As we move into 2026, the capital narrative in the embodied intelligence sector is undergoing a substantive shift, moving away from a blind pursuit of humanoid concepts in the lab towards a focus on mass production and delivery capabilities in real-world industrial settings.

On May 9th, general embodied intelligence technology company Xiaoyu Zhizao announced the completion of a Series B+ financing round worth several hundred million yuan. This round was jointly led by BAIC Capital, FOSUN RZ Capital, and C&D Emerging Industry Investment, with participation from existing shareholders HuaYe TianCheng Capital and Guizhou Science and Technology Innovation Angel Fund. DiDi Global Inc. and Xiaomi co-founder Li Wanqiang also made additional investments. Gengxin Capital acted as the exclusive financial advisor and also participated in the investment.

In 2026, as embodied intelligence accelerates its deployment in real industrial scenarios, this investor list spanning automotive, consumer electronics, and heavy construction equipment signals a clear move by industrial capital to strategically position itself for scaled applications.

The company's consistent fundraising over its three-year history is closely tied to its core team's background, with the founding team all hailing from Xiaomi Corp.. Founder and CEO Qiao Zhongliang is the former head of MIUI R&D at Xiaomi. Co-founder and CTO Wang Wenlin is the former General Manager of Xiaomi's Software System Platform Department. Partner Shi Jiangtong is the former General Manager of Xiaomi Ecosystem Chain Hardware R&D. All possess extensive experience in system development and hardware mass production for hundreds of millions of smart devices.

Unlike the frontier algorithm focus often seen with academic AI researchers, Xiaoyu Zhizao exhibits typical big-tech engineering characteristics.

Its technical approach avoids starting from scratch on the extremely complex challenge of a general-purpose hardware platform. Instead, it has established a "one-brain, multiple-forms" architecture, concentrating core resources on developing the embodied brain and using a single underlying system to adapt to various hardware types. This strategy, reminiscent of the software-defined hardware shift in the smartphone era, offers substantial engineering advantages in cost control and hardware-software decoupling.

The three new lead investors in this round—BAIC, FOSUN, and C&D—represent vehicle manufacturing, diversified industrial ecosystems, and heavy construction, respectively. This deep involvement by industrial capital reflects a core evolution in the embodied intelligence funding market: a shift from paying for technological concepts to paying for on-the-ground efficiency and bulk orders.

According to public information, Xiaoyu Zhizao has avoided the fiercely competitive consumer or light-interaction segments, targeting instead the challenging deep-water area of intelligent welding in heavy industry.

In 2024, the company partnered with Tangshan Panasonic on a large-model robot collaboration project and secured a strategic procurement order for hundreds of units from a leading heavy industry player. The initial batch of mass-produced units is currently being delivered.

The heavy investment by industry giants at this juncture is, in essence, an endorsement of its platform technology and its commercial closed-loop in heavy industrial scenarios. Investors aim to introduce automated production capabilities into their own industrial chains at optimal cost through this investment, seeking efficiency advantages in manufacturing transformation.

This year, the embodied intelligence sector has shown a polarization trend, with funding concentrating at the top and consolidation at the bottom. Xiaoyu Zhizao's completion of its Series B+ round solidifies its position in the industrial embodied intelligence cohort, but significant challenges remain ahead.

The foremost challenge is the engineering barrier to scaled replication. Heavy industrial scenarios impose extremely high demands on equipment fault tolerance in extreme environments and continuous operational stability. Scaling from initial deliveries of tens of units to deployments of thousands presents an exponentially increasing engineering difficulty.

Secondly, there is the challenge of the large model's generalization capability. Achieving smooth system migration across vastly different industries like automotive and shipbuilding requires continuous breakthroughs in underlying generalized reasoning abilities and the accumulation of massive amounts of industrial know-how data.

As the industry's focus shifts from the laboratory to the production line, competition in embodied intelligence has evolved from a singular algorithm contest into a comprehensive commercial race encompassing cost structures and yield assurance. Securing funding is merely obtaining an entry ticket; the true test of mass production is just beginning.