Jinghua Secret Computing announced the completion of tens of millions in seed round financing to promote the implementation of the high-performance encrypted computing industry.

Recently, Beijing Jinghua Cryptographic Computing Technology Co., Ltd. (hereinafter referred to as "Jinghua Cryptographic Computing"), a high - performance encrypted computing technology enterprise, officially announced the completion of a seed - round financing of tens of millions of yuan. This round of financing was led by Inno Angel Fund, and Light Source Capital served as the incubator and exclusive financial advisor. The funds from this round will be mainly used for the research and development of two core engines, the high - performance encrypted inference engine and the high - performance encrypted training engine, as well as the recruitment of product - oriented talents.

It is worth mentioning that with its disruptive technical solutions, Jinghua Cryptographic Computing won the national championship of the "Innovation China 2025 DEMO CHINA" competition organized by StartupsBang at the early stage of its establishment. Its founder, Lin Xiuchun, also won the highest honor, the "DEMO GOD" title, becoming the first post - 2000s person to receive this honor.

The Call of the Times: The "Privacy Paradox" in AI Development and the "Jewel in the Crown"

In today's increasingly advanced artificial intelligence era, there is a great need to use artificial intelligence in privacy - sensitive scenarios. However, due to the serious risk of data leakage, it is difficult for users to hand over sensitive files and problems to cloud - based large models for reading and use. Lin Xiuchun, the co - founder of Jinghua Cryptographic Computing, learned through extensive market research that most people do not need "24 - hour privacy" and can hardly afford the high cost of local deployment of large models. However, everyone may need "10 - minute privacy" at some point in their lives to solve the potential privacy - leakage problems when using cloud - based large models. Meanwhile, in the field of data element circulation, the characteristics of data, such as "zero - cost copying and zero - cost dissemination", make it easy for customers who have purchased data to resell it, causing great risks of interest and privacy losses for data providers.

High - performance encrypted computing is regarded as the ultimate solution to this paradox. It can complete calculations without decrypting data and was once called the "jewel in the crown" in the relevant field. However, the performance loss of traditional solutions is as high as a thousand times or even ten thousand times, making it like a "dud that won't go off for ten years" and difficult to implement in computing - intensive scenarios such as AI. The market urgently needs a truly usable high - performance solution.

In this context, the technological breakthrough of Jinghua Cryptographic Computing is particularly scarce. In September, Jinghua Cryptographic Computing first launched a DEMO prototype, which won a major award in the startup industry and attracted high attention from the media as soon as it came out. Important media such as People's Daily, Xinhua News Agency, and Guangming Daily reported on its award - winning situation and innovative achievements.

Technological Breakthrough: From Ten - Thousand - Fold Loss to Three - Fold, Making Encrypted Computing Truly Usable

The disruptive innovation of Jinghua Cryptographic Computing stems from its top - notch industry - academia - research team. The company's chief scientist is Professor Ren Ju, a Changjiang Scholar, the director of the Institute of Human - Computer Interaction and Media Integration at the Department of Computer Science and Technology of Tsinghua University, and the first - place winner of the First - Class Award for Natural Science of the China Institute of Electronics. The company has gathered a R & D team with a triple background in "data security, artificial intelligence, and software - hardware computing optimization" and has been deeply involved in the cross - field of encrypted computing and AI for many years. In the process of combining academic frontier exploration with high - difficulty engineering implementation, the team has experienced a large amount of development and trial - and - error, ultimately building a very high technological barrier.

After five years of research, Jinghua Cryptographic Computing has uniquely constructed an encrypted computing solution based on new - type cryptography. Without relying on specific GPU - TEE, it has reduced the time loss of AI encrypted computing from more than 1000 times to about 3 times, achieving the best balance among performance, compatibility, and security for the first time, and currently leading both at home and abroad.

This solution abandons the optimization of general scenarios and focuses on AI computing instead. Its advantage lies in that it can guarantee security from a mathematical principle, does not rely on specific hardware (such as GPU - TEE), and can be compatible with various domestic GPUs and NPUs. While ensuring encrypted computing on GPUs/NPUs, it provides a computational cost that the industry can bear, paving the way for the large - scale application of encrypted technology in the AI field.

Addressing Pain Points: "Using AI Large Models in Privacy - Sensitive Scenarios" and "Trusted Circulation of Data Elements"

Based on its core technological breakthrough, Jinghua Cryptographic Computing plans to officially launch two engine products on the market next year:

· High - performance encrypted inference engine: It directly addresses the "10 - minute privacy" needs of individual and enterprise users. When users use cloud - based encrypted large models, the system will automatically encrypt problems and files locally on PCs and mobile devices into an encrypted state and then send them to the cloud for "usable but invisible" computing. The calculated results will also be sent back to the local in ciphertext and then decrypted into plaintext locally. This process can ensure the security of data during transmission, storage, and computing.

· High - performance encrypted training engine: This engine is mainly designed to protect the rights and interests of data - owning enterprises in data transactions. It focuses on solving the problem that customers who have purchased AI training data sets may easily resell the data, causing losses of interests and privacy for data providers. The Jinghua Cryptographic Computing team proposed the term "post - sleep income", expecting to rent out data assets and realize repeated monetization on the premise of "zero data leakage", forming a continuous income pipeline.

Currently, the project in which Jinghua Cryptographic Computing participated in the design has been selected as one of the first - batch "Trusted Spaces" pilot projects in China. It has successfully helped a well - known multinational enterprise transfer hundreds of millions of user data from regions such as the EU and South America to its self - built servers for model training, preliminarily verifying its commercial potential.

The AI encrypted computing engine based on cryptographic principles itself has a high degree of security and universality. If the performance problem that has been restricting its development can be solved, the team believes that this solution will become a new consensus for using AI large models in privacy - sensitive scenarios and "cross - domain circulation" of data elements. Jinghua Cryptographic Computing aims to deepen its technological moat in high - performance encrypted computing, continuously provide advanced solutions for major pain points around the encrypted engine, and ultimately become a pioneer in leading this consensus.

Inno Angel Fund said: "Jinghua has an internationally leading and breakthrough encryption algorithm. This technology has great commercial potential, which can solve the privacy - leakage problem when sensitive data is used in large models, enabling more vertical scenarios to use cloud - based public large models with confidence. At the same time, the encrypted engine can also protect the rights and interests of data owners in data transactions and is expected to become the technical infrastructure for data transactions. The Jinghua project is a typical case of the transformation of cutting - edge academic achievements from universities. With the profound accumulation of the national key laboratory of Tsinghua University and a post - 2000s CEO who understands technology and is full of enthusiasm, Inno is willing to issue the first check to support outstanding young people and believes in the power of innovation."

Huang Xinxin, the person - in - charge of the 3i incubation business of Light Source Capital, said: "We firmly believe in the long - term development of the two major fields of AI and data element circulation. The Jinghua Cryptographic Computing team has solved the world - class problems of the implementation of encrypted computing in privacy - sensitive AI scenarios and the trusted cross - domain circulation of data elements with world - class technological originality. This is a long - awaited breakthrough in the industry. We believe that Jinghua Cryptographic Computing will become an important force in defining the new paradigm of next - generation data security and AI applications."