Synthium

Summary

About

About Synthium

Synthium develops a robot-learning data platform that turns human interaction in simulations into labelled multimodal training data. Participants contribute motion, voice, language and decisions through interactive tasks, providing examples for vision-language-action models and humanoid behaviour. The Boston-based company combines simulation, human participation and model training, with backing from Antler and Liminal, a Temasek-backed venture studio.

The simulation platform is live, while work with physical humanoids focuses on research and sim-to-real validation. Synthium has signed a research collaboration agreement with A*STAR and is developing a Singapore testbed described with Unitree and Weston. Its planned Synthium Brain application for real robots sits on a 2026–2027 roadmap. Co-founders Nicolas Duval and Nicolas Savva bring software-platform, computer-graphics and simulation experience from Autodesk, with Duval also having worked on Adobe Express.

PRICE
Range 1Y
Return
High
Low
Updated
DATA: KORTHOS

Operations

Operating milestones

  • Operating milestone

    Synthium reports more than 5,000 interaction sessions recorded on its live simulation-data platform.

    Source
  • Operating milestone

    Synthium reports more than 200 hours of human interaction collected on its live simulation-data platform.

    Source

Team

  • Nicolas Duval - Co-founder and CEO - Leads Synthium’s vision, business strategy and the platform that turns human interaction into robot-training data.
  • Nicolas Savva - Co-founder and CTO - Owns Synthium’s technical roadmap and core platform architecture for simulation-based robot training data.
  • Endri Deliu - Strategic Advisor and Investor - Advises Synthium on AI systems and machine-learning infrastructure.
  • Philip Chan - Strategic Advisor - Advises Synthium on venture building and early-stage growth.

Timeline

Events

Relations

Relationships

  • MassRobotics

    Massrobotics Resident Startup - Synthium is listed in the MassRobotics resident startup directory.

  • A*STAR

    Research Collaboration - Synthium and A*STAR are collaborating on robot-learning research and real-humanoid validation.

Context

Comparable profiles

  • Lightwheel

    stack adjacency - Robotics Data Infrastructure Layer

  • Xspark AI

    stack adjacency - Robotics Data Infrastructure Layer

  • AstrAlive

    stack adjacency - Robotics Data Infrastructure Layer

  • Mecka AI

    stack adjacency - Robotics Data Infrastructure Layer

  • shotwell.ai

    stack adjacency - Robotics Data Infrastructure Layer

  • 6thSense

    stack adjacency - Robotics Data Infrastructure Layer

  • Cortex AI

    stack adjacency - Robotics Data Infrastructure Layer

  • Physical Turing

    stack adjacency - Robotics Data Infrastructure Layer

  • Valgo

    stack adjacency - Robotics Data Infrastructure Layer

  • Sensei

    stack adjacency - Robotics Data Infrastructure Layer

  • Sureform

    stack adjacency - Robotics Data Infrastructure Layer

  • Kognic

    stack adjacency - Robotics Data Infrastructure Layer

Shared workflows

  • LabyrinthAI

    workflow adjacency - Robot Learning Data Collection

  • Mind Robotics

    workflow adjacency - Robot Learning Data Collection

  • SenseGlove

    workflow adjacency - Robot Learning Data Collection

  • Vision Lab

    workflow adjacency - Robot Learning Data Collection

  • MANUS

    workflow adjacency - Robot Learning Data Collection

  • Trossen Robotics

    workflow adjacency - Robot Learning Data Collection

  • Rokoko

    workflow adjacency - Robot Learning Data Collection

  • StretchSense

    workflow adjacency - Robot Learning Data Collection

  • XDOF

    workflow adjacency - Robot Learning Data Collection

  • Maniformer

    workflow adjacency - Robot Learning Data Collection

  • Jiiov / 极豪科技

    workflow adjacency - Robot Learning Data Collection

  • Hub

    workflow adjacency - Robot Learning Data Collection

Context tags

  • Humanoid training simulation platform - Peer Group - Synthium interactive simulation platform generates humanoid training data. https://www.synthium.xyz/about
  • Robot learning data collection - Workflow - Live human-interaction sessions collect motion, voice, language and decision data in simulation for robot policy training. https://www.synthium.xyz/
  • Robotics data infrastructure layer - Stack Layer - Synthium supplies structured multimodal data for robot learning. https://www.synthium.xyz/about