# You Didn't Play Pokemon GO. It Played YOU.
# Uploader: The Infographics Show
# Upload date: 2026-06-12
# URL: https://www.youtube.com/watch?v=36I3UnYME-A
# Narrated by Josh Risser.
# Timestamped transcript fetched 2026-07-28; seconds from video start.
# Complete: the video carries no sponsor read; chapter titles are spoken
# by the narrator and appear inline below.

[0.16] Pokemon Go was never really about catching Pokemon. It was about something much darker... Mapping the real world.
[5.76] Every park you explored, every street you walked with your camera open, it all trained a massive AI system building
[8.88] a high-resolution 3D model of Earth. And millions of players paid to do it.
[19.52] While you were hunting shiny Charizards, the game's creator, Niantic, was making a digital twin of the physical world.
[24.24] It might be the biggest reverse-labor heist ever. Convincing millions of people to build a global surveillance grid
[29.20] they can never opt out of.
[34.56] Chapter 1 - The Ghost in the Machine. In 2024, Niantic dropped a huge announcement. They were building a "Large
[40.16] Geospatial Model", an AI system designed to understand the physical world in 3D. It wouldn't just recognize locations
[45.84] but analyze and connect them. It would then map millions of real-world places into one massive machine-readable
[52.48] network. This happened right as the AI arms race exploded after OpenAI became one of the most powerful companies in
[57.12] tech. Suddenly, everyone wanted in, even video game companies.
[62.80] But it wasn't the "what" that raised eyebrows. It was the "how".
[70.24] The company revealed that it had trained more than 50 million neural networks, based on 30 billion images taken of the
[74.08] world surrounding them. Those pictures had all been sourced from player-contributed data straight from Pokemon Go.
[79.84] It didn't matter that the players had been inactive for years, their accounts were still sitting around. They were
[88.88] still on the platform and every location scan they had ever submitted was being added to the network. Users probably
[93.28] considered deleting their accounts, but the idea of abandoning that one Eevee no doubt broke their heart. But at
[98.32] least their data would be safe...right? Not quite.
[105.92] The data has all been folded already into the Large Geospatial Model, helping it to create a map of the world around
[110.96] it. And just like other LLMs, it's hard to delete any data from a LGM. Think of it as putting together a jigsaw
[117.44] puzzle. Once you remove one piece, it's not complete anymore. The model can't simply unlearn a data point provided
[122.40] by a specific player. That would be more like trying to completely retrain the model, something that would require a
[128.16] heavy investment in cash. It's estimated it could cost between $6 and 10 million to retrain the model, something that
[138.56] might not even be possible because of how this data mining exists at the core of the operation.
[144.24] Because the model isn't just learning, it's reaching places it could never get otherwise.
[149.84] Chapter 2 - The $15 Billion Land Grab. Since 2007, Google has used Street View cars to photograph roads around the
[159.12] world, creating a map made from real images of real places. But Street View has a major limitation. The cars can only
[164.96] access public roads. That means parks, walking trails, malls, campuses, interiors, and countless private or
[170.80] hard-to-reach spaces remain largely unmapped. They're known as "The Last Meter."
[176.56] But Niantic changed all that. And did it using your foot traffic.
[181.28] It took millions of Pokemon Go players and turned them from a fanbase into a global data-collection network.
[187.20] Suddenly, players were out there mapping every inch of their towns, one walk at a time.
[192.56] And it succeeded beyond the company's wildest dreams.
[196.00] The estimated cost of hiring professional surveyors to map these spaces and build a usable geospatial knowledge base
[202.88] was over $15 billion. But Niantic didn't just avoid paying that, it made a profit.
[208.72] The game has resulted in a net profit of over $6 billion as of 2026. But that's just the cherry on top of the cake,
[215.12] because every one of those players was working miracles for the company. All the data they were collecting, all the
[220.48] locations they were mapping, was being funneled directly back to the company. It would help them develop one of the
[225.52] most ambitious AI projects of all time.
[231.04] Now, Niantic is putting all that data to work for their bottom line.
[234.96] Because it turns out they were never a video game company at all.
[239.36] Chapter 3 - The Intelligence Pedigree. Niantic specialized in Augmented Reality, a type of gaming that involved much
[245.20] more direct interaction with the world. At the time, most immersive experiences were confined to VR arcades: small
[249.92] physical spaces where players wore headsets and explored digital fantasy worlds layered over a controlled environment.
[261.20] But Niantic wanted something much bigger. It wanted to augment YOUR world.
[266.64] The company's first augmented reality game was Ingress and proved that Niantic had what it took to compete in this new
[270.88] field. That proof-of-concept helped them get the Pokemon license from Nintendo. But behind the scenes, there was a
[277.92] completely different goal and it went all the way back to the start.
[282.16] The founder of Niantic, John Hanke, actually began in a completely different field. He was known for founding the
[287.04] company Keyhole, an early digital mapping company founded in 2001. Hanke secured funding for his company through
[291.20] In-Q-Tel, the CIA's venture capital arm, and the technology quickly attracted attention in Silicon Valley. Then
[297.52] Google stepped in, acquiring Keyhole in 2004. The company would eventually become Google Earth, using Keyhole's
[307.12] satellite-imaging tech to create a 3D representation of the planet. That became the foundation for the far more
[312.48] precise mapping systems Google developed for Street View.
[316.72] Eventually, Hanke would combine his two passions.
[319.28] Niantic was formed as an internal startup inside Google in 2010, before becoming an independent company. It got heavy
[323.60] investment from Google, Nintendo, and The Pokemon Company for its biggest project yet. And using Hanke's deep
[333.12] knowledge of geomapping, it would create a new type of gaming experience.
[337.68] Anyone who has ever played an online free-to-play game knows there are always some strings attached. You download the
[341.68] game, and play for free, but then the game starts hitting you up for money. These "Freemium" games exist off
[347.12] microtransactions, and those few dollars here and there add up. But Pokemon Go had its own economy hidden in the game.
[358.32] Users were playing for Pokecoins, an in-game currency that could be used to buy useful items that made it easier to
[363.68] level up, find Pokemon, or just make your collection more attractive.
[368.72] Sure, you could buy them using real money, but there was also a way to earn them through AR mapping tasks.
[375.20] These were often time-consuming, walking-heavy tasks that served a behind-the-scenes purpose. It helped Niantic map
[380.24] the area where Pokemon Go was taking place.
[384.16] And users were all too happy to do that... completely unaware they were being exploited.
[389.84] Chapter 4 - The Poffin Paradox: Reverse Labor. Poffins were one of the most useful items in the game, essentially a
[393.52] super-charged pet treat and they cost 100 Pokecoins. When you factor in the value of Pokecoins and how long it took
[398.80] to earn a Poffin doing these tasks, players weren't actually engaged in gameplay. They were essentially working for
[404.16] the wage of roughly $1 an hour, and they were doing it happily. In the early days, users were doing these tasks
[409.60] regularly, desperate to get their hands on these items by any means necessary.
[420.40] But then, people started talking.
[422.32] The initial novelty of finding Pokemon wore off, and a drop off happened. But if you look at any Pokemon forum, these
[428.00] tasks will be seen as one of the biggest sources of later player frustration. These AR Mapping tasks usually involved
[433.28] physically walking around and recording continuous 3D video scans of existing PokeStops or Gyms using a phone camera.
[443.92] While they weren't essential for playing the game, they were usually the fastest way to get important items if you
[447.92] didn't want to spend real money. So people stopped playing.
[453.28] But by that point, Niantic had gotten what they wanted already.
[456.40] Users may have felt like they were playing a cute game, but what they were essentially doing was data-mining for a
[459.12] powerful tech company.
[463.28] All while being paid pennies on the dollar for the privilege of spending those same pennies in the company store.
[469.76] It calls back to the old "Company Scrip" where laborers in isolated locations weren't paid real money, but tokens
[475.20] that could only be spent at the company's own stores. The result was a closed loop. Labor was exchanged for currency
[480.16] that flowed right back into the company, often at inflated prices. The workers provided essential value, while
[485.60] having no real economic exit.
[491.36] And the data those Pokemon hunters mined turned into an incredibly precise - and potentially terrifying - technology.
[498.40] Chapter 5 - The Patent On Reality. Niantic, as part of its relationship with Google, has been aggressively staking
[503.76] out a space in the AI sector. And one patent might be one of the most significant to make its debut in the field in
[508.00] years. And it might affect you a lot more than you think.
[515.04] Because for the first time, a company is trying to patent reality itself. Can this company really stake a claim to
[520.24] the world we live in, simply because no one else got there first?
[526.56] No...but also yes.
[527.92] The official title of the patent is "Generating 3D Models of Real-World Objects" describes a system for building
[533.52] detailed spatial models from user-generated images. Niantic used data collected through Pokemon Go - photos, movement
[539.20] patterns, and location signals - to help construct what it calls a Visual Positioning System. Instead of just
[544.56] building maps, the goal is to create environments in 3D using crowd-sourced imagery.
[549.60] Over time, that data feeds into a larger system: a digital reconstruction of the real world that exists inside
[555.44] Niantic's infrastructure. It captures both public spaces and privately accessed environments at a level of detail
[559.60] traditional mapping never reached.
[564.56] Now, it all belongs entirely to one company.
[570.00] What Niantic has done is essentially create a digital twin of our world, and that allows them to locate any spot on
[574.08] Earth within seconds. As long as someone is actively using the augmented reality camera, Niantic has access to their
[579.12] location and can find the person in real space, using the digital twin of our world. While standard GPS has been able
[588.72] to track people for decades, it only puts you inside a general radius. This 3D system tracks you down to the exact
[594.40] centimeter. If you're playing Pokemon Go - or any Niantic product - the odds are the company knows exactly where you
[598.56] are right now. Down to the step.
[605.44] Chapter 6 - The Landlord Of Reality. From the start, Pokemon Go was serving two masters. Yes, it was a legitimate
[611.52] Pokemon game and the designers put the effort into appealing to fans. Freemium mobile games often have a reputation
[615.92] for poor quality, either slapping popular IPs onto generic match-three games, or not researching the property
[620.96] effectively. But Pokemon Go showed a strong grasp of the extensive lore of the franchise.
[631.20] The evidence of Niantic's activity since the launch of the company indicates that the mapping was always the root
[635.20] purpose of the game. While fans were frantically searching for a shiny Pikachu, Niantic was collecting every bit of
[639.52] data and using it to create their proof-of-concept for the digital duplicate of the world we live in.
[648.64] Because Niantic doesn't just want to perfect this digital reality, it wants to own it.
[653.92] Niantic has been developing new and powerful tools for Augmented Reality since it launched, with its Lightship
[658.88] platform being used to manage its top games. It was eventually decommissioned and turned into the Niantic Spatial
[664.08] Platform, an interactive tool that allows designers to explore their own virtual reality and craft their own
[668.88] interactive experiences. Niantic can't completely patent the idea of augmented reality, but what it can do is create
[674.00] the best model, the one based on the largest data stores possible, and make everyone else look second-rate.
[685.44] And that means every new experience will have to go through them.
[689.04] Lightship and its spin-off platform is likely to become the definitive platform for designing Augmented Reality apps
[692.64] as long as Niantic keeps up its speed of development. That means the company will likely find itself in a similar
[697.44] situation to Open AI. Unless you have massive resources and a head start, competing with an established model will
[701.76] be extremely difficult. The best way for companies looking to get into the AR game to proceed will be to license
[711.20] Niantic's model and base their app on the pre-existing framework.
[716.88] So Niantic's future could be looking very bright.
[719.84] Pokemon Go showed a deep audience for augmented reality, especially if it's based on beloved properties. The odds are
[724.32] high that other top companies like Disney and Sony will want to parlay this technology to take fans inside the world
[728.08] of their favorite movies and games. And every single one of those apps will be paying rent to Niantic. The company is
[736.96] poised to become not just a dominant force in Augmented Reality, but a monopoly when it comes to the technology the
[741.76] experience is based on.
[746.08] And they owe it all to you, the player.
[748.96] Niantic solved one of the most challenging problems in the digital world - the idea of how to map the last mile. For
[753.76] years, huge swaths of the world were off-limits to digital mapping services, forcing many people to play guessing
[759.28] games when it came to precise locations. Niantic figured out that they didn't have to map the world in order to get a
[764.24] full picture of it. All they had to do was motivate the average person to get in there and do it for them. Niantic
[769.60] got all the data they wanted, and billions of dollars in profit. And the players got to take a selfie with a Squirtle.
[782.08] But now, they might no longer be needed.
[784.40] Chapter 7 - Coda: The Final Evolution. While Pokemon Go is still active, the game currently has around 34.2 million
[793.04] active users, just over 15% of its peak in 2016. But the game is long past its heyday, and Niantic has been
[798.56] increasingly focused on leveraging the data it's collected in order to facilitate new contracts. Many of those deals
[803.36] will be for other games and apps, using the digital world users built for them. But that's just the start of the
[807.92] potential. Pokemon Go has helped to create an incredibly detailed virtual world with all of us in it.
[817.92] And that information is incredibly valuable.
[820.40] It's said that if Google Street View can tell which street you're in, Niantic's Visual Positioning System knows which
[823.92] brick you're looking at. That information could be useful to digital marketers, who would be able to microtarget ads
[829.28] to people as their phones ping right as they walk by a store. But in a more serious usage, this software could be
[838.00] turned over to the police if they need to locate a fugitive before they get out of town. The model is almost
[841.92] complete, but the implications of it are endless.
[847.12] And the user might just become irrelevant.
[849.84] From the day Pokemon Go launched, the fans were out in the streets playing and helping to create a community of
[853.44] trainers. To them, the game was real, and they loved their virtual pets just as much as Ash Ketchum loved his Pokemon.
[857.44] But to Niantic, the fans were always a means to an end. Everyone who played Pokemon Go was part of a massive
[862.56] data-collection operation that leveraged the love of the game to create an army of motivated data-collectors who were
[867.20] helping to build the Digital Panopticon without ever knowing it.
[877.52] But nothing lasts forever - or does it?
[880.16] The gaming industry is littered with free-to-play games that were no longer profitable, and were abruptly taken off
[884.00] the app store. Pokemon Go is profitable now, but that might change and once Niantic has all the data it needs, it has
[888.80] no reason to keep it up. That means that players could one day wake up and find all their hard work has been erased
[893.44] and their beloved pets are now digital vapor, leaving no trace. Well, except for the massive surveillance
[903.52] infrastructure that every Pokemon Go player crafted over a decade.
[908.40] But is everyone chasing a mirage of the next big technology? Find out the shocking truth in "The AI Gold Rush is Dead
[913.12] - Corporate AI Is A DELUSION" or watch this video instead.