Gasgoo Munich-Manifold AI and JD Technology have signed a strategic framework agreement. The partnership will focus on deploying world models across JD’s business scenarios, building compute infrastructure, and providing data services. Together, the companies aim to establish a continuous improvement loop for model iteration, develop embodied AI robotic solutions for vertical industries, and fast-track technical validation and commercial deployment.
The agreement centers on three core priorities. The two sides will deeply integrate Manifold AI’s robotic brain with JD Technology’s compute foundation and data services, tapping domestic compute infrastructure to train state-of-the-art Chinese world models and build industry-leading brain APIs. They also plan to jointly deliver integrated solutions that bundle algorithms, compute, and vertical applications. By leveraging JD’s internal use cases and external market opportunities — and linking up with supply-chain partners — the pair aims to fast-track technical validation and commercial rollouts for the entire suite.
Beyond that, the partners plan to establish a joint laboratory for frontier technology, co-apply for research grants, and help draft national and industry standards — all to strengthen the broader embodied AI ecosystem.
In the real world, deploying world models demands more than just code — it requires synchronized support across algorithm iteration, genuine business demand, hardware supply chains, compute resources, and scenario data. Manifold AI said it will team up with JD Technology and industry partners to align with sector-specific needs, pushing forward scenario validation for world models and embodied AI. The goal is to distill reusable industry solutions, probe sustainable business models, and ultimately drive tangible value from world models in physical settings.
JD Technology has been expanding its footprint in physical AI, using its supply-chain expertise, compute clusters, and scenario data to build world models and embodied AI systems. Manifold AI, for its part, has long focused on world models and robotic brain R&D. By linking models, compute, and scenario data, this partnership bridges the gap — pushing domestic embodied AI from the lab into large-scale commercial validation.
