely available recordings. No, they come from the countless Pokémon Go players. Loud own information features Niantic Spatial over 30 billion images with precise geodata from millions of locations around the globe. This data comes directly from AR scans of PokéStops and Gyms.
Using this information, Niantic Spatial creates a intelligent 3D map of the worldwhich also works where GPS tracking loses precision due to reflections and buildings. The real-world data allows systems to predict where an object is with centimeter precision. Due to the AI combination, the system is also able to recognize the context in the images (tree, park bench, curb) and thus avoid dangers.
Anyone who is now thinking “Are you kidding me” is probably not alone. Billions of people grew under the guise of Pikachu, Taubsi and Charmander willing data collectors repurposed for an AI company. Of course, no one officially knew this, which is why one could certainly speak of a gigantic misuse of data. On the other hand: Every user has read the terms and conditions and thus agreed to the data collection and processing. Your own fault – you could also say.
Voices
- On the first day of his fresh from baptism upscale company said Niantic Spatial boss John Hanke: “While Large Language Models (LLMs) are revolutionary in their own right, the majority of human life and economic activity still takes place in the physical world. As we built Niantic, we were fortunate to build a world-class research and development team, design unique spatial models, and bring early versions of our technology to millions of people. Now we are ready for the next phase of the AI revolution: the leap from the screen into the real world.”
- Coco Robotics’ approximately 1,000 last-mile delivery robots deliver goods in five cities. The company of co-founder and CEO Zack Rash has in March 2026 announced a collaboration with Niantic Spatial: “The development partnership gives us reliable access to localization services that further improve the navigation of our robots. Together we will explore new ways to enable Coco robots to achieve even higher levels of safety and autonomy in any city.”
- Drew Smith compares in its independent portal The Liberty Line Niantic’s work with solving ReCaptchas, whose data was also used for AI training, and gives vent to his displeasure: “We are the unpaid research assistants. The AI gets smarter with every use; the company records this interaction, improves the product and then sells access to it back to us at a high price via a subscription model. We create the content. You own the content. You train the models with this content. You sell the models back to us. We pay.”
Physical AI: Why data will become even more valuable in the future
What can we learn from this story, which is on the one hand so strange and on the other hand not at all strange? For us users, we have to finally stop naive and gullible like a newborn sheep stumbling through the world. There is no fun gimmick in a digital context that is truly about having fun. It’s always about data and your personal information – sometimes disguised as Pokémon and sometimes in the form of talking fruits.
If we put on the professional glasses, we have to note that Niantic Spatial has put itself in an outstanding position with its detailed maps of pedestrian zones and city centers around the world. The magic word is: Physical AI.
As more and more autonomous systems – from humanoid robots to self-driving cars – walk or drive through our streets, a Google Maps for AI increasingly important. And that’s exactly what Niantic Spatial already is.
A little consolation for us: Maybe in the future AI robots will recognize us when we deliver our Amazon packages and thank us for covering the kilometers in Pokémon Go so diligently.
Also interesting:
- Europe in the AI dilemma: What happens when the bubble bursts?
- AI waste without a final storage facility: 19 percent of the internet is absolute junk
- Ready for the stock market at any price: OpenAI is making a radical change in strategy
- New fuel price rules decided: This is now changing at the gas station
The post Niantic uses Pokémon Go data to build AI cards for robots appeared first on BASIC thinking. Follow us too Google News and Flipboard or subscribe to our newsletter UPDATE.
As a Tech Industry expert, I find Niantic’s use of Pokémon Go data to build AI cards for robots to be a fascinating and innovative application of augmented reality technology. By leveraging the vast amount of data collected through the popular game, Niantic is able to create AI cards that can help robots navigate and interact with the physical world more effectively.
This use of data demonstrates the potential for cross-industry collaboration and the power of leveraging existing resources to drive technological advancements. By incorporating elements of gaming technology into robotics, Niantic is pushing the boundaries of what is possible in terms of AI development and robotics integration.
Overall, I believe that this initiative has the potential to revolutionize the way robots are able to interact with their environment and perform tasks. It highlights the importance of leveraging data and technology from diverse sources to drive innovation and create new opportunities for advancement in the tech industry.
Credits