In 1854, Cyrus West Field set out to lay a submarine cable across the Atlantic Ocean. After several failed attempts, and with nearly every shareholder calling for the project to be abandoned, the cable was completed in August 1858. A message from Queen Victoria traveled through it to the US, establishing a telegraphic link between the continents.
More than a century and a half later, the world’s submarine fiber optic cables stretch more than 1.4 million kilometers and carry over 95% of intercontinental data traffic. Maintaining that network remains a difficult task.
“Routine maintenance of submarine cables is virtually nonexistent worldwide,” said Guo Chunyu, founder of deep sea robotics company Gongzhi Marine Equipment Technology. “Once a cable breaks, repairs take months, and the cost can run into an eight-figure RMB sum.”
The challenge extends beyond submarine cables. Complex ocean currents, highly corrosive seawater, and the buildup of marine organisms on equipment limit what conventional underwater systems can do.
Guo’s team is working on a solution.
“We’ve spent more than 20 years doing research and have already solved 70–80% of the problems involved in deep sea operations,” Guo said. “Now is the time to turn those technical capabilities into productive capacity as quickly as possible.”
Guo founded Gongzhi Marine in 2025. Its proprietary propulsion system combines flexible undulating fins with vector pump jets. Less than a year after its founding, the company had secured about RMB 100 million (USD 14.9 million) in prospective orders with a team of fewer than 20 people. Its customers work in technically demanding fields, including deep sea mining, natural gas hydrate extraction, submarine fiber optic cable maintenance, and polar scientific research.
Decades of research into quieter underwater propulsion
Gongzhi Marine draws on decades of academic research. Harbin Engineering University, where Guo is a professor, traces its origins to national military engineering. Its naval architecture discipline ranks among the country’s leading programs. Guo’s teacher, professor Huang Sheng, was among the first group of Chinese scholars sent overseas after the founding of the People’s Republic of China to study shipbuilding technology and became an expert in underwater propulsion.
Guo is also a doctoral advisor at Harbin Engineering University and directs the laboratory overseeing underwater propulsion technology at the Ministry of Industry and Information Technology. He was named among the world’s top 2% of scientists in both 2024 and 2025.
He also serves as an expert on the Ministry of Industry and Information Technology’s special evaluation group for improving ship energy efficiency and as a member of the ship mechanics academic committee of the Chinese Society of Naval Architects and Marine Engineers. He was also among the earlier researchers internationally to systematically study the theory and experimental methods of underwater flow fields around ships.
Drawing on more than four decades of research by his academic predecessors and team, Guo identified equipment noise and limited visibility as persistent problems across almost all forms of deep sea operations.
Conventional propellers provide strong thrust, but their noise can interfere with sonar. They also stir up seabed sediment, reducing visibility around the equipment. Alternatives that cause less environmental disturbance often sacrifice thrust and speed, making it harder to hold position against currents and perform demanding operations.
Gongzhi Marine’s answer is what it describes as the world’s first system coupling flexible undulating fins with vector pump jet propulsion.
The flexible fins mimic the swimming motion of the Persian carpet flatworm, propelling the robot through wavelike movements that reduce disturbance to the surrounding water. The pump jet system draws on propulsion technology used in military submarines and significantly reduces the robot’s own noise.
The combined system allows the robot to reach speeds of up to 15 knots and switch to a low-noise mode during operations. Its endurance exceeds 72 hours. Radiated noise can fall as low as 86 decibels (dB), compared with 100–170 dB for conventional systems.
The company said its sediment plume is just 0.2% that of conventional equipment, resulting in almost no disturbance and allowing operations in ecologically sensitive waters.
Beyond propulsion, Gongzhi Marine has developed its own sensing, control, navigation, communications, waterproofing, and intelligent systems, integrating them into a platform for underwater operations.
Its sensing system draws on the team’s datasets spanning multiple sensing methods and operating conditions, along with 200 teraflops of aggregate computing power. An image enhancement model combines physical principles with artificial intelligence algorithms and a five-megapixel high-definition camera, extending perception range by 25% and raising target recognition accuracy above 95%.
From academic research to a commercial business
Turning that technology into a business has required Guo to adjust his approach as he moves from academia into entrepreneurship.
“An academic is quiet, even stubborn. You need to shut out outside interference to do research,” he said. “But an entrepreneur has to be open enough to let more people help and willing to listen to different opinions. Those two ways of thinking are at opposite extremes. I’m evolving too.”
During Gongzhi Marine’s seed round, a listed company offered a valuation two to three times higher than that of another early-stage investor. After thinking about it all night, Guo chose the latter.
“Gongzhi Marine doesn’t need to prove itself through its valuation,” he said. “What matters is being able to learn things from investors that go beyond money.”
Deep sea technology appeared in China’s 2025 government work report alongside commercial spaceflight and the low-altitude economy. Private investors’ interest rose quickly as a result, but much of it remained concentrated in shallow-water applications such as hull cleaning and consumer underwater drones. Deep sea technology remained a niche field.
The early investor introduced Gongzhi Marine to Beijing’s venture capital community. Those introductions, combined with the scarcity of comparable deep sea technology, helped bring the company to investors’ attention.
Gongzhi Marine said more than a dozen investment firms approached it within a year, despite the company having made almost no public statements.
Guo said the company has now completed three funding rounds. Its Series A round, completed in August, included investment from an embodied intelligence fund affiliated with China-Singapore Suzhou Industrial Park Development Group, as well as Xiao Ventures and Yinfeng Capital.
He is also candid about the team’s weaknesses.
“We’re not really low profile. We’re just not good at commercialization,” Guo said. “Almost all 20 people on the team are researchers. We don’t have spare hands for publicity, and we haven’t even attended any major trade shows.”
To demonstrate the commercial potential of its technology, Gongzhi Marine chose a reservoir as a testing ground.
A reservoir might seem a modest testing ground for deep sea technology. But Guo said its confined space requires precision near walls and presents complicated water disturbances. Fishing nets, branches, and other obstacles add to the difficulty. In some respects, that unpredictable environment makes operations harder than in the open deep sea.
“The difference between a reservoir and the deep sea is like asking a humanoid robot to run on an empty sports field versus making it work inside a room packed with clutter,” Guo said. “Which do you think is harder?”
In the reservoir trial, Gongzhi Marine reported a reduction in cost per operation from RMB 5 million (USD 743,000) to RMB 700,000 (USD 104,000). The full inspection cycle fell from six months to 20 days, and the company reported a 90% increase in efficiency.
“This is our ticket to quickly entering the government procurement and major state-owned enterprise infrastructure maintenance markets,” Guo said.
Lower prices and a growing order pipeline
A customer’s experience helped shape Gongzhi Marine’s approach to product design.
“The customer had previously bought two underwater robots from overseas companies at a cost of an eight-figure RMB sum,” he said. “Management hoped the machines would reduce costs, but the frontline engineers said they actually had to hire several more people just to maintain them. They weren’t user-friendly, and when they broke, they couldn’t be repaired. In the end, they basically sat idle.”
Gongzhi Marine therefore set itself a clear-cut product goal:
“Make underwater robots like smartphones. Turn them on and they’re ready to use.”
One example is a handbag-sized underwater robot under development for COSCO Shipping. It weighs only a few kilograms and is expected to sell for less than RMB 100,000 (USD 14,900).
Gongzhi Marine’s current lineup includes a small, 25-kilogram biomimetic remotely operated vehicle (ROV), a large, 1,000-kilogram biomimetic autonomous underwater vehicle, and several proprietary systems. Their prices are reportedly less than one-tenth those of conventional underwater robots.
Guo attributed the lower prices mainly to in-house development, from design drawings to final assembly, and the company’s own manufacturing of key components.
But he is clear-eyed about the limits of that advantage.
“One major reason we can price them this way is that we haven’t included more than 20 years of research costs,” Guo said. “But no technology can remain closed forever. We have to keep introducing new things and maintain our technological edge.”
Gongzhi Marine is developing components including underwater robotic manipulators. It is also working with a team led by an academician on technology combining underwater inertial navigation with a Doppler velocity log, aiming to address challenges associated with underwater loads.
To turn its academic expertise into a viable business, the company will also need to develop its engineering and mass production capabilities.
Gongzhi Marine has secured nearly RMB 100 million in prospective orders for deep sea mining applications.
In other technically demanding fields, including remote equipment maintenance and military maritime reconnaissance, it has confirmed technical specifications and completed commercial negotiations with customers. It has also partnered with the 704th Research Institute of China State Shipbuilding Corporation to provide technical support for China’s national polar scientific research programs.
The company has also made progress overseas. It has signed a framework agreement for underwater cleaning ROVs with a Singapore-based marine engineering service provider, with the first batch of deliveries valued at an eight-figure RMB sum. It has also signed an RMB 30 million (USD 4.5 million) exclusive supply agreement for key components with Canada’s APG-Neuros.
To handle larger orders, Gongzhi Marine began construction of an intelligent factory in the county of Qidong in Nantong, Jiangsu, in July. The facility is expected to produce nearly 100 underwater robots a year.
Guo expects the company to generate RMB 30 million in revenue in 2026. He said revenue could exceed RMB 100 million in 2027, when the company could also reach profitability.
Asked about his longer-term vision, Guo pointed to Huawei.
“Huawei’s development has a lot in common with the history of our team,” he said. “I especially hope Gongzhi Marine can become the Huawei of the deep sea.”
KrASIA features translated and adapted content that was originally published by 36Kr. This article was written by Yang Yuexin for 36Kr.
Note: RMB figures are converted to USD at rates of RMB 6.73 = USD 1 based on estimates as of September 16, 2026, unless otherwise stated. USD conversions are approximate and, where appropriate, rounded for ease of reference. They may not fully match prevailing exchange rates.