What Is Spatial Intelligence and Why Did It Just Go Public?

Hangzhou-based AI startup Manycore went public on the Hong Kong Stock Exchange on April 16, raised roughly $130 million, and watched its shares close 144% above the offer price. The company became the first of Hangzhou's so-called "Six Little Dragons" AI startups to list. The trading day told investors something about the appetite for what the company says it is building.

What Manycore is building, according to chairman Victor Huang, is "spatial intelligence." The phrase sounds like marketing, but there is a real technical concept underneath it. Spatial intelligence means AI that can understand and navigate physical environments rather than just process text or images. "When you enter a room," Huang explained, "you can understand where you are and what's in front of you. And if you want to take a seat, you can understand which seat is empty."

That kind of environmental awareness is what separates a chatbot from a robot that can operate in the real world. It is also what makes it so difficult to train. Most AI training data is text or video. Manycore's argument is that 15 years of running Kujiale (a 3D interior design platform serving China) and its international equivalent Coohom (now available in over 200 countries) has given it something unusual: roughly 500 million 3D spatial assets. Chairman Huang believes these are better training data for physical AI than video, because they capture geometry, scale, and object relationships directly rather than as a camera's interpretation of them.

This is a genuinely hard problem. Robotics companies and autonomous vehicle developers have struggled with spatial understanding for years. Large language models have shown that scale and data can solve problems that looked intractable, but 3D spatial reasoning involves different challenges from predicting the next word in a sentence.

The IPO also reflects something larger happening in Asian markets. AI companies raised $14 billion in Hong Kong in the first quarter of 2026 alone, as Chinese tech firms increasingly prefer Hong Kong over US exchanges. Whether investor enthusiasm matches the difficulty of building AI that actually understands physical space is a question Manycore will spend the next several years answering.