Visualizing Large Scale Gaussian Spalts on Mobile with @sparkjsdev RAD.
This beast was built by the @OVRtheReality — kudos for pushing mobile 3D this hard!
Test it yourself 👉: https://t.co/R5V6Lg2T3K
P.S. Got 360 videos? Turn them into Gaussian Splats FOR FREE at https://t.co/rIolLrh46H (also by the same team)
The first global, city-scale 3D map is taking shape, and it’s machine-readable.
City-scale environments can now be reconstructed in high-fidelity, reaching a level of detail that is becoming almost indistinguishable from reality.
This leap is made possible by large-scale mapping using just an Insta360 X5 and Over the Reality technology.
So far, our dataset includes:
- 220,000+ 3D mapped locations
- 97M images
- 1,000TB of spatial data
Growing by more than 10,000+ newly mapped locations every week.
Enabling Visual AI, Robotics, VPS, XR, and Digital twins.
Big thanks @DAnconia_Crypto for the shoutout 🙏
Calling @OVRtheReality the “sleeper pick of DePIN” hits different.
From an AR/XR NFT project in 2020 → quietly built one of the largest real-world image datasets: 95M+ images, 900TB, 200K locations.
Now those pixels are feeding next-gen world models at the biggest AI labs.
The future ahead is massive.
Real data × DePIN × AI = unstoppable.
Grateful to be seen. The journey is just getting started 🔥
$OVR #DePIN #PhysicalAI
Thanks for the shoutout @leonidas_xyz! At @OVRtheReality, we’re proud to offer a non-extractive approach to data collection, rewarding our mappers with $OVR tokens for their efforts. Join us building the DePIN foundation of Physical AI, one 3D map at a time.
it was about time to finally get attention on this. might be a huge catalyst for DePIN and data collection networks like @OVRtheReality
when i last checked the data quality on both platforms, Over actually had way better imagery than Niantic.
the video below is a random example of two data sets near Times Square and Union Square that i picked for comparison.
explore it yourself, it is wild how much high quality 3d imagery they have already collected
some more background on Over:
Over has built a highly advanced 3D imagery data engine with more than 195,000 mapped locations and 93 million images collected.
by coordinating a global community of mappers to scan high-traffic locations, Over may have built one of the largest 3D mapping datasets in the world.
data is scarce, and real-world data is even harder to gather at scale.
this proprietary dataset can power a wide range of Physical AI use cases, from training vision foundation models to enabling accurate visual positioning systems and supporting AI-driven XR publishing and browsing.
One of the Top 5 most valuable companies in the world has just acquired a license for Over the Reality 3D mapping dataset to train next-generation vision foundation models.
This is a huge milestone for OVER 🧵👇
Mapping with Over the Reality is now powered not only by smartphones, but also by Insta360 X5.
If you own an Insta360 X5 and want to be part of it:
👉 Apply here: https://t.co/kH8r7Y1p3M
With 360 capture, contributors can cover significantly larger areas in a single session, turning daily walks into true city level mapping and building 3D infrastructure for AI, XR, and Robotics.
Rewards are active, and we’re opening the Insta360 X5 Alpha Mapper Program to early contributors.
🌍 City-scale 3D mapping is live with Insta360 X5.
A new mapping technology unlocks full urban city scale 3D maps with Over the Reality, dramatically improving the mapping experience and enabling exponential scalability.
Powered by Insta360 X5, we can now move beyond single locations and map entire cities, with rewards active for contributors.
Own an Insta360 X5? Become an Alpha Tester
📩 email [email protected]
💬 or DM us here on X
2026 Roadmap: Scaling the Spatial Web 🌐
We’re going all-in on building the world’s largest machine-readable 3D map.
What’s coming:
📍 City-Scale Mapping10-20x faster via 360 cams
✨ Free Gaussian SplattingTurn your uploaded videos into 3D assets
👓 Production XRSeamless publishing + Unity SDK
🤖 Robotics StackVPS for autonomous machines
🧠 AI DatasetsTraining the next gen of world models
Execution mode: Activated.
The spatial future starts now. 🚀
We’re excited to share the OVER roadmap for 2026.
The roadmap is focused on one core objective: scaling a machine-readable, world-scale 3D map of the physical world.
Everything we’re building in 2026 ties back to this 🧵👇
🇪🇺 Today it has been all about demos, and yes, our indoor navigation is finally live.
We scanned and mapped the UnitedXR venue with centimetre-level precision.
Key elements:
🔹 Digital Twins as enabling technology for Smart Cities
🔹 Visual Positioning Systems (VPS)
🔹 Accessibility and inclusivity as core design principles
The future of Smart Cities is being built by you, one map at a time.
Tomorrow: important meetings that will shape future European projects 🇪🇺.
We are proud to announce our first contribution to open source Computer Vision research! 🚀
Today, we’re publishing on @huggingface a 1,000 3D Maps Dataset under CC-BY-NC license, designed to accelerate progress in 3D perception and spatial AI. Our way to give back to the global AI community and show how DePIN can make an impact. 🧵
👉 https://t.co/on6Jy3oGyo
Mind-blowing tech alert!
Imagine transforming industrial ops combining 3D Mapping, Machine Perception and IoT-driven machine measurements. No sci-fi.
This is happening now!
After 1X released its Neo for pre-orders two days ago, robotics has gone absolutely viral again. Robot memes everywhere, and more and more people are waking up to how massive this secular growth trend is going to be over the next decade.
A lot of people have reached out asking which crypto x robotics positions I’m holding to bet on this trend and narrative.
Let me break it down one last time for you, here are the positions in my portfolio and why I hold them:
1) $PEAQ / @peaq
Building the global coordination layer for the machine economy and robotics. My largest position and highest conviction bet with the best risk/reward profile.
2) $ROBOT / @RoboStack_io
A cloud platform for simulating and testing robots with its own protocol and tooling stack.
3) $SLC / @silencioNetwork
Crowdsourcing real-world audio data to become the “ears” of AI and robots.
4) $OVR / @OVRtheReality
Crowdsourcing 3D spatial data to train robotic perception systems.
5) $BREW / @homebrewrobots
An “App Store for robotics” offering pre-built motions, behaviors, and gamified control modes.
6) $SHOW / @SHOW_ROBOTICS
A collective building robots that combine hardware, software, and unique AI-driven personalities.
7) $NATIX / @NATIXNetwork
Streetmapping real-world data to train autonomous driving and physical AI systems.
8) $DEUS / @xmaquina
Not liquid yet, but wanted to mention as they are preparing for their next public token sale. It’s essentially an on-chain robotics DAT giving exposure to top humanoid companies like Figure, Apptronik, 1X etc.
Beyond these, a few other strong contenders I’m keeping an eye on:
9) $CODEC / @codecopenflow
Developing vision-language-action (VLA) agents aka AI systems that can see, think, and act, enabling advanced automation across robotics, digital operations, and gaming.
10) $GEOD / @geodnet
A decentralized RTK (Real-Time Kinematic) network delivering centimeter-level GPS accuracy for robots, drones, and autonomous machines, critical for precise navigation.
11) $AUKI / @Auki
Building Posemesh, a decentralized machine perception network that enhances real-time spatial awareness for robots and AI. Effectively aggregating all relevant data from networks like for example Geodnet and Natix as well as generating their own data sets for richer perception.
What else did I miss?
Just saw @alescontrela's GaussGym—mind-blowing photorealistic sims for #robotics!
Imagine scaling it with OVER's massive dataset: 159K+ real-world Gaussian Splats across 100K+ 3D maps (88TB training data) crowdsourced via #map2earn.
Zero-shot transfer to reality?
Game-changer for AI agents. 🚀
Check it: https://t.co/bp0JEvUoNb
Simulation drives robotics progress, but how do we close the reality gap?
Introducing GaussGym: an open-source framework for learning locomotion from pixels with ultra-fast parallelized photorealistic rendering across >4,000 iPhone, GrandTour, ARKit, and Veo scenes!
Thread 🧵
Thousands of people map the world with Over the Reality. New friends are joining us. 🤖
Watch this Unitree Go2 robot dog capture camera feeds and generate a 3D map.
This is just the beginning. An early test.
But soon, fleets of robots will autonomously map the world, constantly updating and expanding the global 3D layer, and share and selling their data in $OVR on @peaq.