Shanghai-based AI firm Infinigence AI has secured an additional 500 million yuan in Series A+ funding, led by state-owned capital. This brings its total financing to nearly 1.5 billion yuan. By leveraging heterogeneous scheduling technology, it harnesses domestic computing power and takes the lead in the large-model infrastructure sector.

On November 27, Infinigence AI, an AI infrastructure company established just two and a half years ago, announced the completion of a Series A+ funding round, raising nearly 500 million yuan. This round was jointly led by Zhuhai Technology Group and Futeng Capital, with existing investors Legend Capital and Shenwan Hongyuan also increasing their stakes.

With this latest funding, the company's cumulative financing over the past two and a half years has approached 1.5 billion yuan, making it the first unicorn in China's 'large-model computing power operator' sector.

The new funding will primarily be used to expand its multi-heterogeneous technology stack, accelerate the deployment of 'Infinigence AI Cloud' nodes in the Beijing-Tianjin-Hebei and Guangdong-Hong Kong-Macao Greater Bay Areas, and invest in the research and development of the next-generation Agentic Infra. The goal is to equip cloud and edge infrastructure with self-evolving capabilities.

The head of the leading investor, Zhuhai Technology Group, stated that the entry of state-owned capital was motivated by the scarcity of 'unified domestic computing power scheduling.' This approach 'not only addresses high-end chip restrictions but also lowers the computing power threshold for small and medium-sized enterprises.'

According to a recent report by industry analyst IDC, China's AI computing power market is projected to reach $19 billion by 2025. Third-party independent clouds will account for less than 10% of this market, and hybrid training using domestic chips remains an untapped area. Leveraging its first-mover advantage and rapid implementation, Infinigence AI has become the highest-valued startup in this niche sector.

Founded in May 2023 in Shanghai, Infinigence AI originated from the NICS Laboratory at Tsinghua University's Department of Electronic Engineering. Co-founded by Professor Wang Yu from Tsinghua's Electronic Engineering Department and PhD graduates Xia Lixue and Dai Guohao, the company focuses on the AI Infra intermediate layer, which lies 'above chips and below models.'

Its core product, the Infinigence Infini-AI heterogeneous cloud platform, already manages 25,000P of diverse heterogeneous computing power across 26 cities nationwide. It supports hybrid training and inference for over 20 large models and more than 10 chip types. Officially disclosed data shows a maximum utilization rate of 97.6% for thousand-card heterogeneous clusters.

In February 2025, the company led the way in adapting and optimizing DeepSeek-R1 and V3 across seven major domestic hardware platforms, including Biren, Hygon, and Moore Threads. This enables developers to deploy 'model + computing power' packages with a single click in the cloud.

Data from the Shanghai Municipal Commission of Economy and Informatization reveals that as of the end of October, Infinigence AI has connected to over 12,000P of intelligent computing power in the Yangtze River Delta, accounting for 38% of Shanghai's operational domestic computing power. It provides training and inference services to over 20 leading model enterprises, including Yuezhian, Baichuan Intelligence, and Li Auto.

Co-founder and CEO Xia Lixue disclosed that the platform's daily inference call volume has surged 17-fold since the beginning of the year. The company projects revenue exceeding 500 million yuan in 2025, with sustained positive cash flow.