On April 10th, Zhiyuan Kuto held its inaugural media briefing in Shanghai, where COO Qiu Heng comprehensively outlined the strategic positioning and commercial ambitions of this new, independent subsidiary of the Zhiyuan Group. Zhiyuan Kuto was officially registered on April 9th, having been spun off from the Group's Lingxi product line. Qiu Heng explained the rationale for the spin-off directly: the quadruped business is growing too rapidly and would be overshadowed if it remained under the umbrella of the humanoid robot division. The Group's assessment is that the quadruped segment is itself a giant and needs to operate independently to reach its full potential. Qiu Heng repeatedly emphasized a core concept: Zhiyuan Kuto is not building robotic dogs, but quadruped robot "persons." The distinction lies in Artificial General Intelligence (AGI). His exact words were striking, stating that robotic dogs currently on the market are "all brainless," and that "there is simply no competition in making quadruped robots."
This statement reflects confidence, underpinned by technological reuse from the Zhiyuan Group. The "brain" and "cerebellum" are almost entirely shared, training data is highly transferable, and the workload for training quadruped operation models using Genie Studio is reduced by 70% compared to starting from scratch. The financial targets are equally ambitious. The revenue goal for 2026 is 500 million yuan, aiming for 10 billion yuan by 2030, with overseas revenue contributing 35%. Sales in the first quarter have already surpassed the total for the entire previous year. Medium-sized quadruped robots are selling out, requiring the recall of demonstration units for a partner conference on April 17th. Overseas orders accounted for over 40% of Q1 sales. However, Qiu Heng also set clear boundaries. There are no immediate plans for fundraising, citing leadership's view that "this market is huge, worth a great deal of money, so don't sell out too early." The focus is also not on the consumer market for now; instead, the priority is to deeply penetrate industrial scenarios like power grids, mines, security, and logistics. "What we want to create is a productivity tool," he stated. When asked about competition with companies like Unitree and Yun Shen Chu, Qiu Heng's response was thought-provoking. He said he currently "does not see intense competition" in the market, as customers are purchasing intelligent capabilities. "If a customer buys a 'dumb dog,' they won't want to buy a second one." Regarding supply chain costs, he cautioned against looking solely at quadruped shipment volumes, noting that humanoid and quadruped robots share a supply chain. The total volume of joints across both categories represents the true scale. The upcoming D2 Max is positioned as a key product for differentiation, claimed to be the first all-terrain L3 autonomous quadruped robot, capable of visual perception, autonomous navigation, and multi-task generalization within a single industry. Qiu Heng predicts that 2026 will be the inflection point for quadruped robots. Following is an edited excerpt of the conversation: On the Logic of the Spin-off Qiu Heng: Zhiyuan Kuto is the most important subsidiary of Zhiyuan, different in nature from previous ventures like hands, data, or leasing. It was spun off from a core product line. The future plan is for Zhiyuan to focus on humanoid robots and Kuto on quadruped robots. Chairman Deng has high expectations for us; Kuto is destined to become very, very large. On Fundraising and IPO Qiu Heng: We暂时 won't pursue fundraising because we don't want to dilute equity. Leadership believes this market is enormous, worth a lot of money, so we shouldn't sell out prematurely. As for an IPO? We look forward to that day. On Competition Reporter: Companies like Unitree already have significant shipment volumes, and Yun Shen Chu is also large. How do you address their cost advantages? Qiu Heng: Comparisons shouldn't focus solely on quadrupeds, because humanoids and quadrupeds share a supply chain. Consider how many joints a humanoid has versus a quadruped; the total combined volume determines who has greater scale. 2024 is Zhiyuan Kuto's first year of creation. We aren't claiming we'll be number one in year one, but I can say our pace to becoming number one will be faster than any other company. Reporter: Will prices fall? Qiu Heng: Costs and prices will certainly decrease rapidly. But at least for now, we don't see intense competition in the market. Most often, customers prioritize whether you can serve their needs well, not a price difference of one or two thousand. If a customer buys a 'dumb dog,' they won't want to buy a second one. On the Core Difference of Intelligence Qiu Heng: All existing robotic dogs in the industry are essentially 'blind.' They only use reinforcement learning to adjust their gait after a misstep. A normal person seeing a pit would either go around it or jump over it, not step into it. What Kuto is building looks at the path first with vision, then chooses a route, and then walks. Reporter: Is intelligent capability a real demand or a fabricated one? Qiu Heng: If someone says they don't need intelligence, the only reason is that they believe true intelligence is unattainable or they can't afford it. It's like asking someone if they want to live in a 1,000-square-meter house? Probably everyone does, but they might think it's too expensive. Our intelligent capabilities are about unlocking this latent demand. On L3 Autonomy Reporter: L3 for cars is just beginning trials. How do you ensure safety for L3 in quadruped robots? Qiu Heng: One advantage for quadruped robot autonomy is their limited operational scope; autonomy operates within a confined area. The challenge is all-terrain capability. Cars don't drive into mountains or ditches, but quadrupeds could very well go up hills, into gullies, or rivers. We describe it as 3D autonomy, whereas car autonomy is relatively two-dimensional. On Shipments and Industry Penetration Qiu Heng: We already have shipments in six scenarios: patrol, inspection, emergency response, firefighting, education, and entertainment. But all sectors are at a very early stage; overall market penetration is less than 1%. Our Q1 volume equaled last year's total, and we expect Q2 to exceed the full-year sum. Reporter: When is the tipping point? Qiu Heng: 2026 is already a tipping point. The feedback loop for quadruped robots might close faster than for humanoids because they are relatively simpler. On the Consumer Market Qiu Heng: There are暂时 no plans for a B2C approach. It's B2B first, then potentially B2C, which relates to technological maturity. Industrial environments are relatively constrained, making it easier to achieve a closed loop. Home environments are highly diverse and require more complex intelligence. We aim to create productivity tools. First, we'll master power grids, mines, security, and logistics before considering whether to work in homes. On Data Reuse Reporter: Zhiyuan has over 3 million real-world data samples. Given the different forms of humanoids and quadrupeds, how is data reused? Qiu Heng: We are training the same 'brain.' Humanoid robots don't all look identical either; Genie, Yuanzheng, and Lingxi have different forms, but they all contribute to data collection for training the same core intelligence. The difference isn't that great; it might require a little extra time, but not starting completely from scratch. On Cost Reduction Qiu Heng: As shipment volumes increase severalfold over the coming years, a cost reduction of 20-30% should be normal. Cost reduction primarily comes from scale itself. On 6G and Communication Qiu Heng: 6G standardization is about 9 years away; we hope to be among the first to adopt it post-standardization around 2029. Integrated sensing and communication in 6G is particularly important for us. Robots need radar, ranging, sensing, and terrain scanning; ideally, communication happens simultaneously during scanning. On Future Forms Reporter: Why not develop hexapods? Qiu Heng: The future could include hexapods or octopods, but first, we must perfect the quadruped form. A vast number of animals in the world are quadrupeds, indicating this form has high adaptability in nature. Once the intelligent brain is developed, it could control 100 legs. The key is developing the brain first; once that's achieved, it can be installed on any platform, turning it into whatever 'person' is needed.
Comments