China Launches 2026 Industrial Innovation Open Competition Initiative

China's Ministry of Industry and Information Technology and six other ministries have launched the 2026 'open competition' initiative targeting 24 innovation topics across six sectors, including AI, robotics, and advanced manufacturing. The campaign sets concrete 2028 performance benchmarks—such as achieving over 96 percent recognition accuracy in embodied AI—and introduces third-party evaluation to select top-performing industrial and research teams.

The initiative reflects China's shift toward goal-oriented, metric-driven industrial policy, using explicit quantitative targets and third-party assessments to accelerate the commercialization of cutting-edge tech like embodied AI and edge computing.

China's Ministry of Industry and Information Technology (MIIT) and six other government departments have officially launched the 2026 'open competition' initiative—a mechanism known in Chinese policy as *jiebang guashuai*—to accelerate breakthroughs in key industrial and information technology fields, state media reported on July 29, 2026. The program is designed to deepen the integration of scientific research and industrial innovation, utilizing targeted technological R&D to support high-quality economic development. The newly released scheme establishes a framework covering 24 specific topics across six priority directions: future industries, equipment manufacturing, information technology, communications, artificial intelligence, and consumer goods. Under the future industries track, the plan establishes five targeted topics, which encompass humanoid robotics and embodied intelligence, atomic-scale manufacturing technology, high-performance separation and purification equipment for biomanufacturing, brain-inspired intelligence, and cutting-edge laser manufacturing technologies. The artificial intelligence track is divided into four main topics covering foundational infrastructure technology, application-enabling technology, intelligent products and equipment, and common foundational support technology. Accompanying application guidelines set clear quantitative targets for project applicants. For embodied intelligence, the guidelines mandate that by 2028, multimodal large models must achieve integrated processing across at least four modalities—including vision, language, force and tactile sensing, and internal body state—while achieving a multi-scenario perception and recognition accuracy rate of over 96 percent. Additionally, applicants under relevant AI tasks are tasked with engineering edge-side world model systems with independent intellectual property rights, supported by a full development and deployment toolchain. To execute the project lifecycle, MIIT will lead the initial review process to select and publish an official shortlist of qualified candidate units. Once these shortlisted teams complete their respective research and development assignments, MIIT will commission independent third-party professional institutions to evaluate performance outcomes and select the final winning entities based on merit.