Hangzhou is positioning itself as a leading hub for artificial intelligence (AI) development, underscored by a recent wave of investment totaling US$3.71 billion (CNY25.5 billion) in 12 AI projects. The announcement, made following an AI development summit at the Hangzhou Civic Center on February 28, signals a strategic push by the city and the broader Chinese government to bolster domestic capabilities in the critical AI sector. A key component of this initiative is a focus on securing the supply of essential computing power, particularly in the form of advanced graphics processing units (GPUs).
The largest single project within this investment is a US$3.71 billion research and development initiative focused on high-performance GPUs and inference chips, led by Sunrise, a Zhejiang-based company. This project reflects a growing trend in China: a move away from reliance on a single supplier for AI chips and towards a multi-vendor ecosystem. The development comes as China seeks to reduce its dependence on foreign technology, particularly from the United States, amid ongoing geopolitical tensions and export controls.
Sunrise Leads Hangzhou’s GPU Push
Sunrise, which was spun off from SenseTime at the end of 2024, is developing high-performance GPUs and multimodal inference chips for commercial applications. According to Sunrise Co-CEO Wang Zhan, the future of industrial development hinges on computing power, not just model capability, and Hangzhou’s AI strategy provides a supportive environment for the company’s expansion. The company has cited multiple generations of validated products, though large-scale delivery claims, such as “ten-thousand-card-scale delivery,” originate from corporate and local disclosures rather than independent verification.
This investment in Sunrise is part of a broader effort by local governments to support domestic AI computing providers. Yet, Sunrise is not alone in this endeavor. China’s AI chip deployment is evolving into a multi-player structure, with several companies vying to meet the growing demand for AI processing power.
A Multi-Vendor Approach to AI Chips
Huawei’s Ascend series, including the 910C, is emerging as a primary domestic platform for large-model training, and inference. Reports indicate Huawei is preparing for large-scale shipments to compete with Nvidia, though manufacturing capacity and policy constraints remain challenges. US officials have estimated that Huawei’s 2025 shipments may be capped at around 200,000 units due to these limitations, highlighting ongoing supply chain and technological hurdles. Reuters reported on the challenges and opportunities facing Huawei in the AI chip market.
Beyond Huawei, companies like Cambricon are gaining prominence. Its MLU series of AI accelerators has been included in government and enterprise procurement programs, according to The Financial Times, alongside Huawei chips. Alibaba’s T-Head, MetaX, and Biren are also contributing to the domestic AI chip supply, as evidenced by their deployment in the China Unicom data center in Xining, Qinghai, Reuters reported.
Moore Threads and Enflame Technology are also expected to supply chips for future phases of the China Unicom project and related cloud initiatives. Enflame is reportedly advancing an IPO on the STAR Market, indicating continued investment and growth in the sector.
Expanding Domestic CPU Alternatives
The push for domestic AI chip solutions extends to central processing units (CPUs) as well. Companies like Loongson and Phytium are being positioned as alternatives to Intel and AMD in government, state-owned enterprises, and industrial server environments. China is progressively limiting the use of foreign-made chips in government-funded systems to promote domestically developed alternatives, providing policy support for these vendors.
The shift towards a multi-vendor AI chip ecosystem in China represents a significant strategic move. While gaps remain in computing scale, manufacturing capacity, and global ecosystem integration compared to established international players like Nvidia, the country is making substantial progress in building a more self-reliant and competitive AI industry. The Hangzhou investment and the broader national strategy signal a long-term commitment to innovation and technological independence.
Looking ahead, the success of this multi-vendor approach will depend on continued investment, overcoming manufacturing challenges, and fostering a robust ecosystem of software and support services. The coming years will be crucial in determining whether China can achieve its goal of becoming a global leader in AI.
What are your thoughts on China’s strategy to develop its own AI chip industry? Share your comments below.
