We invested in @utopiadata.
Every robotics team we meet is compute-rich and data-poor.
Utopia pays people to record the places machines move through then converts each capture into a 3D mesh that AI labs and robotics teams can use, each one can be verified as genuine.
A mapping car photographs a street once and calls it mapped.
6 months later the café is a pharmacy and a bike lane sits where parked cars used to be, the model is still training on last year's world.
Fleets produce a snapshot, people produce a constant feed.
The data bottleneck in physical AI is real, not hype. Robot datasets top out in the low thousands of hours while LLMs train on trillions of tokens. Physical interaction and scene data can't be scraped; it has to be collected.
Utopia's model—bounties for phone footage turned into verified textured meshes—targets a real gap: ground-level places that satellites and mapping cars miss. Robotics labs need scene diversity for navigation and world models. The training-data market is already scaling into the high hundreds of millions and headed for billions. Early, but demand is live. PMF will turn on mesh fidelity, density of coverage, and whether labs become repeat buyers.
Utopia has raised $1M.
Thanks to Gate Labs, Monarch Group, Ascentis Capital and RyzeLabs for backing us.
Everyone agrees physical AI is next but it has almost no real-world data to learn from.
We turn the phones people already carry into an economic network that collects it.
Robots and AI need data that barely exists today.
For users: pick a request, record a location on your phone, get paid in USDC when a buyer accepts your clip.
For AI labs and robotics teams: post a bounty for any place you need captured, accept a clip and get back a 3D model (textured mesh, GLB) with proof of exactly when and where it was filmed. Want it in your pipeline? Pull the verified spatial data over our API too.
Real, verifiable and captured by people, not synthetic.
Our app is coming soon.
The industry is waking up to a problem we have been building for since day one.
One of us will be in San Francisco on August 27.
If our thesis resonates with you, come talk.
in sf next week for @infodrivercap AI data event
Compute got cheap, models got good and now everyone's discovering the same thing: robots/models are starving for data nobody has collected yet
come say hi if you're around✌️
Join the event: https://t.co/fSrPJ38PzU