Tsinghua-educated Fu Zhi, born 2000, raises $45.1M for compute startup Gongji, backed by CATL
What's the deal? Gongji Technology, a Chinese distributed-compute startup founded by Tsinghua University doctoral student Fu Zhi, has raised US$41.3M (≈$45.1 million) in an early VC round. Investors include the China Internet Investment FundDealroom has a profile for this one. Try Dealroom →, CATL-affiliated vehicles Ningbo Wending and Puquan CapitalDealroom has a profile for this one. Try Dealroom →, Oriza Puhua, Saimo Technology, and Deshi Investment. Existing backers Chunhua Venture, Lihe Zhongke, and Xinjin Venture joined again.
What's the endgame? Gongji taps GPU resources scattered across regions and clusters, then offers on-demand compute for model training, inference, and AI applications. Fu, born in 2000, likens the model to building a power grid — idle data-centre capacity gets pooled and dispatched where it is needed.
By the numbers: The company, founded in 2023, says it has served more than 10,000 companies and teams. It reported annual revenue growth above 800%, with quarterly revenue doubling into 2026 and annual recurring revenue topping US$27.5M.
Why now? Most capital in China's compute sector has chased large-scale training clusters built on heavy assets. Gongji bets on the opposite end: light-asset scheduling of fragmented, bursty inference demand, which it argues will spread across regions as AI applications proliferate.
The round follows a tens-of-millions-CNY angel round in October 2024 and a near-100-million-CNY Pre-A in December 2025. It widens the cap table to national funds, industrial capital, and market-driven firms.
What could go wrong? Delivering stable service from volatile, distributed supply is the core technical challenge. The company says its scheduling platform maintains an SLA above 99.9%, but scaling across varied chip architectures and nodes remains demanding.
The signal: The round sits in the 94th percentile by size among early VC deals in Chinese enterprise software over the trailing 48 months, underscoring how much capital is flowing to compute infrastructure. CATL's participation — alongside its earlier moves into robotics and sensors — points to a battery giant extending toward downstream AI applications, though with no disclosed stake or partnership terms, its strategic weight here stays unproven.
Read more: 36kr.com
Image credit: Gongji Technology
