The human layer of Embodied AI

Scale AI for humanoids

Human-robot interaction data for humanoid robots, generated in simulation.

Users play. Robots learn. We sell the data. Players do it for free, and every session becomes training data for the teams building humanoids.

Synthium

Backed by

The problem

Humanoid progress is gated by interactive human-robot data, not model size or compute.

There aren't enough physical humanoids in the world to collect the data needed to train them. So Vision-Language-Action (VLA) models starve for the diverse, interactive, multimodal data they depend on, because it barely exists yet.

How it works

We generate the decision-rich data teleop and video can't.

  1. 1.Play

    People play and interact with robots in simple simulations.

  2. 2.Capture

    We record it all: motion, voice, language, decisions.

  3. 3.Structure

    Each session becomes clean, labeled training data.

  4. 4.Train

    Robot models learn from it, and we measure the gains.

  5. 5.Deploy

    The skills transfer to real and simulated humanoids.

The moat

15× cheaper data that compounds while hardware burns cash.

Cost per usable data-hour

$135

Teleoperation

$9

Synthium

≈ 15× lower

Players generate the data for free because it's genuinely fun, like a Roblox for robotics.

No teleop wages, near-zero marginal cost, and every new player makes the dataset stronger. The gap compounds into a data moat.

Proven on real hardware

Signed agreements, live humanoids, real deployments.

Unitree

Collaboration on real humanoid hardware.

A*STAR

Research Collaboration Agreement (RCA).

Google Singapore

Engineering support via Google for Startups.

MassRobotics

On the ground in Boston.

Our edge

We capture the signal everyone else skips.

Active, not passive

The field scrapes video and teleop snapshots. We run a live loop of humans and robots interacting in real time.

Human + synthetic

Pure synthetic data is hollow. We're native to simulation and add what it can't fabricate: real human behavior and decisions.

Interaction, not dexterity

Billions chase robot hands. We own how humanoids behave safely around people, where they actually have to work.

The human know-how inside every robot.

We start as the dataset others can't make.
Play a session, or partner with us on data.

Get Access