Light Origins closes RMB-hundreds-of-millions Pre-A, ex-OpenAI researcher Jiang Xu builds embodied AI foundation model
On August 14, Light Origins announced a Pre-A round of several hundred million RMB, led by Casstar with participation from China Merchants Venture Capital and Xianghe Capital. Founder Jiang Xu previously contributed to OpenAI's GPT-4 development and was named 8 times in the GPT-4 technical report.
On August 14, Light Origins announced the closing of a Pre-A round of several hundred million RMB, led by China Science and Technology Investment (Casstar) with participation from China Merchants Venture Capital and Xianghe Capital. Proceeds will fund ultra-large embodied intelligence model training, multimodal data infrastructure, full-stack software-hardware platform R&D, and high-talent-density team expansion.
Light Origins was founded at the end of 2024. Founder and CEO Jiang Xu previously served as an OpenAI algorithms expert, joining in 2019 and experiencing the full evolution from InstructGPT and ChatGPT to GPT-4. He contributed to large-model training, inference optimization, and RLHF. In the GPT-4 technical report, Jiang was named eight times, tied with co-founder Greg Brockman for the most mentions.
The company proposes and practices a three-stage paradigm: Scalable Pre-Training, Scalable Alignment, and Scalable Deployment, aiming first to build capabilities through massive data training, then elevate via reinforcement learning, and finally push foundation models into the real world to continuously generate high-quality data and form a flywheel. The team has grown to over 100 people with R&D accounting for more than 90 percent, and the hardware team includes alumni of Huawei and DJI.