Dialogue with Zhiyuan Kutuo's COO: There is basically no competition in making quadruped robots right now

April 10th, Zhiyuan Kuto held its first media briefing in Shanghai, where COO Qiu Heng comprehensively explained the strategic positioning and business ambitions of Zhiyuan Robotics, a brand new independent subsidiary. Zhiyuan Kuto was registered on April 9th, having been spun off from the Lingxi product line of Zhiyuan Group. Qiu Heng stated the reason for the spin-off very directly: the quadruped business is growing too fast, and if it remains under the halo of humanoid robots, it will be overshadowed. The Group’s judgment is that quadruped itself is a giant and should run independently. Qiu Heng repeatedly emphasized a core concept: Zhiyuan Kuto is not making robot dogs, but quadruped robot “humans”. The difference lies in AGI. His words were impactful: the robot dogs on the market “all have no brains”, “when it comes to quadruped robots, there is simply no competition”. This confidence comes from Zhiyuan Group’s technology reuse. The brain and cerebellum are almost fully shared, training data is highly universal, and using Genie Studio to train quadruped work models reduces workload by 70% compared to starting from scratch. Financial targets are equally aggressive. Revenue target for 2026 is 500 million RMB, aiming for 10 billion by 2030, with overseas accounting for 35%. The first quarter has already surpassed last year’s total, with medium-sized quadruped robots sold out of inventory; for the partner conference on April 17, sample machines even needed to be recalled. Overseas orders accounted for more than 40% in Q1. But Qiu Heng also drew clear boundaries. There will be no fundraising for now, “the boss said this market is very, very big, worth a lot of money, don’t sell it off just yet”. There will be no ToC for now, focusing first on industry scenarios like power grid, mines, security, and logistics, “what we want to make is a productivity tool”. Regarding competition with Unitree and Yunshenchu, Qiu Heng's response was thought-provoking. He said there is “no intense competition” currently in the market; customers are buying intelligent capabilities, “if someone buys a dumb dog, they won't want a second one”. In terms of supply chain costs, he noted not to focus only on quadruped shipments—humanoid and quadruped share the same supply chain, and the total number of joints combined is the true scale. The soon-to-be-launched D2 Max will be Zhiyuan Kuto’s key differentiated product, claimed to be the first all-terrain L3 intelligent driving quadruped robot, capable of navigating with open eyes, autonomous navigation, and generalizing multi-tasking in single industries. Qiu Heng predicts that 2026 will be the explosive point for quadruped robots. Below is an edited transcript of the conversation: On the logic of independence Qiu Heng: Zhiyuan Kuto is the most important subsidiary of Zhiyuan—unlike the previous hand, data, or leasing types. It’s spun off from the main product line. In the future, humanoid robots will be done by Zhiyuan, quadruped robots by Kuto. President Deng has expectations for us; Kuto will certainly become very, very big. On fundraising and IPO Qiu Heng: No fundraising will be done for now, because we don’t want to dilute shares. The boss said this market is very big, worth a lot of money, and don’t sell now. Will we go public? We look forward to that day. On competition Reporter: Unitree’s shipment volume is already very big; Yunshenchu is also big. How to deal with cost advantage? Qiu Heng: When comparing, you can’t only look at quadrupeds, because humanoids and quadrupeds share the supply chain. One humanoid has many joints, so does a quadruped; combine them to see who’s bigger. This is only Kuto’s first year; we haven’t said we want to be number one right away, but our speed to number one will be faster than any others. Reporter: Will prices decrease? Qiu Heng: Cost and price will surely decrease quickly. But at least for now, there’s no fierce competition—the customer mostly looks at whether you can serve them well, not whether the price is a little higher. If someone buys a dumb dog, they won’t want another one. On the core difference of intelligence Qiu Heng: All existing robot dogs are blind—they use reinforcement learning to adjust gait only when they step in a hole. Normally, seeing a pit, a person either detours or jumps, won’t step inside. What Kuto wants to do is to look ahead with open eyes, pick a route, then walk. Reporter: Is intelligence a true demand or a false one? Qiu Heng: If someone doesn’t need intelligence, the only reason is they think it’s unattainable or they can’t afford it. It's like asking someone if they'd like to live in a 1,000-square-meter house—everyone wants one, but it's expensive. Our intelligence is about releasing people’s potential demand. On L3 intelligent driving Reporter: Car L3 is just starting trial; how do you guarantee quadruped robot L3 safety? Qiu Heng: Quadruped robots have an easier part—the activity range is limited, so intelligent driving happens in a defined range. The hard part is all terrain, as cars don’t go into hills and ditches, but quadrupeds can. We call it 3D intelligent driving; cars are relatively planar. On shipments and industry penetration Qiu Heng: Patrol, inspection, emergency, fire-fighting, education, and entertainment—all six scenarios have shipments now, but all are in very early stages. Market penetration is not even at 1%. This first quarter matched last year’s total volume; Q2 is expected to exceed the whole year. Reporter: When is the explosion point? Qiu Heng: 2026 is already an explosion point. Quadruped robots may have a faster closed-loop than humanoids since they are simpler. On ToC Qiu Heng: No plans for ToC yet. First ToB, then ToC, depending on tech maturity. Industry environments are relatively defined and easier for closed-loop; home environments are diverse, requiring more complex intelligence. What we want to make are productivity tools—after we tackle power grids, mines, security, logistics, then we’ll see if we need to go into homes. On data reuse Reporter: Zhiyuan has over 3 million real-world materials—humanoid and quadruped are different forms, how to reuse? Qiu Heng: Training is for the same brain. Humanoid robots aren’t exactly the same—Jingling, Yuanzheng, and Lingxi aren’t identical, but they all gather data, training the same brain. The difference isn’t that big, only a little time needed, not necessary to start over. On cost reduction Qiu Heng: Shipments should multiply in the coming years; a 20-30% cost reduction is normal. Cost reduction mainly comes from volume. On 6G and communication Qiu Heng: 6G won’t be standardized for another 9 years—we hope to be among the first users when it is in 2029. The integrated sensing and communication in 6G is especially important to us—robots need radar, ranging, sensing, terrain scanning, and doing communication while scanning. On future forms Reporter: Why not make six-legged robots? Qiu Heng: The future can be six legs, eight legs, but first we need to excel at quadrupeds. Most animals in the world are quadrupeds, indicating this form is highly adaptive in nature. When the intelligent brain is done, even a hundred legs is possible. The key is to get the brain ready—once it’s made, whatever it’s installed in becomes a certain “person”. This article comes from [WeChat Official Account "Hard AI"](https://mp.weixin.qq.com/s/2rfpVmedXuAVybneRhpkIg). For more AI frontier news, [click here](https://mp.weixin.qq.com/s/2rfpVmedXuAVybneRhpkIg). Risk warning and disclaimer: The market carries risks, and investment should be cautious. This article does not constitute personal investment advice, nor does it take into account the individual user's specific investment goals, financial situation, or needs. Users should consider whether any opinion, view, or conclusion in this article fits their particular situation. Investing accordingly is at your own risk.