Zhongke Shitong Hengqi has raised nearly 500 million yuan in Series B financing, according to September 7 reports from 36Kr and Taibo. The company builds software that connects business data, organizational rules and permissions, aiming to make AI agents useful inside operations where an incorrect action can have consequences beyond a bad answer on a screen.
Founder Huang Yu told 36Kr that its AOS system does not replace a large language model or a company's existing information systems. Instead, it gives agents a common representation of business objects and rules, with approval, audit and rollback controls. One example in the interview is matching customers to products under many business rules, rather than asking a model to improvise an unconstrained recommendation.
Taibo's reporting describes a related distinction in remote sensing: Shitong Hengqi works with commercial satellite companies and combines data from different sources, rather than trying to own the satellites or all the underlying imagery. Its role is the software layer that helps turn those inputs into decisions. That separates the business from a spacecraft manufacturer or a company selling launch capacity.
The company's organization profile on Kechuang China dates its founding to 2016 and describes a Chinese Academy of Sciences-incubated team and the Lingjing product family. It identifies decision support in specialized sectors as the focus. Its defense origins are relevant background, but claims that demanding customers establish universal reliability still need to be tested in each new deployment.
Reported participants in the financing include Beijing Information Industry Fund, China Unicom Venture Capital, GF Xinde and existing investor Leli Capital. The reports say the capital supports product development and expansion into areas including telecommunications, finance and energy. The amount is disclosed as nearly 500 million yuan; dollar conversions circulating in feed headlines should not be mistaken for an exact additional financing amount.
The civilian expansion presents a concrete commercial test. A telecom operator's rules, a utility's approvals and a financial firm's data permissions are not interchangeable. Selling this kind of software requires fitting the customer's actual workflow, not just connecting a capable model. The important question is whether deployments can repeat across organizations without each installation becoming an entirely bespoke project.