Alibaba open-sources RynnBrain to challenge Google, Nvidia robotics tech

Alibaba has on Tuesday launched RynnBrain, a new embodied AI model that helps machines handle real-world tasks.

RynnBrain is open-source, and it’s already up on GitHub and Hugging Face, with the Chinese giant claiming it thing can help robots figure out how to grab stuff, understand where objects are, and plan what to do next.

Basically, Alibaba is turning robots into workers. RynnBrain was built by the company’s internal research group called DAMO Academy. It’s trained to recognize objects, understand how they move, and make decisions based on space and time.

That’s the kind of stuff robots need if they’re going to be useful outside labs. The company says the model performs better than Google’s Gemini Robotics‑ER 1.5 and Nvidia’s Cosmos‑Reason2 in benchmark tests.

Alibaba releases multiple versions and trains AI model on Qwen3-VL

The RynnBrain AI model is trained on Alibaba’s Qwen3‑VL, and comes in different sizes, starting with 2 billion parameters, and there’s also a version built using a mixture-of-experts design for better efficiency.

Developers can try out whichever version fits their project best. Alibaba made sure the model can be used for robotics in real settings like factories and kitchens. It’s built to predict where objects might go, avoid crashes, and plan what action to take.

This kind of AI is exactly what Beijing is focusing on right now. The Chinese government is putting money and attention into physical AI, especially robots that can work in manufacturing and hospitality. The goal is simple: beat the U. S. in the next round of tech battles.

Jeff Zhang, who is the Chief Technology Officer at Alibaba, also runs the DAMO Academy. He’s leading the team behind this release and the labs they’re building next.

“The DAMO Academy will be at the front of developing next-generation technology that supports Alibaba and our partners. We want to discover new tech that helps users and businesses work faster and more securely,” said Jeff.

Alibaba expands global labs and launches hiring for 100 researchers

Alibaba is building seven new labs around the world in Beijing, Hangzhou, San Mateo, Bellevue, Moscow, Tel Aviv, and Singapore.

Their work includes machine learning, fintech, network security, visual computing, quantum computing, and human-machine interaction.

Alibaba also plans to recruit 100 researchers from around the world who specialize in IoT, data intelligence, and natural language processing.

The DAMO Academy is also working with schools. One of their main partners is the University of California, Berkeley, where they’re teaming up with the RISE Lab on secure real-time computing. That means their work won’t stay inside corporate walls.

Jeff said:-

“We’ve spent 18 years building a strong tech base that keeps up with our business growth. Now we’re expanding and want researchers to help us build new tools that improve lives and support small businesses.”

This is all part of Alibaba’s long-term plan. The company says it wants to serve 2 billion customers and create 100 million jobs in the next 20 years.

Right now, it already has a tech team of 25,000 engineers and scientists. But with the release of RynnBrain, they’re making it clear they’re not done building.

Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.