Jinghua Confidential Computing raises nearly RMB 100M Pre-A for private AI infrastructure

Zhongguancun Science City and Beijing Venture Capital co-led the Pre-A round. Huacang Capital and Jizhi Investment joined, and existing shareholders Inno-linked capital and Shengjing Jiacheng participated again. The amount was given as nearly RMB 100M, roughly $14M. Valuation was not disclosed. Jinghua builds confidential AI infrastructure so companies can run models on sensitive documents, financial data and customer records without exposing them. Products include encrypted model services, an agent development platform, protected chat applications and secure meeting recording. It says the stack runs on Chinese hardware including Huawei Ascend and Hygon systems, and reports token volumes and overseas deployments, none independently verified. For AI builders, the round shows privacy-preserving inference becoming a funded product category, in the same week Anthropic licensed technology and hired staff from US peer Opaque. Regulated buyers want proof that data stays protected, so teams selling into finance or healthcare should be ready to document where their models run.