Light Origins Raises Hundreds of Millions in Pre-A Round

Summarized by NextFin AI

Embodied-intelligence startup Light Origins closed a Pre-A financing round worth several hundred million yuan, led by Guoke Investment.
The funding will support large-scale model training, multimodal data infrastructure, full-stack software-hardware development, and technical team expansion.
Founded in late 2024, the company employs more than 100 people, with over 90% dedicated to research and development.
Led by former OpenAI algorithm expert Jiang Xu, Light Origins aims to extend large-model capabilities into physical environments through cross-embodiment applications.

NextFin News — Embodied-intelligence startup Light Origins has closed a Pre-A financing of several hundred million yuan, the company announced on Thursday.

Guoke Investment led the round, with participation from China Merchants Venture Capital and Xianghe Capital. Proceeds will fund large-scale training of embodied-intelligence models, construction of multimodal data infrastructure, development of a full-stack software-hardware platform and expansion of a high-density technical team.

Founded at the end of 2024, Light Origins is led by Chief Executive Jiang Xu, a former OpenAI algorithm expert who contributed to the full-stack development of InstructGPT, ChatGPT and GPT-4. The company, which now employs more than 100 people with over 90 percent in research and development, aims to extend large-model capabilities from digital environments into the physical world and plans to release scaled pre-training results and cross-embodiment demonstrations.

Explore more exclusive insights at nextfin.ai.

Insights

What does embodied intelligence mean, and how does it differ from AI systems that work only in digital environments?

What technical foundations are needed to move large-model capabilities from software into the physical world?

How do multimodal data infrastructure and large-scale training support embodied-intelligence models?

Why are investors putting hundreds of millions of yuan into embodied-intelligence startups at this stage?

What does Light Origins' rapid hiring of a research-heavy team suggest about the current talent race in embodied AI?

How important is Jiang Xu's OpenAI background to Light Origins' credibility and strategy?

What are the main industry trends driving interest in full-stack software-hardware platforms for embodied intelligence?

What recent breakthroughs or policy changes could affect embodied-intelligence development in China?

What should readers watch for in Light Origins' planned pre-training results and cross-embodiment demonstrations?

How might embodied-intelligence models evolve over the next few years as computing power and robotics improve?

What long-term impact could embodied intelligence have on manufacturing, logistics, healthcare and home robotics?

What are the biggest technical bottlenecks in training AI systems to operate reliably in physical environments?

Why is building a full-stack embodied-intelligence company more difficult than building a software-only AI startup?

What safety, cost or deployment concerns could slow adoption of embodied-intelligence products?

How does Light Origins compare with other Chinese and global embodied-AI startups in funding, team background and technical goals?

What lessons can Light Origins learn from earlier AI or robotics companies that scaled quickly after major funding rounds?

NextFin.Al

No Noise, only Signal.

Open App