Jinghua Misuàn (荆华密算) closes angel round of “tens of millions” RMB to push AI confidential computing

Funding and focus

It has been reported that Jinghua Misuàn (荆华密算) has closed an angel financing round worth several tens of millions of RMB to accelerate development and commercialization of its AI confidential computing stack. The round was reportedly led by Shengjing Jiacheng Venture Capital (盛景嘉成创投), with strategic participation from Guolian Co., Ltd. (国联股份), Beyondsoft (博彦科技) and returning investor Inno-Creation Fund; Guangyuan Capital (光源资本) served as financial adviser. Proceeds will go toward R&D of a high‑performance AI confidential computing platform, verification of a domain‑specific secure compute chip, and commercial roll‑out across government, finance, healthcare and legal sectors.

Technology and partnerships

Jinghua Misuàn says it focuses on privacy‑preserving computing for AI — enabling secure model training and inference on sensitive data without exposing plaintext to external models or services. It has been reported that the company, working with Professor Ren Ju’s lab at Tsinghua University, reconstructed AI operators to run confidentially on domestic heterogeneous GPUs without relying on GPU‑TEE, and claims to have cut AI confidential‑computing time costs by 3–4 orders of magnitude; a prototype chip design has already been validated on FPGA. The company offers localized deployment for enterprises, a confidential inference platform for developers and consumers, and a confidential training engine that supports encrypted data exchange; multiple pilot projects are underway.

Market implications and geopolitical context

Why does this matter? As AI agents gain autonomous decision‑making and file‑handling abilities, traditional “encrypt‑before‑upload” approaches struggle to contain millisecond‑scale overreach and data exfiltration risks. Jinghua Misuàn positions confidential computing as the underlying trust layer that could make enterprise AI adoption feasible. Geopolitically, the work also aligns with China’s push for compute and chip self‑reliance amid export controls and broader supply‑chain pressures: compatibility with domestic GPUs and a homegrown secure‑compute chip are strategic as well as commercial priorities. With industrial investors on board who bring customer access — notably Guolian’s network of manufacturing clients and Beyondsoft’s overseas compliance expertise — the company aims to couple technical claims with route‑to‑market partners rather than merely raise capital.