Hangzhou's Six Little Dragons have made quite a splash, with Qunhe Technology emerging as the first to go public, giving Silicon Valley a run for its money! It's neither DeepSeek nor Unitree but Qunhe, specializing in spatial intelligence, that's stealing the spotlight.

So, what exactly is spatial intelligence?

And how did Qunhe Technology secure the title of both 'First of the Six Little Dragons' and 'First Spatial Intelligence Stock'?

I'll address these questions one by one. This video dives deep, so give it a thumbs up and watch it thoroughly to grasp the latest and most advanced AI technology trends.

First, let's delve into who Qunhe, one of Hangzhou's Six Little Dragons, truly is.

While Qunhe Technology may not be a household name, its flagship product, 'Coolhome', is a giant in the home improvement market and is exploring full-scene applications across offices, retail chains, e-commerce, and beyond!

Qunhe Technology stands as the world's largest spatial design platform, boasting over 45,500 enterprise customers across more than 200 countries and regions, with an average monthly active user count of 86.3 million.

In the first nine months of last year, the company generated revenue of 550 million yuan. Over the past decade, it has concentrated on spatial cognition-related technologies, culminating in the release of a multi-modal CAD large model based on three-dimensional space, further enhancing its spatial intelligence solutions.

Its three founders hail from Zhejiang University and Tsinghua University, with Chairman Huang Xiaohuang previously involved in CUDA development at NVIDIA. This trio has always been at the forefront of using high-performance computing to simulate the physical world.

So, what exactly is spatial intelligence, and why is Qunhe the 'First Spatial Intelligence Stock'?

Spatial intelligence forms a crucial part of embodied intelligence and is also the main focus of NVIDIA and Jen-Hsun Huang for 2025.

Whether it's a quadruped robot, wheeled robot, or tracked robot, their instructions, behaviors, and cognition originate from a single digital chip. The brain resides in the virtual digital world, while the body operates in the real physical world.

Consequently, a large model is essential to map the data of the physical world into the digital world, enabling robots to accurately comprehend the physical world.

In simple terms, spatial intelligence allows robots to emerge from the virtual digital world and seamlessly integrate into the real physical world.

While this concept may seem straightforward, its practical implementation is far from easy.

Even Jen-Hsun Huang has only just outlined a grand vision for 'physical AI', which is still years away from becoming a reality.

Currently, Qunhe Technology may be the only company globally capable of providing large-scale, high-quality synthetic data that mirrors real physical laws.

Over a decade of industrial applications has enabled Qunhe to amass the world's largest interactive 3D dataset of indoor scenes, comprising over 360 million 3D models.

These spatial data can be leveraged to train large models and even robots.

This is precisely what Qunhe's SpatialVerse is doing – enhancing the cognitive abilities of intelligent robots, AR/VR systems, and embodied AI through virtual simulation training.

Previously, robot training primarily relied on real-world scenarios, with robot companies constructing their own training grounds.

Not only was this costly, but it also posed significant safety risks.

A 300-400-pound robot, if it were to malfunction, could turn into a scene straight out of a Hollywood horror movie.

However, with the world's largest 'interactive 3D data capabilities', Qunhe has bridged the gap between the virtual and physical worlds. It not only provides a vast amount of multi-modal motion data for robot training but, more importantly, has shifted robot training entirely into the virtual realm.

In the virtual world, Qunhe can assign real properties, equipping models with authentic physical property information such as density, friction, elasticity, damping, and more.

For instance, a US robot company recently unveiled a robot capable of folding blankets and performing household chores. This required a specific scene dataset to furnish the robot with more detailed and precise 3D data of beds, pillows, blankets, and other fundamental elements.

Training with 100 synthetic datasets yields higher accuracy compared to training in a single physical space.

In my view, virtual robot training is indispensable. At the very least, it eliminates the need for staff to worry about restraining a robot that might malfunction or go rogue at any moment.

The IPO of Hangzhou's first Six Little Dragon, Qunhe Technology, marks a significant milestone, signaling a shift in the offensive and defensive dynamics of the AI world. China's AI has transformed from being an 'imitator' to a 'leader'. Hangzhou, the 'City of Artificial Intelligence', is once again at the forefront of innovation, even carrying the banner of 'one city versus Silicon Valley'.

Behind this success stands the full support of two national-level large scientific facilities, 17 national key laboratories, and 5 provincial laboratories.

Today, they are the Six Little Dragons of AI, and tomorrow, they will be six 'Oriental Dragons'. By pooling their wisdom and strength, they will strive for ultimate victory in the AI world.