(Yicai) July 16 -- Huawei Cloud is applying the collaborative approach behind its automotive partnership with Seres Group to the embodied artificial intelligence sector, teaming up with humanoid robot developer Yijiahe Technology to jointly develop and commercialize robotics products.

The 'Seres Model' draws more on the open collaboration philosophy of the intelligent automotive industry, Wu Xiaodong, director of Huawei Cloud's Embodied AI Innovation Lab, told Yicai ahead of the World Artificial Intelligence Conference. It continuously improves platform capabilities through joint research and development and joint validation, rather than simply copying the automotive business model, he added.

Huawei Technologies' collaboration with Seres illustrates the model Wu is referring to. The technology equipment giant and automaker began collaborating on vehicle development in 2021. Unlike Huawei's traditional component supply business or its Huawei Inside model, the partnership involves Huawei in product definition, research and development, brand marketing, and omnichannel sales, while Seres is responsible for vehicle manufacturing, quality control, and supply chain management. Their jointly developed Aito brand delivered 423,000 vehicles last year and 161,000 in the first half of this year.

That approach is reflected in Huawei Cloud's partnership with Yijiahe, signed earlier this month. The collaboration combines the Nanjing-based robot maker's expertise in industries including power utilities and healthcare with Huawei Cloud's CloudRobo embodied AI development platform, computing resources, AI models, and underlying capabilities such as data governance, model optimization, and robot operation management. Rather than following a traditional supplier relationship, the companies aim to establish a viable business model in one industry before expanding it to others.

Building an Open Robotics Ecosystem

Wu said 2026 will be a pivotal year for the embodied AI industry, arguing that the biggest challenges extend beyond shortages of high-quality data and advanced models to broader gaps in development platforms, engineering systems, and commercialization capabilities.

High-quality multimodal data remains insufficient, while simulation environments still differ significantly from real-world conditions, Wu said. "How to quickly deploy large models to robots after training and continuously optimize them, as well as development processes such as data governance, model training, and motion control, remain highly fragmented," he said. "These issues are becoming major obstacles to large-scale commercialization."

Huawei Cloud expects the embodied AI industry to evolve into an open ecosystem rather than being dominated by a handful of companies with end-to-end capabilities. The ecosystem is expected to include AI model developers, robot manufacturers, and numerous companies specializing in specific industry applications.

Wu said the CloudRobo platform will remain open to partners with stronger data generation capabilities, AI models, or development tools. He also stressed that the embodied AI industry is still at an early stage and that meaningful results will require sustained investment and long-term cooperation rather than short-term breakthroughs.

Commercialization Begins With Power Sector

Huawei Cloud is focusing first on practical industrial applications.

Huawei Cloud selected the power sector as its first area of cooperation because embodied AI should initially be deployed in hazardous environments, including live electrical operations, where people cannot safely work, as well as in labor-intensive jobs that are difficult to fill, rather than simply replacing human workers, Wu said.

He added that robotics commercialization has generally begun in highly standardized industries with clear economic value, including industrial manufacturing, energy inspection, warehousing, and logistics, both in China and overseas.

Editor: Emmi Laine