In an inspiring move that's set to accelerate robotics innovation globally, AGIBOT Innovation Technology Co. unveiled its groundbreaking Genie Sim 3.0 platform at CES 2026, offering the robotics community a powerful new tool that's completely open-source and accessible to all.

The Shanghai-based company, which has achieved remarkable growth since its founding in February 2023, is sharing the fruits of its research with developers and researchers everywhere. This generous approach represents a refreshing commitment to collaborative innovation in the rapidly evolving field of embodied intelligence.

At the heart of Genie Sim 3.0 lies an elegantly unified workflow powered by NVIDIA Isaac Sim technology. The platform brilliantly combines digital asset generation, scene creation, data collection, automated evaluation, and physics-based simulation into a single, user-friendly toolchain. What makes this particularly exciting is how it removes traditional barriers that have slowed robotics development, making sophisticated simulation accessible to teams of all sizes.

The platform's crown jewel is its intelligent natural-language scene generation capability. Imagine describing a robot environment in plain conversation and watching as the system automatically creates detailed, accurate simulations complete with thousands of semantic variations. This intuitive approach means developers can focus on innovation rather than getting bogged down in complex coding, truly democratizing advanced robotics development.

AGIBOT's commitment to the community shines through in its release of over 10,000 hours of open-source synthetic data spanning more than 200 different tasks. This treasure trove includes multi-sensor data from RGB-D cameras, stereo vision systems, and whole-body kinematics, providing researchers with an invaluable foundation for their work. The intelligent data-collection toolkit even features clever recovery mechanisms that resume work after task failures, saving countless hours of frustration.

The platform's comprehensive evaluation system is equally impressive, offering more than 100,000 simulation scenarios through the Genie Sim Benchmark module. Rather than relying on narrow metrics, it builds complete capability profiles for embodied models, helping developers understand not just what works, but why and how to make it better.

AGIBOT's real-world success adds credibility to their simulation platform. The company recently celebrated rolling out its 5,000th mass-produced humanoid robot and launched the updated AGIBOT A2 humanoid model. This practical experience in manufacturing and deploying robots ensures that Genie Sim 3.0 addresses real-world needs rather than theoretical scenarios.

By partnering with Skyland Innovation's MetaCam technology, the platform can capture real environments with stunning accuracy using handheld 3D laser scanning combined with high-resolution imaging and centimeter-level positioning. Even more remarkably, interactable objects can be transformed into simulation-ready assets from just 60 seconds of video, dramatically speeding up the development process.

The decision to make all simulation assets, datasets, and evaluation source code freely available on GitHub represents a powerful vote of confidence in open collaboration. By reducing dependence on expensive physical hardware and lengthy validation cycles, AGIBOT is helping to level the playing field for robotics innovation worldwide.

This bold initiative promises to spark creativity and accelerate breakthroughs across the global robotics community, bringing us closer to a future where intelligent robots enhance human capabilities and improve lives everywhere.