Analysis: Alibaba’s New Processor Shows Applications Are Key to AI Chip Success

1. On January 29, Alibaba Group's chip subsidiary, T-Head, unveiled the Zhenwu 810E, a high-end artificial intelligence (AI) chip. This marks the confirmation of industry speculation about Alibaba creating its own proprietary processing power unit (PPU). A PPU is a highly integrated circuit specifically designed for large language model (LLM) inference, focusing on achieving high computational performance while maintaining low energy consumption.[para. 1][para. 2]

2. The Zhenwu 810E chip incorporates both AI training and inference functions and is recognized as the PPU previously featured in a CCTV broadcast from September 2025. Hardware-wise, its performance places it between Nvidia’s modified chips for China—the A800 and H20. The Zhenwu 810E has 96GB of HBM2e memory, matching Nvidia’s H20 and surpassing the A800’s 80GB. Its interconnect bandwidth is 700GB/s, exceeding the A800’s 400GB/s but just below the H20. The chip uses a PCIe 5.0 interface, equaling H20 and surpassing the A800.[para. 3][para. 4]

3. Demand for the Zhenwu 810E has outstripped supply over the last year, establishing it as one of the most shipped domestically produced AI accelerators. Large-scale deployments, with clusters exceeding 10,000 chips each, serve over 400 customers. Notable clients include the State Grid Corp. of China, the Chinese Academy of Sciences, and XPeng Inc., indicating substantial market penetration and trust in the product.[para. 5]

4. This milestone highlights the full emergence of Alibaba’s “AI golden triangle,” which consists of Tongyi Lab, Alibaba Cloud, and T-Head. Alibaba is now only the second tech company in the world, after Google, to possess fully integrated, in-house capabilities covering LLMs, cloud computing, and AI chip production.[para. 6]

5. According to domestic reports, the cumulative shipments of the Zhenwu PPU have reached several hundred thousand units. This volume has surpassed those of Cambricon Technologies (688256.SH), previously one of China’s leading AI chipmakers. Many domestic competitors, like Moore Threads (688795.SH), Metax Integrated Circuits (688802.SH), and Shanghai Biren Technology, face challenges in ensuring sustainable growth without stable, continued demand from application scenarios, despite respectable financial metrics at the time of listing.[para. 7][para. 8]

6. T-Head’s rapid success is closely tied to Alibaba’s strong position as one of China’s top internet firms. Alibaba Cloud, serving as one of the country's largest cloud service providers, generates immense AI computing needs—ranging from e-commerce recommendations to intelligent customer service—thus providing T-Head with a reliable deployment base. T-Head was founded in 2018 after Alibaba acquired Hangzhou C-Sky Microsystems and combined it with its DAMO Academy chip team.[para. 9][para. 10]

7. Since its inception, T-Head has launched a series of successful chips, including the Xuantie 910 processor, Hanguang 800 AI inference chip, and the Zhenyue 510 controller. All these have been effectively deployed within Alibaba Cloud, facilitating continuous improvement and performance optimization through access to a vast user base and diverse AI applications.[para. 11]

8. A key success factor is the deep synergy between Alibaba Cloud’s full-stack technical capabilities and T-Head’s hardware, resulting in technical and commercial spillover that accelerates commercialization. The deep ecosystem integration seen at Alibaba is mirrored at other Chinese tech giants, such as Huawei and Baidu, who also benefit from in-house chips tightly coupled with their software and cloud services.[para. 12][para. 13]

9. Internationally, Google’s development of the Tensor Processing Unit (TPU) illustrates a similar model—embedding chips across cloud services and products while leveraging massive internal demand for scale and integration. T-Head follows suit, packaging its chips within Alibaba’s offerings to generate tangible commercial value derived from close access to data, traffic, and end-users.[para. 14][para. 15]

10. According to company executive Qi Xiaoning, Alibaba’s semiconductor approach is to focus on real-world scenarios and application-driven innovation, distinguishing it from traditional chipmakers. The sector's competitive landscape in China looks to increasingly prioritize ecosystem coordination and integration over cutting-edge process nodes or theoretical peak performance. Scaled AI chip deployment, industry experience shows, hinges more on stable application scenarios than raw technical prowess.[para. 16][para. 17]