DIMENSIONAL × ADVENTUREX · 2026
Let your agents touch the world
Over the past three years you have built chatbots, AI SaaS, and agentic consumer apps. Your agents have called thousands of APIs and orchestrated complex workflows. But they have never driven an actuator, pushed open a door, or seen the real world with their own eyes. This year, let's do something radical: give your agents a body.
THE TRACK
A million agents in the cloud. Almost none in the real world.
On Dimensional (DimOS), programming a robot feels exactly like programming any other agent, except this one can see you and walk over. If you can write a basic Python function or prompt Claude Code, you can already teach robots new skills. 15 minutes from pip install to a robot dog that finds you in a room, walks over and says hi. What you build after that is up to you.
AT YOUR STATION
Program atoms.
██████╗ ██╗
██╔═████╗███║
██║██╔██║╚██║
████╔╝██║ ██║
╚██████╔╝ ██║
╚═════╝ ╚═╝
Unitree Go2 ↗
A robot dog and a charging station per team. 12 joint motors (up to 45 N·m of torque), a 360°×90° 4D LiDAR, and a 120° wide-angle front camera: everything your agent needs to see the room and move through it.
25 STATIONS
██████╗ ██████╗
██╔═████╗╚════██╗
██║██╔██║ █████╔╝
████╔╝██║██╔═══╝
╚██████╔╝███████╗
╚═════╝ ╚══════╝
Galaxea A1Z ↗
A 6-axis force-controlled arm at your station, 700 mm of reach and up to 5 kg of payload. Your agent can pick things up, hand things over, sort, assemble, and work with objects on a table.
10 ARMS
██████╗ ██████╗
██╔═████╗╚════██╗
██║██╔██║ █████╔╝
████╔╝██║ ╚═══██╗
╚██████╔╝██████╔╝
╚═════╝ ╚═════╝
DimOS
The agent-native OS for the physical world. Navigation, perception, memory, voice, and LLM integration work out of the box.
OPEN SOURCE
██████╗ ██╗ ██╗ ██╔═████╗██║ ██║ ██║██╔██║███████║ ████╔╝██║╚════██║ ╚██████╔╝ ██║ ╚═════╝ ╚═╝
On-site engineers
The Dimensional engineering team is at the venue throughout, with a support group and daily office hours. Day 1 workshop: a live walkthrough from zero to a robot saying hi.
ALL WEEK
THE BRIEF
Give your robot a job.
We are not going to tell you what to build. But we will tell you what gets us excited: a robot that does something useful enough that someone would pay for it. Bonus points if you can tell us who that someone is. Best if it solves a problem you experience first-hand, something a venue, a warehouse, a school, or a store might actually want running. If you need a starting point, look around the venue: what job here could a robot do better?
Autonomy
The agent perceives, decides, and acts. If you are driving it with a joystick or running a hardcoded script, that's not an agent.
Cross-track projects welcome
Building something for another track? Give it a body here. Your productivity tool, your game, your social app: any of them can grow a physical presence.
Part of a bigger system
Your robot doesn't have to work alone. A cloud agent plans, another pulls the data, and the robot dog walks over and does the part that needs a body. Think of it as giving your existing agent stack a pair of legs and hands.
DIMENSIONALOS · IN PRODUCTION
What Dimensional provides out of the box
Production use cases
JUDGING
Winning criteria.
Autonomy
The more your agent perceives, decides, and acts on its own, the better.
A working prototype
Your deliverable needs to show that it works on presentation day.
Creative use
We have been thinking about this space for a long time. Show us something we hadn't considered.
Business use case
Optional. A use case you could deploy in production today, for a business that would pay for it.
PRIZES
¥30,000 across four awards.
| AWARD | PRIZE | NOTES |
|---|---|---|
| Best Physical Agent | ¥15,000 | Grand prize |
| Best Worker | ¥5,000 | Does real work in a real setting |
| Best Performer | ¥5,000 | Live audience vote in the arena |
| Best Hands | ¥5,000 | Best use of the Galaxea A1Z arm |
Ready to give your agent a body?
Join the track group on Feishu for setup help, office hours, and announcements. Docs contain everything you need to run DimOS on your chosen platform.
HARDWARE PARTNERS
