Alibaba Plans 10 Trillion Parameter AI Models with New Chip

Alibaba Deepens Artificial Intelligence Push with Models, Chips and Data Centers

Quick Summary

Alibaba is accelerating its artificial intelligence expansion across models, chips, cloud services, and infrastructure. The company plans future Qwen models containing between 5 trillion and 10 trillion parameters, significantly exceeding its current flagship model. Alibaba also introduced the Zhenwu V900 AI chip, which reportedly delivers three times its predecessor’s performance. The new processor can operate within clusters containing up to 500,000 chips for large AI workloads. 

Additionally, Alibaba plans to increase its global data center capacity beyond 20 gigawatts by 2032. Strong customer demand continues supporting growth across its cloud business, although supply constraints could slow infrastructure expansion. Following these announcements, Alibaba’s Hong Kong-listed shares climbed 5.1% to their highest level in about one month.

Introduction

Alibaba is accelerating its AI push across models, chips, cloud services, and data centers. The company announced several major developments during its annual cloud conference in Hangzhou. These plans come as Chinese technology companies develop domestic alternatives to advanced artificial intelligence processors. The broader shift follows tighter United States restrictions on advanced AI chip exports. 

Alibaba is now expanding its capabilities across several important layers of the artificial intelligence ecosystem. Its strategy combines larger foundation models, specialized computing hardware, and additional infrastructure for growing AI workloads.

Alibaba AI Push Targets Larger Qwen Models

Alibaba’s Qwen team plans to develop models containing between 5 trillion and 10 trillion parameters. The company expects these larger models to handle increasingly complex and longer-horizon artificial intelligence tasks. Its current Qwen 3.8 Max flagship model contains 2.4 trillion parameters. Future Qwen 4.5 and Qwen 5 models could therefore operate at substantially larger scales. Alibaba is currently training its next-generation Qwen 4 model as part of this development roadmap. 

The company also reported progress toward systems that can improve models with limited human involvement. These systems can identify weaknesses, conduct experiments, and generate training data during the development process. Such capabilities could support more sophisticated artificial intelligence applications across different industries and use cases.

New Zhenwu V900 Chip Expands Alibaba AI Capabilities

Alibaba also unveiled the Zhenwu V900 AI chip through its T-Head semiconductor division. Company executives said the processor delivers three times the performance of its predecessor. The chip can connect within clusters containing up to 500,000 processors for large-scale AI computing. These clusters could help train and operate some of the world’s largest artificial intelligence models. Alibaba expects mass production and commercial availability during the first quarter of 2027. 

The company also expects significant growth in annual AI chip shipments following the new processor’s launch. This development expands Alibaba’s strategy beyond artificial intelligence software and into specialised domestic computing hardware. It also strengthens the company’s efforts to build an integrated technology stack for advanced AI workloads.

Alibaba Targets 20GW Data Centre Capacity By 2032

Alibaba plans to expand its global data centre capacity beyond 20 gigawatts by 2032. The company said customer demand for artificial intelligence remains exceptionally strong across its cloud business. This demand continues supporting faster revenue growth while increasing requirements for computing infrastructure. However, supply chain constraints could restrict how quickly Alibaba expands its available capacity. Company executives said industry demand currently exceeds their existing supply capabilities. 

Alibaba plans to bring AI supernodes online at commercial scale during the current quarter. These systems could provide additional computing resources for training and deploying increasingly advanced artificial intelligence models. The announcements also affected Alibaba’s Hong Kong-listed shares during Tuesday’s trading session. The shares rose 5.1%, reaching their highest level in approximately one month.

FAQs

What is Alibaba’s new AI model target?

Alibaba plans future Qwen models containing between 5 trillion and 10 trillion parameters for complex artificial intelligence tasks.

What is the Zhenwu V900 chip?

The Zhenwu V900 is Alibaba’s next-generation AI processor designed for large-scale model training and deployment.

When will the Zhenwu V900 launch?

Alibaba expects mass production and commercial availability of the Zhenwu V900 during the first quarter of 2027.

How much data centre capacity does Alibaba target?

Alibaba aims to increase its global data centre capacity beyond 20 gigawatts by 2032.

Why is Alibaba expanding its AI infrastructure?

Alibaba expects strong artificial intelligence demand to increase cloud workloads and drive additional infrastructure requirements.

Key Takeaways

• Alibaba plans future Qwen models containing between 5 trillion and 10 trillion parameters.
• The current Qwen 3.8 Max flagship model contains 2.4 trillion parameters.
• The Zhenwu V900 reportedly delivers three times its predecessor’s performance.
• Alibaba expects commercial Zhenwu V900 availability during the first quarter of 2027.
• The company targets more than 20 gigawatts of global data center capacity.

Conclusion

Alibaba is expanding its AI strategy across foundation models, specialized chips, cloud services, and data centres. Its planned Qwen models could significantly increase computing scale for complex artificial intelligence applications. The Zhenwu V900 adds another major component to Alibaba’s growing domestic AI hardware capabilities. 

Its data center expansion could also support increasing demand for model training and deployment. Together, these developments highlight Alibaba’s broader efforts to build an integrated artificial intelligence technology ecosystem.