Tongzhiyuan and CIDI Sign Strategic Cooperation to Advance Autonomous Driving in Mining

On October 19, Tongzhiyuan and CIDI Intelligent Driving signed a strategic cooperation agreement. The two sides will collaborate around unmanned mining scenarios, with the first project expected to land in 2026. On the same day, Tongzhiyuan also announced the completion of an angel round of financing. The investors include well-known domestic investment institutions and senior industry figures, and the company's valuation has already reached the level of a unicorn. Public information shows that Tongzhiyuan was established in August 2026. Its business focuses on the implementation of physical AI in vertical scenarios, and it has simultaneously received support from both industry and capital. The essence of this investment is not betting on the track itself, but on the team behind it. After the founders left Huawei, they brought out a team that can be counted on one hand. The founding team has a clear technical background and is a team with a "hardcore proof" from Huawei's system. An investor said in an interview that the founders are top-tier talents who have experienced the main battles of Huawei's smart car business. The core members of the team come from Huawei's 2012 Lab, the Intelligent Automotive Solution BU, Tsinghua University, the Chinese Academy of Sciences, and other institutions, and have previously worked on AI research and system development. Tongzhiyuan's founder and CEO is Dr. Tan Jianjun, who has industry, academia, and research experience and serves as an adjunct professor at a university. After leaving Huawei, he founded Tongzhiyuan in 2012. The company's core team participated in the development of ADS 2.0 and 4.0 and is advancing pre-research on ADS 6.0. The team's core members have served as committee members of academic committees and have been deeply involved in the implementation of end-to-end large models. During their time in the BU, the team experienced the transition from rule-based to end-to-end large models, and possesses a system that has undergone millions of kilometers of real-world road testing and closed-loop training. Tan Jianjun said that the team's goal in autonomous driving is not to repeat the past, but to use Tongzhiyuan's methodology to solve the real problems of mining scenarios. In the field of intelligent driving, they want to create a "sample." The company's name, "Tongzhiyuan," is officially interpreted as exploring the path to the ultimate in intelligence and connecting the paths to general intelligence. It has no mysterious meaning. The technical system is called PSI, which is derived from the Greek letter psi and represents the bridge from the physical world to the digital world, as well as the closed loop from multimodal perception to real-world implementation. The data collection in the mining area is relatively simple, and the cost of expert annotation is low, making it suitable for large-scale implementation. It can also allow the industry to use the data collected in the mining area to train models. Tan Jianjun said that the biggest problem in the implementation of autonomous driving is the structural deviation between simulation and reality. For example, the abnormal behavior of equipment is a deviation that is difficult to predict in advance, and the model lacks the ability to respond to it. Tongzhiyuan hopes to use a path from Real to Sim to connect the entire process, and to use the data generated in the mine to drive the model, thereby reducing the friction between simulation and reality. The goal is to use visual data to supplement scientific learning and sign off on the implementation. Tan Jianjun's judgment on this is: the data in the real world can clearly tell you what is what, what is abnormal, and what is a long-tail scenario. The PSI system connects data, planning, execution, and real-world learning and verification into a closed loop. The goal is not to repeat a fixed trajectory at a single point, but to turn the failures that occur in the mine, the abnormal indicators that humans have identified, and the process into training experience, and then transfer it to the mine. Tan Jianjun said that the mine is the first battlefield, and there is currently a lack of real data and strong methods. The China Association of Automobile Manufacturers released the "Intelligent Connected Vehicle Development Report (2026)" on October 16, pointing out that the large-scale implementation of intelligent driving in mining areas still faces shortcomings such as data, algorithms, and generalization. This change requires new data collection methods, subtle improvements, and a model that has not yet formed a scalable and transferable engineering paradigm for every road. This is also one of the reasons why Tongzhiyuan chose the mining area as its entry point. The mining area is a scenario where the value of unmanned driving is most easily highlighted. At the same time, the implementation of regulations, standards, and safety in the mining area is also a test of the effectiveness of visual perception. Tongzhiyuan plans to complete the PoC of the entire vehicle, module, and components in the mining area, and expects to launch the first project in 2026. Tongzhiyuan's testing of the mining area is not limited to the mine itself. The battlefield is whether it can form a closed loop of data, business, and responsibility. The first PoC project can be delivered at a single point, which is Tongzhiyuan's first verification. The model can be transferred from the mine to the real battlefield. (Reposted from the Smart Car Platform. The purpose of this repost is to disseminate information. The views in the article do not represent this website. The information has not been verified. Investors are advised to refer to it. If there is any infringement, please notify this website and we will handle it as soon as possible.) Keywords: , This is an original article on smart manufacturing. Author: Xiaobian. This article is an original article on the intelligent manufacturing network. Please indicate the source when reprinting. This article is an original article on the intelligent manufacturing network. Please indicate the source when reprinting. All rights reserved. 1. All works marked "Source: xxx" on this website are reprinted from other websites. The purpose of reprinting is to disseminate information. The copyright belongs to the original author. If there is any infringement, please contact us. 2. All works marked "Source: xxx" on this website are reprinted from other websites. The purpose of reprinting is to disseminate information. The copyright belongs to the original author. If there is any infringement, please contact us. 3. All works marked "Source: xxx" on this website are reprinted from other websites. The purpose of reprinting is to disseminate information. The copyright belongs to the original author. If there is any infringement, please contact us. 4. All works marked "Source: xxx" on this website are reprinted from other websites. The purpose of reprinting is to disseminate information. The copyright belongs to the original author. If there is any infringement, please contact us. Hot words: speed reducer, harmonic drive, Nabtesco, RV reducer, cycloidal pinwheel reducer, special grease for reducers, precision reducer, European reducer. Phone: 0571-87759904 Customer service hotline: 0571-87756395 Procurement hotline: 0571-87759926 Consulting QQ: 1060309408 Website customer service QQ: 386392707 Media QQ: 1271141964 Procurement consulting QQ: 342806899