Today, @LightwheelAI is announcing a strategic partnership with @ManusMeta to build the data infrastructure backbone for Physical AI. In Physical AI, the quality of a robot's behavior follows directly from the quality of the human demonstration data behind it. Dexterous manipulation — the hardest and scarcest capability in robotics — depends on capturing the human hand with precision. That is where this partnership begins. MANUS captures human dexterity at the source, at 25 degrees of freedom per hand and millimeter precision. Lightwheel turns that high-quality data into robot-learning-ready datasets, then scales it through simulation, evaluation, and real-world deployment — an end-to-end Real2Sim2Real pipeline. As a core capture partner on Lightwheel's Human Data Capture Platform, MANUS brings its high-precision hand capture to an open foundation for standardized human data capture. The next generation of Physical AI will run on high-quality data — and we are building that foundation together. Read the full announcement: lightwheel.ai/media/lightwhe…

Lightwheel builds robotics sim & data infra Raised $145m, based in China 🇨🇳 🔴 Turns physical objects into sim assets 🔴 Collect high-quality egocentric data 🔴 Evaluate data in sim environments They partner with top AI labs (OpenAI, NVIDIA, Alibaba, ByteDance, Manus, Deepmind) and manufacturers (Figure, Toyota, GM, Samsung, Galbot, Agibot) I talked to @jonstephens85 from @LightwheelAI

Just came across @LightwheelAI YCB assets Why did I not hear about this before?

Geely's humanoid robots are now working on a live automotive production line — sorting and sequencing parts. The path from months per task to weeks: Lightwheel's Real2Sim2Real infrastructure. One continuous learning system — human data, simulation, deployment, and back: → EgoSuite captures how Geely's own operators work — first-person human data from the live line. → SimFoundry turns it into physics-grounded simulation, calibrated to Geely's real equipment. → Every failure on the line becomes the next lesson — and each cycle ships a stronger policy. Built on NVIDIA GR00T and NVIDIA Isaac. @NVIDIARobotics This is our continuous learning system for Physical AI — running in production. Full story 👉 : lightwheel.ai/media/geely-hu…

又有新的一鲨

Took a chance to introduce A3 to everyone at the @LightwheelAI Physical AI Night during Automate~ @AGIBOT_US

.@ManusMeta and @LightwheelAI have announced a strategic partnership to accelerate the development and deployment of dexterous robotic systems. The partnership connects two foundational layers of the Physical AI data stack: ✔️ MANUS provides high-precision hand-motion capture as the source signal that feeds the stack. ✔️ Lightwheel's data infrastructure spans real-world human demonstration capture and large-scale simulation. “The most important bottleneck in Physical AI today is the quality and scale of dexterous manipulation data. MANUS captures what the human hand actually does; Lightwheel scales that knowledge into the millions of training examples that robot policies require. Together, we are building the data infrastructure that the next generation of Physical AI will run on,” said Stephan van den Brink, Co-Founder and CEO, MANUS. Read the announcement here: humanoidroboticstechnology.c… #humanoids #humanoidrobotics #humanoid #humanoidrobot

Way to go, guys!

Today, MANUS is announcing a strategic partnership with @LightwheelAI to build the data infrastructure backbone for Physical AI. The partnership connects two foundational layers of the Physical AI data stack. MANUS provides the source signal: human hand capture at 25 degrees of freedom per hand, millimeter precision. Lightwheel turns that capture into robot-learning-ready human demonstration data, through consistent tracking and high-quality annotation at scale. The technologies are already natively interoperable through the @nvidia Isaac ecosystem. MANUS is the official data glove of Isaac Teleop; Lightwheel's SimReady assets sit at the core of Isaac Sim. This partnership formalizes that link into an end-to-end Real2Sim2Real pipeline. High-quality demonstration data is where robot policies begin. This is the infrastructure that makes it scale. Read the full announcement: manus-meta.com/blog/manus-an…