Chinese Professor Takes Underwater Robotics From Lab to Market With a Bet on Quiet Propulsion

NextFin News -- For more than two decades, Guo Chunyu worked at the frontier of marine propulsion, developing technologies involving oscillating fins, pump-jets and underwater robotics at Harbin Engineering University.

When he began meeting investors after deciding to start a company, however, the conversations often stalled.

Guo could explain the technology. He could point to papers, patents and national-level research projects. But the investors sitting across the table wanted answers to a more basic question: Could it actually be sold?

That question exposed what Guo calls the “wall” between university research and commercial products.

Late one night in December 2025, that wall began to look more like a bridge.

Guo came across a video by Chen Haibin, initiator of the Beijing Scientists Entrepreneurship Group, about scientists becoming entrepreneurs. The suggestion that scientists could make successful founders caught his attention. Guo left a comment on WeChat Channels, and within minutes Chen accepted his friend request.

It was after 11 p.m. Chen told him there was no reason to wait until the next day. They spoke by phone that night and agreed to meet in Qingdao.

A week later, Chen visited Gongzhi Ocean, the company Guo had founded. He listened to the team's technical roadmap—combining oscillating fins with pump-jet propulsion—and watched a prototype move through a test pool without a conventional propeller.

The machine moved quietly, almost like a fish gliding close to the bottom.

Chen did not focus on asking whether the technology could be commercialized. Instead, he saw something else: a research team at Harbin Engineering University with decades of accumulated expertise, a professor who had spent much of his career working on marine propulsion, and a propulsion technology that differed fundamentally from the conventional approaches used in underwater robots.

After returning to Beijing, Chen decided to bring the Beijing Scientists Entrepreneurship Group into Gongzhi Ocean, providing resources and incubation support.

Over the following months, investors began arriving.

Gongzhi Ocean recently announced an angel financing round of nearly 100 million yuan, with investors including Yinfeng Capital, Leaguer Financial, China-Singapore Smart Land, Small Is Big Ventures and Harbin Engineering Asset Management.

For investors, the appeal lies not simply in a professor's academic credentials. It is the combination of technical depth, a differentiated product and early evidence that customers will pay.

As of August 2026, Gongzhi Ocean said its intended orders had exceeded 160 million yuan, spanning applications from polar-vessel propulsion and nuclear-radiation detection to reservoir dams and overseas customers.

The company is betting that underwater robotics is entering an inflection point, as commercial demand, technological advances and China's push to become a maritime power converge.

But Gongzhi Ocean's larger ambition is more audacious. Guo hopes that it can eventually become a “Huawei of the deep sea.”

The company is still a tiny fraction of the scale of Huawei, and it faces the difficult transition from laboratory technology to mass production and global sales. Yet its early orders are beginning to provide a test of whether its technological bet can translate into a business.

Breaking Through the Wall Between Research and Business

Guo, born in Bengbu, Anhui, in 1981, entered Harbin Engineering University in 1999 and completed his undergraduate, master's and doctoral studies there. He later conducted postdoctoral research at Shanghai Jiao Tong University and spent time as a visiting researcher at University College London before returning to Harbin Engineering University as a doctoral supervisor.

Counting from his graduate studies in 2003, he has spent about 23 years immersed in marine propulsion.

His academic credentials are formidable. He is a recipient of the National Science Fund for Distinguished Young Scholars, served as chief scientist of the National Defense Basic Strengthening Program and won the First Prize for National Defense Science and Technology Progress. He was also named among the world's top 2% of scientists in 2024 and 2025.

More important for Gongzhi Ocean, his career was built inside one of China's longstanding marine-engineering research groups.

His adviser, Professor Huang Sheng, graduated from the former Harbin Military Engineering Institute and belonged to an earlier generation of Chinese researchers working on underwater propulsion and aircraft-carrier-related technologies.

The expertise accumulated across generations eventually became one of the university's leading research capabilities in underwater propulsion, sensing and intelligence.

Guo recalls Huang telling a story about the early development of China's aircraft carriers. Chinese researchers discovered that foreign carriers used four propellers with identical blades, although their rotation directions differed. At first, the arrangement seemed counterintuitive. Later, during development, the Chinese team encountered a power mismatch caused by factors it had initially overlooked and had to work through the problem from first principles.

The lesson stayed with Guo.

“You can’t buy this kind of technology, and no one will hand it to you,” he said. His generation had to build it from nothing.

But the professor eventually encountered a different problem.

As the head of a laboratory, he saw expensive prototypes—some costing tens of millions of yuan—sit unused for years.

Guo began asking why.

If researchers were afraid that equipment would be damaged or lost, he reasoned, perhaps the technology had not yet been sufficiently engineered for real-world use.

“It’s wrong for lab technologies to get published in papers, win awards, and then be put on the shelf,” he said.

A prototype that operates for 100 hours in a controlled test pool is not necessarily ready for 1,000 hours in the ocean. Saltwater, pressure, low temperatures and unpredictable currents create a very different engineering environment.

Guo believed the timing finally became right in 2024.

“The technology had matured, the policy window had opened, and market demand had arrived as well,” he said. “If we didn’t do it at that moment, we might miss it.”

He registered Gongzhi Ocean that August.

Soon afterward, someone addressed him as “President Guo.” He was briefly stunned. But the title captured the change in his role. One side of his job still demanded scientific truth; the other demanded market proof. Papers and patents were no longer enough. Customers wanted orders, deliveries and revenue.

He had crossed the wall.

The transition was difficult. Guo recalls that the company's business logic was initially rough and that investors often struggled to understand the technology. There was even a saying circulating online that investors should avoid startups founded by scientists.

The broader market, however, was changing.

ICV Tank projected China's underwater-robotics market to grow at a compound annual rate of about 46.93% from 2025 through 2030. In March 2026, China's outline for the 15th Five-Year Plan was approved, calling for accelerated development of the country as a maritime power.

From the second half of 2025, some investment firms began forming dedicated ocean-technology teams, while investors previously focused on artificial intelligence and robotics began paying greater attention to underwater applications. According to incomplete statistics cited in the company's materials, disclosed financing in China's underwater-robotics sector in 2025 exceeded the combined total of the previous three years.

The investment thesis is increasingly centered on one problem: underwater robots need to move without interfering with their ability to perceive their surroundings.

That is where Gongzhi Ocean believes its technology has an advantage.

A Robot Designed to Move Like a Fish

The company's flagship underwater robot, called Hetu, looks unlike most commercial underwater robots.

There is no conventional propeller.

Instead, flexible fins undulate slowly, generating thrust in a motion inspired by fish. The result is a platform designed to operate quietly and close to the seabed without generating the turbulence associated with high-speed propellers.

The company says the technology has been under development for about 18 years, evolving from a research prototype into a product it says is ready for mass production.

The name Hetu has multiple meanings. Its pronunciation is one letter away from HEU, the abbreviation for Harbin Engineering University. In Chinese tradition, the Hetu is associated with the origins of yin-yang, the Five Elements and numerological systems.

For Gongzhi Ocean, the name reflects the robot's intended ability to detect patterns in chaotic underwater environments, make autonomous decisions and navigate without constant human control.

The technology starts with a fundamental problem.

Underwater robots have difficulty both seeing and hearing.

Light is quickly absorbed by water, and visibility can fall to roughly a meter in turbid conditions. In rivers, suspended sediment can reduce visibility even further. A robot that stirs up the seabed can make the problem worse by generating its own cloud of sediment.

Then there is sonar.

Sonar is essential to underwater robots because electromagnetic signals cannot travel through water in the same way they do through air. But sonar is highly sensitive to noise generated by the robot itself.

That creates a fundamental conflict between propulsion and perception.

Traditional propellers have been used for more than a century, but Gongzhi Ocean argues that they have reached the limits of what incremental optimization can achieve for intelligent underwater platforms. The company says conventional propellers can generate radiated noise of roughly 100 to 170 decibels. High-speed blade rotation can create cavitation, while collapsing bubbles generate noise and erode blades. Propellers can also disturb sediment on the seabed.

The result is that the machine's own propulsion system can interfere with its sensors.

“Propulsion and sensing are naturally at odds,” Yinfeng Capital said in its assessment of Gongzhi Ocean. Reducing noise, in its view, is not merely a matter of stealth. It can improve the signal-to-noise ratio needed for underwater multimodal perception.

Previous biomimetic approaches have not completely solved the problem.

Fish-like tail propulsion can generate speed but may cause the robot's body to sway, potentially disrupting sonar and magnetometers. Manta-ray-inspired systems can operate quietly and efficiently but may have difficulty performing complex tasks.

Gongzhi Ocean's approach combines two propulsion systems.

Its flexible oscillating fin provides gentle, quiet thrust for low-speed and stationary operations, while a submarine-style pump-jet provides higher cruising speeds and directional control.

The company describes the combination as one quiet propulsion system and one dynamic system on the same platform.

The goal is to allow one robot to both explore and operate, rather than relying on one autonomous underwater vehicle for reconnaissance and a separate remotely operated vehicle for physical work.

Gongzhi Ocean says Hetu's radiated noise can be reduced to 86 decibels, sediment disturbance to one five-hundredth that of conventional solutions and propulsion efficiency above 90%. It can also pivot with a zero turning radius.

But the technology brings its own engineering difficulties.

The flexible fin changes shape as water speed, direction, depth and temperature change. Traditional rigid-body mechanics cannot fully describe how a flexible surface deforms in an unsteady flow.

Gongzhi Ocean says it has developed a hyperbolic-surface wave dynamic equation based on fluid mechanics and combined it with real-time closed-loop attitude control, allowing the propulsion system to adapt dynamically to changing water conditions.

The pump-jet is another complex engineering problem.

Its design includes a duct, stator and multi-blade rotor. According to the company's investors, the number of design variables can rise from a dozen or so for conventional propellers to hundreds, while interactions among components become exponentially more complicated.

That accumulated expertise is part of what investors describe as the company's technological moat.

From Deep-Sea Data to a Commercial Flywheel

The company is also building the software and sensing systems around its propulsion technology.

Gongzhi Ocean says its six core underwater operating systems—power, control, sensing, communications and navigation, waterproofing, and intelligence—are developed in-house, with a 100% localization rate.

Its self-developed Luoshu research-and-development software is designed to shorten design iterations that once took months to 48 hours.

The company says it has more than 200 terabytes of annotated deep-sea data, recognition accuracy above 95%, onboard computing power of 200 teraflops, a 42% improvement in perception-distortion correction and a 25% increase in perception range.

The more important asset, however, may be the data generated after deployment.

Traditional underwater-robot systems typically train algorithms offline and then deploy them. Over time, the model can become less representative of actual operating conditions.

Gongzhi Ocean is attempting to build a feedback loop.

Its initial models are trained using years of accumulated research data. Once a robot is deployed, it collects navigation, terrain and target information at millisecond-level frequencies. After the mission, that data is fed back into the training system. The model is retrained and upgraded before the next deployment.

The objective is a self-reinforcing cycle: more operating miles generate more data; more data improves the model; improved models make subsequent missions more accurate and stable.

That could eventually create a barrier for competitors, because newcomers would have to reproduce not just the hardware but years of real-world sea-trial data.

Guo is more cautious than some investors about the current enthusiasm around underwater “embodied intelligence” foundation models.

Underwater robots cannot rely on the same cloud infrastructure as land-based machines. Electromagnetic signals attenuate rapidly in water, while underwater acoustic communications can have bandwidths of only a few dozen kilobits per second.

A robot therefore has to carry much of its own intelligence.

“Every underwater robot has to carry its brain and cerebellum inside the body and run autonomously,” Guo said.

For now, he believes the priority should be autonomous navigation, target recognition and operational control rather than large underwater foundation models.

That pragmatism is reflected in the company's commercial strategy.

Gongzhi Ocean sees the domestic industry as broadly divided into three groups: traditional integrators using propeller-based systems for large ROVs and high-power applications; consumer and light-duty companies serving shallower waters; and companies attempting to innovate at the level of propulsion principles and system architecture.

Gongzhi Ocean is pursuing the third path.

Orders Put the Technology to the Test

The central question for a hard-tech startup remains the same one Guo faced in his first investor meetings: Who will pay?

Some investors still see underwater robotics as a niche market characterized by long development cycles and heavy technology requirements.

Wang Xueyang, investment director at the Beijing Scientists Entrepreneurship Group, disagrees.

After studying more than a dozen underwater-robotics companies, Wang argues that demand isn't necessarily absent. It is simply hidden.

As offshore wind farms, reservoirs and other underwater infrastructure age, operators increasingly need inspection and maintenance technologies that can reduce labor and operating costs.

The commercial logic is shifting from one-off scientific projects to return-on-investment calculations.

“If your equipment can help them save money, time and labor, or replace people getting into the water to do the work, the orders will follow,” Wang said.

Gongzhi Ocean's reservoir-inspection project provides one example.

In the second half of 2025, the company took on a reservoir inspection project. The conventional approach involved renting a large ROV and support vessel at a cost of about 5 million yuan for a single operation, with a six-month timeline.

Hetu completed the same job in 20 days for about 700,000 yuan, according to the company, cutting costs by 86% and increasing efficiency by 90%.

China has roughly 94,000 reservoirs. Even if only 1% required annual inspection, that would represent more than 900 inspections a year.

The reservoir project also provided an important engineering test. Hetu briefly encountered sensor-data drift near the reservoir bottom during its first days in the water. The team worked through the night to adjust the algorithms.

“Every delivery is an iteration,” Guo said.

For a startup, that learning process is important because customers do not lower their standards simply because the company is young or founded by a professor.

The polar market has provided another test.

Gongzhi Ocean says it secured almost all domestic propulsion-system orders for high-ice-class vessels in 2025. Its polar propulsion system uses stainless steel and rare-metal alloys rather than conventional nickel-aluminum bronze and is designed to withstand collisions with ice 3 to 4 meters thick while meeting the PC1 ice-class standard established by the International Association of Classification Societies.

A Canadian deep-sea mining company provided perhaps the clearest example of technology turning into commercial trust.

The company initially approached Gongzhi Ocean for consulting work to assess whether a particular technical solution was feasible. Technical discussions continued online for about a month and a half, with the Chinese team answering questions in detail.

After due diligence confirmed that Gongzhi Ocean was founded by university professors and that its technology was genuine and advanced, the consulting engagement was upgraded into an exclusive supply agreement for core components worth tens of millions of yuan.

The company, less than two years old at the time, had won a contract in an overseas subsea-mining market long dominated by established European and American offshore-engineering companies.

“At the end of the day, technology is what you build your footing on,” Guo said.

The company is now moving into another closely watched application: hull cleaning.

Marine biofouling increases drag and fuel consumption, while regulations governing underwater hull cleaning are becoming more stringent. Yet propellers and the curved areas around ship bows and sterns remain difficult for existing suction-based cleaning robots to reach.

Gongzhi Ocean has attached a robotic arm to a Hetu platform and plans to open public pre-orders in mid-September. The robot is designed to hover with centimeter-level precision while autonomously performing cleaning, grabbing and cutting operations in a highly disturbed underwater environment.

The broader commercial strategy remains deliberately staged.

Polar propellers, 3D-printed propellers and related products are intended to generate cash flow today.

Reservoir inspection and ship cleaning are flagship applications designed to validate the technology and open markets.

Deep-sea resource development and seabed mining are longer-term strategic positions.

That sequencing may be one of the company's most important differences from a typical university spinout.

Gongzhi Ocean's chief executive, Li Honglei, brings more than 20 years of industry and capital-management experience. Other senior members oversee intelligent algorithms, hydrodynamic testing, intelligent control and engineering implementation.

The structure is designed to address a common weakness of professor-led startups: strong technology but insufficient commercial and operational capabilities.

Yinfeng Capital described Gongzhi Ocean not simply as a professor starting a company, but as an institutionalized team with deep technical accumulation and dedicated commercialization leadership.

China-Singapore Group, meanwhile, has spent the past year mapping more than 300 companies across the embodied-intelligence supply chain and established a 250 million yuan fund focused on the sector.

Gongzhi Ocean was selected because its investors saw a niche market with clear pain points, deep technical expertise and potential synergies with industrial resources.

The company now has a triangular operating structure: its Suzhou headquarters focuses on engineering and development, a Nantong subsidiary handles manufacturing and a Sanya subsidiary conducts sea trials.

The new financing will primarily support mass production, product iteration, research and development, order delivery, market expansion and expansion of the core team.

For an angel-stage company, having too many orders rather than too few may be an unusual problem.

But it is also a more concrete problem than the one Guo faced when investors struggled to understand his technology.

Can It Become the “Huawei of the Deep Sea”?

The “Gongzhi” in Gongzhi Ocean comes from Harbin Engineering University's motto: “Great engineering strives for the utmost good; great universities pursue the utmost truth.”

For Guo, the phrase represents the transition he has spent years trying to make—from discovering scientific truths to turning those truths into engineering systems that work in the real world.

When he discusses oscillating-fin modeling, the hundreds of variables involved in pump-jet design or the transient forces generated when ice strikes a propulsion blade, his pace quickens. He gestures with his hands as if tracing water flowing around a submerged machine.

Then the conversation turns to the future.

Over the next five years, Guo hopes Gongzhi Ocean can become a strategic cornerstone in China's marine sector.

He also repeats his more ambitious goal: to become a “Huawei of the deep sea.”

The comparison is, by any measure, audacious. Gongzhi Ocean remains an early-stage company, and it has yet to prove that it can manage large-scale manufacturing, supply chains and global sales.

Those are precisely the areas in which professor-turned-entrepreneurs are often considered weakest.

But the company's ambition is no longer based solely on papers and patents.

There is a robot that can glide quietly through a test pool. There are four decades of accumulated marine-engineering expertise behind its core research group. There is a propulsion architecture designed around the fundamental conflict between movement and perception. There are commercial orders spanning civilian and defense-related applications. And there is a team combining scientific research, engineering and business experience.

Each of those provides a measurable test of the company's thesis.

Guo spent more than 20 years building knowledge before deciding that the technology needed to leave the laboratory.

Now the company has to prove that knowledge can survive the much harsher waters of the market.

Can Gongzhi Ocean become the “Huawei of the deep sea”?

For now, the answer remains underwater.