Infinigence AI

Infinigence AI was founded in May 2023 by Professor Wang Yu of Tsinghua University's electronics department together with his students Xia Lixue (co-founder and CEO) and Dai Guohao, as a technology-transfer venture of Tsinghua University and Shanghai Jiao Tong University. The company positions itself as an AI-native infrastructure operator that neither makes chips nor trains models: its compute hub reaches about 37,000P of heterogeneous compute across 16 chip types and runs thousand-card mixed training on domestic chips (Ascend, Tianshu Zhixin, MetaX, Moore Threads) alongside AMD and NVIDIA hardware, with utilisation up to 97.6%; its Token factory provides an Agentic MaaS platform whose cross-cluster heterogeneous prefill-decode architecture cut first-token latency by 51.5% and per-token cost by 37.5%. In October 2026 it completed a RMB 2.3 billion Series C, taking cumulative funding to RMB 4.3 billion, a record for China's AI infra sector, and has reportedly filed confidentially for a Hong Kong IPO.

• 01Completed a RMB 2.3 billion Series C in October 2026, taking cumulative funding to RMB 4.3 billion - the largest single round and total for China's AI infra sector.
• 02Investor roster spans state capital (Hangzhou Hi-tech Jintou, social-security Zhongguancun fund), corporates (Tencent, Baidu, Lenovo) and financial investors (Sequoia China, Qiming, Hillhouse-linked Shunwei).
• 03Mixed training across Ascend, Tianshu, MetaX and Moore Threads chips with AMD/NVIDIA reached up to 97.6% utilisation; a 4,000+ km cross-domain training validation was done with China Telecom.

Not yet listed

Not yet listed

Not yet listed