Longgang district of Shenzhen in South China's Guangdong province has launched the OpenIndustry Community, a platform aimed at using industrial knowledge graphs to equip artificial intelligence with an "industrial brain". 

OpenIndustry Community addresses the "misfit" of general AI models in manufacturing, where data bias and logical fallacies hinder meeting the high precision and safety demands of industrial production.

An aerial view of Longgang urban area.

The platform features three core components: the Industrial Ontology-based Knowledge Graph Center, the Industrial AI Application Center, and the Industrial AI Talent Center. These components utilize multimodal extraction, 3D analysis, and human-machine interface to transform fragmented engineering experiences into structured knowledge graphs.

Additionally, the platform introduces a dual-track points-sharing mechanism, allowing knowledge contributions to be exchanged for computational power, technical services, and even commercial opportunities, alleviating companies' concerns about sharing knowledge.

The platform was launched in Longgang due to its strong industrial foundation. The district hosts nearly 30,000 industrial enterprises across 133 categories, with an industrial output exceeding 1 trillion yuan (147 billion) for two consecutive years. 

From R&D to manufacturing, and from supply chain management to equipment maintenance, Longgang boasts a comprehensive real-world manufacturing environment that serves as the most critical "testing ground" for general-purpose AI models.

The output value of the AI industry in Longgang has already exceeded 4 billion yuan, and it aims to reach a trillion-yuan scale in AI and robotics by 2030.

At the launch ceremony of the platform, Longgang also released its first list of real-world "Industrial+AI" scenario needs, covering areas like integrated circuit testing, 3D printing visual calibration, and predictive maintenance.