Over $200 Million in Funding in a Month: Why Has QiuZhi Technology Become the New Darling of Embodied Intelligence Investors?

On August 3, 2026, QiuZhi Technology—which had been in business for just over three years—once again found itself in the spotlight of the embodied intelligence industry.

这家从清华大学智能产业研究院孵化而来的mechanical person创业公司宣布完成1亿美元天使+轮financing. Industrial and financial investment institutions such as IDG Capital, Xinglian Capital, Wuyuefeng Sci-Tech Innovation, Dacheng Capital, Jiuyang Venture Capital, Huaying Capital, and Wuxi Binhu District Industrial Group participated in this round of financing, and existing shareholders also made follow-on investments.

What’s even more noteworthy is that this comes less than a month after QiuZhi Technology announced the completion of an angel round of financing exceeding $100 million.

In late June 2026, QiuZhi Technology secured a joint investment from institutions including Legend Capital, Xincheng Investment, Lenovo Ventures, Bai Capital, Puhua Capital, Lingang Sci-Tech Venture Capital, Amber Capital, and Gengxin Capital. This funding round has been described by the company and some of the investors as the largest single angel round in China’s consumer-grade embodied robotics sector.

With cumulative funding from two rounds totaling more than $200 million, QiuZhi Technology has made the leap from an early-stage research-oriented startup to a star project in the capital markets in a very short period of time.

But the real question we should be asking isn’t why a robotics company was able to raise so much money, but rather: What exactly does capital see in Qiushi Technology?

The answer may not be just a single robot, but rather an embodied intelligence infrastructure that seeks to connect models, data, simulations, robot hardware, and consumer scenarios.

I. Behind the Two Rounds of Funding Lies an “Early Bet”

In terms of its funding history, QiuZhi Technology is still a relatively young company.

The company was incubated by Tsinghua University’s Institute for Intelligent Industry in 2023 and initially focused on providing the AIRBOT series of products to universities, research institutions, and robotics developers. According to publicly available information from Tsinghua University’s Institute for Intelligent Industry, QiuZhi Technology’s vision is “to bring robots into every household,” and its technology and team are closely tied to the institute’s Collaborative Vision and Robotics Laboratory (DISCOVER Lab).

Typically, hardware startups must go through several stages—including prototype validation, small-batch production, market validation, and mass production—before they can secure funding in the hundreds of millions of dollars.

QiuZhi Technology, whose products are still in the transition phase from the research and development market to the consumer market, has successfully completed two consecutive rounds of funding, each worth $100 million. This means that investors are not merely buying into the company’s current revenue, but are investing based on their expectations for its future technology platform, product forms, and market access.

The complexity of these expectations is also evident from the list of investments.

Firms such as Legend Capital, IDG, Dachen, and Huaying are typically characterized by their focus on financial investments; Lenovo Ventures and Joyoung, on the other hand, possess resources in consumer electronics, supply chains, manufacturing, and distribution channels; while Xincheng Investment, Lingang Science and Technology Venture Capital, and the Binhu District Industrial Group are backed by industrial policies, production bases, and regional ecosystem development initiatives.

In other words, these two rounds of financing are not merely an injection of capital; rather, they are akin to assembling an industrial network in advance to prepare for future mass production: research institutions provide the technological foundation, venture capital assumes the risks of long-term R&D, industrial capital assists in building supply chains and distribution channels, and local state-owned enterprises provide production space, policy support, and industrial infrastructure.

For a robotics company planning to enter the home and consumer markets, this kind of partnership is often more important than simply securing financial investment.

II. A Robotics Company Born in a Tsinghua University Lab

The origins of QiuZhi Technology’s technology can be traced back to the DISCOVER Lab, which is part of Tsinghua University’s Institute for Intelligent Industry.

Zhou Guyue, Chief Scientist at QiuZhi Technology, is an associate professor at Tsinghua University’s Institute for Intelligent Industry and the head of the Collaborative Vision and Robotics Laboratory. According to information disclosed by the Tsinghua University Institute of Intelligent Industry, Zhou Guyue has long been engaged in research on intelligent drones, autonomous driving, computer vision, and intelligent robots. He has participated in the research and development of multiple civilian drone and educational robot products, published more than 80 papers in the fields of robotics and computer vision, and holds dozens of domestic and international invention patents.

This experience explains the difference between QiuZhi Technology and many startups that “started in mechanical manufacturing and moved into robotics.”

Rather than first building a robotic body and then searching for artificial intelligence algorithms, the approach starts with visual perception, robotic learning, and intelligent control, and gradually extends to the robot itself and product engineering.

This has also become QiuZhi Technology’s most distinctive hallmark: the combination of an academic technical background and experience in consumer electronics engineering.

The company’s academic background enabled it to get an early start in embodied foundational models, simulation-based training, and robotic learning; its product engineering experience, in turn, has led the team to place greater emphasis on cost, reliability, the supply chain, development interfaces, and mass production.

In the embodied intelligence industry, it is often difficult to possess both of these capabilities at the same time.

Some research teams have advanced algorithms but lack the capabilities for mass-producing hardware and defining products; some traditional robotics companies can reliably manufacture equipment but struggle to build data-driven, general-purpose intelligent systems. QiuZhi Technology aims to occupy the intersection between these two areas.

III. Don’t rush to create a human figure; start by making “moving arms” instead.”

Currently, the product form attracting the most attention in the embodied intelligence industry is the humanoid robot. Bipedal locomotion, human-like appearance, and full-body control have a strong visual impact and are more likely to attract capital and public attention.

QiuZhi Technology, however, has taken a relatively restrained approach.

Its flagship product, the AIRBOT MMK2, features a mobile chassis, a lifting mechanism, and dualRobotic ArmA modular system, rather than a complete bipedal humanoid structure. According to official technical documentation, the MMK2 is equipped with two lightweight robotic arms with six degrees of freedom, multiple vision sensors, a LiDAR, a mobile chassis, and a lifting mechanism. It can be used for tasks such as data collection using both arms, algorithm deployment, flexible manufacturing, and household chores.

This approach may not seem as “sexy” as humanoid robots, but it is more pragmatic from an engineering standpoint.

For many tasks in home, laboratory, food service, retail, and light industrial settings, what’s truly needed is mobility and operational capability—not necessarily the ability to move on two legs. Wheeled platforms typically offer better stability, longer operating times, and easier control, and they’re also more cost-effective.

The lifting mechanism can adjust to different heights, such as tabletops, counters, shelves, and the floor; the dual robotic arms can perform tasks such as handling, organizing, delivering, opening doors, retrieving items, and collaborative operations.

Therefore, the mobile dual-arm platform is essentially a product that falls somewhere between industrial robotic arms, service robots, and humanoid robots.

Although it sacrifices some of its human-like appearance, it is able to enter real-world environments more quickly.

From a business perspective, this decision also helped QiuZhi Technology avoid direct competition in the bipedal humanoid robot market. Faced with numerous competitors such as Yushu, Zhiyuan, UBTech, Galaxy General, and Stardust Intelligence, directly entering the competition for full-scale humanoid robot platforms would not only require massive capital investment but would also quickly lead to a “red ocean” of competition in areas such as actuators, joints, batteries, motion control, and cost management.

QiuZhi Technology prefers to first use a mobile operating platform to accumulate data, train algorithms, and validate user experience before deciding on the final form of its future consumer-grade products.

IV. AIRBOT is not a final product, but rather a data gateway

QiuZhi Technology’s first product, AIRBOT Play, is a lightweight robotic arm designed for education, scientific research, and robotics algorithm development.

At first glance, the market for this type of research-grade robotic arm appears limited and unlikely to support a valuation in the billions of dollars. However, from the perspective of embodied intelligence, AIRBOT’s significance extends far beyond simply selling hardware.

Basic robot models require a large amount of real-world operational data.

Large language models can access vast amounts of text from the internet, but robots cannot learn how to unscrew bottle caps, fold clothes, or tidy up a desk directly from web pages. Robots must acquire multimodal data, including visual information, joint states, force feedback, motion trajectories, and task outcomes.

This presents a fundamental challenge for embodied intelligence: high-quality real-world data is scarce, and the cost of collecting it is extremely high.

QiuZhi Technology’s AIRBOT robotic arm, MMK2 dual-arm platform, and teleoperation and teaching systems essentially constitute a data acquisition terminal. When developers and researchers use this equipment to conduct experiments and perform tasks, they can also generate standardized robotic operation data.

According to the company’s official documentation, its products support ROS, ROS2, Python, C++, and Web APIs, and are compatible with open-source embodied intelligence models and algorithm ecosystems such as ALOHA, Mobile ALOHA, RT-X, VoxPoser, RDT, and π0.

The significance of open interfaces is that AIRBOT is not merely a closed-off device, but can serve as a hardware platform for different algorithms, models, and data formats.

If the scale of deployment is large enough, Quzhi Technology could create a cycle in which:

Sell robot hardware to expand the installed base; the devices are used for scientific research and scenario development, continuously generating data; the data is used to train and improve foundational models; as the models become more capable, this in turn enhances the value of the robot products.

This bears some resemblance to the cycle of “vehicle sales—real-world road data—autonomous driving model upgrades” in the smart car industry.

What truly determines QiuZhi Technology’s long-term value may not be how many robotic arms it sells, but whether it can transform disparate robotic devices into a network that continuously generates production data.

5. “Model—Data—Simulation” Is the Core of the Pitch

In its latest funding announcement, QiuZhi Technology highlighted three key technology modules: ORION, a layered embodied intelligence foundation model; LOOP, a closed-loop system for virtual-physical hybrid data; and DISCOVERSE, a high-fidelity simulator.

These three correspond to the three key issues facing embodied intelligence.

ORION aims to address how robots can understand tasks, break them down, and execute actions; LOOP is responsible for integrating real-world data, teleoperation data, synthetic data, and training workflows; and DISCOVERSE enables low-cost, large-scale training of robots in virtual environments.

Investment firms describe it as a full-stack technological foundation comprising “models, data, and simulation.”

This combination of technologies is not merely a collection of concepts thrown together.

There are currently two main approaches to foundational models for embodied intelligence: one is the end-to-end model, in which the model directly generates robot actions based on visual and language instructions; the other is a layered architecture, in which higher-level models handle understanding, planning, and reasoning, while lower-level models handle motion control and skill execution.

Household chores are often time-consuming, involve many steps, and are fraught with uncertainty.

For example, asking a robot to “clear the table” may involve multiple subtasks, such as recognizing objects, identifying trash, picking up utensils, moving to the kitchen, opening the dishwasher, and placing items inside. If we rely entirely on a single end-to-end model to directly generate all actions, not only would the training data requirements be enormous, but it would also be difficult to pinpoint errors.

A layered architecture breaks down complex tasks into a number of reusable skills, invoking lower-level action modules through high-level planning.

In its *Report on the Development of Embodied Intelligence (2025)*, the China Academy of Information and Communications Technology (CAICT) noted that the algorithmic approaches for embodied intelligence have not yet converged, and that end-to-end “unified” models and hierarchical collaborative models continue to develop in parallel; at the same time, data, models, and robot hardware have become the three key areas of competition within the industry.

QiuZhi Technology’s choice of a layered model actually reflects an engineering judgment: at this stage, robots are still far from achieving truly general-purpose end-to-end intelligence, and separating task planning from action execution may make it easier to achieve stable deployment.

VI. DISCOVERSE: Why Are Simulators So Important?

DISCOVERSE is an often-overlooked component of QiuZhi Technology’s technical framework.

Training robots in the real world is extremely costly.

Robots may collide with objects, fall, or damage them; collecting real-world data requires a testing site, equipment, and human intervention; and there are significant variations in lighting, materials, object locations, and friction across different environments.

Simulators can run tens of thousands of robot instances in parallel within a virtual environment, allowing models to repeatedly attempt different tasks. Failures do not damage equipment, and the environment and objects can be changed quickly.

In its product documentation, QiuZhi Technology describes DISCOVERSE as a Real2Sim2Real solution, which involves first mapping real robots and scenarios to a simulation environment, training them in the virtual world, and then deploying the models back to the real robots.

However, simulation-based training faces a classic problem: the simulated world does not fully correspond to the real world.

Subtle differences such as lighting, friction, object weight, sensor noise, and mechanical errors can all cause a model that performs well in simulation to fail in a real-world environment. This is known as the “Sim-to-Real” gap.

Therefore, the value of a simulator depends not only on the realism of the visuals, but also on whether the physical laws, sensor models, robot dynamics, and real-world equipment are accurately aligned.

QiuZhi Technology develops robotic arms, mobile platforms, control systems, and simulators simultaneously, which theoretically enables stronger hardware-software coordination: data generated by real robots is used to calibrate the simulation environment, and models developed through simulation training are then fed back into the same hardware for validation.

This also constitutes a key competitive advantage for its full-stack approach.

VII. Why Capital Is Placing Concentrated Bets in 2026

QiuZhi Technology's funding round is not an isolated incident.

Since 2025, investment in embodied intelligence has picked up significantly. According to the China Academy of Information and Communications Technology (CAICT), as of December 2025, there had been 744 investment deals in China’s embodied intelligence and robotics sectors, with total funding reaching approximately 73.5 billion yuan. However, the academy also cautioned that the commercialization of the industry remains in the early exploratory stages, and that the technical maturity and actual delivery capabilities should be assessed objectively.

As we enter 2026, the focus of the capital markets has shifted from “whether embodied intelligence is viable” to “which company will be the first to achieve large-scale deployment.”

This is also one of the reasons why QiuZhi Technology has attracted attention.

On the one hand, the company possesses algorithms and research capabilities from Tsinghua University’s laboratories; on the other hand, it has already launched robotic arms, dual-arm mobile platforms, data acquisition systems, and simulation software, rather than remaining merely at the stage of concept videos or laboratory prototypes.

At the same time, Wuxi and the Yangtze River Delta region boast well-established supply chains for machining, precision components, electric motors, gearboxes, and consumer electronics. By establishing its headquarters in Binhu District, Wuxi, QiuZhi Technology is able to connect with research talent in Beijing, the electronics industry chain in Shenzhen, and manufacturing resources in the Yangtze River Delta.

The participation of local governments and industrial capital in the financing also suggests that QiuZhi Technology may take on another role: that of a system integrator and platform provider within Wuxi’s embodied intelligence industry cluster.

What capital is actually betting on is a complex proposition:

If foundational robot models continue to improve, the cost of core components continues to decline, and demand for home services becomes increasingly clear, then a company that possesses expertise in models, data, simulation, hardware, and the supply chain could be the first to launch a consumer-grade robot that can be sold at scale.

VIII. Consumer Robots: More Challenging Than Industrial Robots

QiuZhi Technology is focusing on consumer-grade embodied intelligence and home service robots. This market offers enormous potential, but the challenges are far greater than those in structured industrial settings.

Factory environments typically feature fixed workstations, standardized materials, well-defined processes, and safety barriers. Robots can be repeatedly trained to perform a single task.

It's completely different in a family setting.

Every household has its own unique spatial layout, furniture heights, types of items, lighting conditions, and lifestyle habits. Robots must not only recognize objects but also understand human intentions and operate safely in environments where children, the elderly, and pets coexist.

More importantly, consumers have very little tolerance for unreliability in robots.

A lab robot with a success rate of 90% may well be of research value, but if a home robot fails once for every ten tasks it performs, users will likely stop using it very quickly.

Home robots must also address issues such as noise, battery life, design, privacy, maintenance, after-sales service, and price. Even if the technology is up to the task, it remains to be seen whether consumers are willing to pay tens of thousands of yuan—or even more—for them.

The China Academy of Information and Communications Technology (CAICT) also pointed out that for embodied intelligence to move from demonstration to practical application, engineering challenges—such as ensuring thousands of hours of stable operation, hardware-software compatibility, security safeguards, and sustained operation in real-world scenarios—still need to be addressed.

Therefore, the biggest challenge facing QiuZhi Technology is not putting on another impressive robot demonstration, but ensuring that the robots operate reliably every day.

IX. Continuous fundraising also means greater pressure

The more than $200 million in funding has provided QiuZhi Technology with ample resources, but capital does not come without a cost.

First is the pressure to ramp up production.

The company needs to establish assembly and testing production lines, optimize its supply chain, reduce product costs, and transform laboratory prototypes into standard products capable of long-term, stable operation. According to information disclosed by local media, the funds raised in the first round of financing will be used for core algorithm research and development, new product development, building assembly and testing capacity on the scale of tens of thousands of units, and expanding overseas channels.

Next is the pressure associated with product definition.

“Bringing robots into every home” is a vision, but not a specific product. Household tasks such as cleaning, organizing, providing companionship, cooking, and delivering items vary greatly, and it would be difficult for a single robot to address all of these issues simultaneously in the early stages.

QiuZhi Technology must answer: What need does its first true consumer-facing product actually address? Is it a household robot, an assistive device for the elderly, a smart mobile device, or a developer platform?

Third is the pressure to close the data loop.

Simply owning robotics hardware does not automatically guarantee high-quality data. The data must be cleaned, labeled, filtered, and standardized, and issues related to user authorization, privacy protection, and the boundaries of commercial use must be addressed.

Fourth is valuation pressure.

In a press release, Lenovo Ventures referred to Qiushi Technology as an “embodied unicorn”; however, as of now, publicly available funding announcements have not disclosed a specific post-investment valuation. Therefore, this designation stems primarily from the investor’s promotional materials, and there is currently a lack of independently verifiable valuation data.

When a company secures substantial funding and a high valuation in its early stages, the market quickly raises its expectations regarding orders, revenue, shipment volumes, and product development progress. If the pace of commercialization falls short of expectations, the advantages gained from funding may turn into operational pressures.

X. Who Are QiuZhi Technology’s Real Competitors?

In terms of product types, QiuZhi Technology’s competitors include companies that develop mobile robots, dual-arm robots, and home service robots.

However, from the perspective of a long-term technological roadmap, its real competition may not be any single robot, but rather the gateway to the era of embodied intelligence.

In the future, the embodied intelligence industry may develop across multiple levels:

The bottom layer consists of motors, gearboxes, sensors, and chips; the middle layer comprises the robot chassis and control system; the top layer includes foundational models, data platforms, simulators, operating systems, and a skills ecosystem; and the outermost layer consists of application scenarios in the home, manufacturing, logistics, elderly care, and commercial services.

QiuZhi Technology is attempting to establish a presence in several of these areas simultaneously.

The advantage of this full-stack approach is that it creates a closed-loop system, reducing reliance on external platforms; the downside is that it requires significant capital investment and involves organizational complexity, and every stage may face competition from specialized players.

In the field of foundational models, it must compete with large-model companies and leading robotics firms; in the field of simulation, it must compete with mature platforms such as NVIDIA Isaac Sim, MuJoCo, and Habitat; in the robotics hardware sector, it must compete with complete-robot manufacturers that hold advantages in manufacturing and cost; and in the consumer market, it must contend with the channel and brand advantages of home appliance, smartphone, and internet companies.

It is difficult for QiuZhi Technology to be the industry leader in every aspect.

A more realistic approach is to achieve a clear competitive advantage on a specific type of robotics platform through the integration of hardware and software, and then build a data and developer ecosystem around that platform.

XI. The Real Story Behind the Funding News

The story of QiuZhi Technology is a microcosm of the current wave of embodied intelligence startups in China.

It combines the technological heritage of Tsinghua University’s laboratories, the manufacturing capabilities of the Yangtze River Delta, the product expertise of its consumer electronics team, and an industry network supported by a multi-tiered capital structure.

However, what makes it particularly noteworthy is its assessment of the sequence for commercializing embodied intelligence.

Instead of focusing first on developing a robot with a highly human-like appearance, QiuZhi Technology started with lightweight robotic arms, data collection, mobile dual-arm platforms, and simulators; rather than merely discussing basic models, the company sought to validate them within real hardware and data loops; and instead of jumping directly into the consumer market, it first served universities, research institutions, and developers.

While this approach does not guarantee success, it is relatively consistent with the patterns of development in the robotics industry.

The key to the robotics industry has never been solely about “whether the AI is advanced enough”; it also depends on whether the hardware is stable, whether costs are manageable, whether data is sustainable, whether the tasks are realistic, and whether users are willing to pay.

The more than $200 million in funding provided QiuZhi Technology with a rare window of opportunity.

Over the next two to three years, it needs to prove three things: that ORION, LOOP, and DISCOVERSE are not just technical concepts confined to promotional materials; that AIRBOT and MMK2 can evolve from research equipment into large-scale platforms; and that consumer-grade robots can deliver significant user value in specific use cases.

If these conditions are met, QiuZhi Technology may become more than just a robot manufacturer; it could evolve into a hardware and software platform for the era of embodied intelligence.

If this proves unachievable, two consecutive rounds of massive funding may turn out to be just another example of capital speculation left over from a period of industry hype.

In the race for embodied intelligence, the outcome will ultimately not be determined by the amount of funding raised.

The real turning point is whether robots can leave the lab and exhibition halls to perform valuable work day in and day out in real-world environments.