Vound

Active

Data & Deployment Infrastructure for Physical AI.

What it does

Robots learn by watching people work. But the footage they need, real people doing real jobs in real workplaces, barely exists, and it can't be scraped or simulated. Vound supplies it at scale from hard-to-access commercial environments: first-person video of skilled work in factories, farms, hotels, construction sites and dozens of other settings, captured across industries and geographies that others can't reach, and delivered with the task labels and quality control that robotics labs need to train general-purpose models. That access is the foundation of a bigger plan. Every environment we film in is a place robots will eventually work, so the same relationships that power training become the path to deployment: testing, integrating and operating robots in the businesses that already know us. Founded by three Egyptian co-founders from Harvard, MIT and Stanford, Vound launched in the summer of 2026 and reached a $2M revenue run rate within eight weeks.

Next to it in Data for AI · Real-world robotics training data

  • Praxis RoboticsY Combinator S26

    Turns real businesses into robotics data vendors

  • DeepReach Inc.Y Combinator S26

    Network for Physical AI Data Business

  • Cortex AIY Combinator F25

    Large-scale real-world robot & human data for embodied AI

  • Munaria16z speedrun SR007

    Improving data collection for robotics.

  • Liva AIY Combinator S25

    Building socially intelligent AI

  • SureformY Combinator X25

    Real-world data and RL environments for frontier AI

  • HubY Combinator X26

    Real-world training data for frontier AI labs and robotics.

  • Measurement LabsPlug and Play PnP 2025

    Measurement Labs is building the independent intelligence and performance layer for enterprise AI agents, helping companies evaluate, benchmark, and optimize agent performance against real business KPIs across any vendor stack.