Lightwheel
Lightwheel builds Physical AI data, simulation and evaluation infrastructure for robot learning. Its platform spans human demonstration data, SimReady simulation assets, synthetic-data scaling and policy evaluation, with reported USD $100M+ Q1 2026 orders and a MANUS data-glove partnership.
Market context
Lightwheel sits in the missing infrastructure layer between model training and real-world robotics deployment. Its value is the continuous loop across simulation, data generation, evaluation, and deployment rather than a single robot or model.
Facts
- Website: https://www.lightwheel.ai/
- HQ: Santa Clara, United States
- Founded: 2023
- Segment: physical-AI data infrastructure
- Known funding rounds: 1
Investors
- Baotong Technology - Participant - Angel / Individual - 1 funding event - first seen 2026-06-23
- CIC Industrial Investment - Participant - Strategic Industrial - 1 funding event - strategic/corporate - first seen 2026-06-23
- Giant Network - Participant - Angel / Individual - 1 funding event - first seen 2026-06-23
- Quantum Intelligence Strategy - Participant - Corporate / Cvc - 1 funding event - strategic/corporate - first seen 2026-06-23
- Shandong Development Venture Capital - Participant - Vc - 1 funding event - first seen 2026-06-23
- Sichuan Development Science and Technology Innovation Fund - Participant - Vc - 1 funding event - first seen 2026-06-23
- Yusys Technologies - Participant - Angel / Individual - 1 funding event - first seen 2026-06-23
- Zhongguancun Science City Fund - Participant - Vc - 1 funding event - first seen 2026-06-23
Commercial
1 commercial event; 1 commercial order; latest event 2026-05-06
- 2026-05-06 - Commercial order - Lightwheel reports ~$100M in Q1 2026 Physical AI infrastructure orders - source: prnewswire.com
Simulation and synthetic-world infrastructure for robotics development
Lightwheel AI develops simulation and real-to-sim infrastructure for robotics and embodied AI. Its platform provides simulation-ready assets, synthetic environments and tooling that robotics teams use to train, test and validate physical AI systems.
- Target environment: Robotics development labs, simulation environments, synthetic-data pipelines, embodied AI teams and digital-twin workflows.
- Deployment model: AI infrastructure model built around simulation assets, real-to-sim workflows, robotics data tooling and developer-facing platform services.
- Customer context: Robot builders, AI labs, simulation teams, robotics developers and companies building or validating embodied AI systems.
- Workflow context: Robot simulation, synthetic data, real-to-sim conversion, training environments, digital twins, embodied AI testing and robotics development tooling.
- Commercial maturity: Lightwheel is early-stage in the robotics simulation layer, with maturity centered on platform development rather than broad disclosed customer deployment.
- Market position: Robotics simulation infrastructure company in the synthetic-environment layer for embodied AI and physical-world model training.
- Adoption constraints: Simulation fidelity, asset quality, toolchain integration, customer adoption, model-transfer performance, compute cost and competition from broader simulation platforms.
- Adjacent context: Simulation, embodied AI, synthetic data, real-to-sim, robot development platforms and digital twins.
- Source confidence: medium
Events
- 2026-07-03 - Partnership - MANUS and Lightwheel partner on Physical AI data infrastructure - MANUS contributes high-precision 25-DoF-per-hand capture for dexterous human demonstrations. Lightwheel turns the source signal into robot-learning-ready data through annotation, simulation and synthetic scaling. The companies frame the workflow around NVIDIA…
- 2026-06-23 - Funding - Lightwheel raises RMB 1B strategic financing for physical-AI infrastructure - Capital is earmarked for robot-learning data, synthetic and simulation data, capability evaluation infrastructure and real-world scenario deployment tooling. Investors include Zhongguancun Science City Fund, Sichuan Development Science and Technology Innovati…
- 2026-03-04 - Product Launch - LightwheelAI releases RoboCasa365 simulation framework - • LightwheelAI released RoboCasa365, a large-scale simulation framework for training and benchmarking generalist robots. • The framework includes 365 everyday tasks across 2,500 diverse kitchen environments, with over 600 hours of human demonstration data and…
- 2025-12-24 - Partnership - Lightwheel AI integrates LeIsaac with Hugging Face's LeRobot EnvHub - • Lightwheel AI integrated its LeIsaac platform into Hugging Face's LeRobot EnvHub, enabling developers to launch high-quality Isaac Sim environments with a single line of code. • This integration allows for seamless robot teleoperation, instant demo capture,…
- 2025-12-24 - Product Launch - Lightwheel AI launches EgoSuite for Embodied AI - • Lightwheel AI introduced EgoSuite, a high-quality, multimodal, globally scalable egocentric human data solution for Embodied AI and world models. • EgoSuite delivers over 20,000 hours per week of continuous egocentric data collection across more than 10,000…
- 2025-12-15 - Product Launch - Lightwheel AI Introduces RoboFinals for Evaluating Robotics Models - • Lightwheel AI launched RoboFinals, a simulation evaluation platform for robotics models. • The platform included 100 industry-aligned tasks and cross-simulator generalization. • RoboFinals was built on NVIDIA Isaac Lab – Arena with the SimReady Asset ecosys…
- 2025-12-15 - Product Launch - Lightwheel AI Launches RoboFinals Simulation Platform - • Lightwheel AI launched RoboFinals, an industrial-grade simulation evaluation platform. • RoboFinals includes 100 industry-aligned tasks and supports cross-simulator generalization across platforms like Isaac Lab and MuJoCo. • The platform is now available f…
- 2025-11-12 - Demonstration - Lightwheel's SimReady asset featured in NVIDIA GTC keynote - • Lightwheel's SimReady asset was showcased during NVIDIA CEO Jensen Huang's keynote at NVIDIA GTC. • The demonstration featured cross-platform cloth simulation using Lightwheel's asset within NVIDIA Isaac Sim and Newton. • The cloth-folding task exhibited co…
- 2025-05-19 - Partnership - NVIDIA powers humanoid robot industry with Isaac GR00T platform adopted by Agility, Boston Dynamics, Fourier and others - NVIDIA released Isaac GR00T N foundation models enabling humanoid robots to perform complex manipulation tasks via natural language commands. Agility Robotics, Boston Dynamics, Fourier, Galbot, Mentee Robotics, NEURA Robotics, General Robotics, Skild AI, and…
Team
- Steve Xie - Founder and Chief Executive Officer
Supply chain relationships
- NVIDIA - Integration - Lightwheel is an early adopter of NVIDIA GR00T N models for humanoid robotics
Relationships
- NVIDIA - Partner - NVIDIA powers humanoid robot industry with Isaac GR00T platform adopted by Agility, Boston Dynamics, Fourier and others
- MANUS Meta - Physical AI Data Stack Partnership - MANUS and Lightwheel are linking high-precision hand capture with Physical AI data infrastructure for Real2Sim2Real robot-learning workflows.
Context tags
- Robotics data infrastructure layer - Stack Layer - The financing is tied to robot-learning data and deployment tooling.
- Physical AI data infrastructure - Stack Layer - Lightwheel builds Physical AI data, simulation and evaluation infrastructure.
- Industrial synthetic data - Capability - Lightwheel scales demonstrations through synthetic data generation.
- Physics AI simulation - Capability - Lightwheel works with SimReady assets and Isaac Sim workflows.
- Robot learning data collection - Workflow - Lightwheel turns human demonstrations into robot-learning-ready data.
- Robot foundation model platform - Stack Layer - The data and evaluation layer supports robot-learning model development.
Customers and deployments
- Named Customer Site - GR00T N1 humanoid system in Geely automotive factory - 2025-05-19 - Geely - automotive production tasks - Lightwheel deployed a GR00T N1 humanoid system using Unitree H1 robots in Geely's live auto…
Lightwheel canonical Korthos profile