At a berth in Shenzhen, the cranes keep working while a small machine drops into the water beside a container ship's hull. There is no diver, no hose team on the quay, no shutdown. The machine presses itself against the steel with its thrusters, works along the plates, and a technician on the dock watches the hull appear on a screen.
Above the waterline, nothing about the operation looks dramatic. Below it, the ship is carrying a garden — barnacles, mussels and algae that add drag, burn fuel and, increasingly, break the law.
Key takeaways
- On 5 June 2026, Shenzhen Port Group's marine development arm completed the city's first live trial of a domestic ROV for removing biofouling from a ship's hull, reporting efficiency more than five times that of conventional cleaning.
- In March, Hebei's maritime authority ran a standard-setting field test for robotic hull cleaning at Caofeidian, using two robot models on the same vessel.
- Tianjin-based Shihang Intelligent (世航智能) says its orca-shaped cleaner has serviced more than 1,000 ships, works at roughly 50 times the rate of a human diver, and is supported by service teams in more than 20 Chinese ports.
- The commercial driver is regulatory: the International Maritime Organization aims to complete mandatory biofouling rules by 2030.
Why a clean hull is worth money
A ship's fuel consumption is set by how easily it slides through water. A moderate layer of marine growth can raise fuel use by 20–40% on a mid-sized container ship, according to industry estimates cited in Chinese naval-engineering coverage — which is why hull cleaning has always existed, usually in dry dock or by divers with brushes.
Both traditional routes are awkward. Dry-docking takes the ship out of service and costs a fortune in lost voyages. Diver cleaning happens at anchor or in port, is weather-dependent, and puts people in the water next to propellers and thrusters.
A robot that can work while cargo is being handled removes the scheduling problem entirely. That is the actual product: not a cleaner, but a cleaning that fits inside a port call.
Three deployments, one direction
The Shenzhen trial used a domestically built ROV that holds itself against the hull with propeller thrust rather than magnets, which matters because a hull is curved, coated and rarely flat exactly where the growth is. The machine used an adaptive cleaning head, was driven along a planned path, and sampled water quality while it worked — an attempt to show that the removed growth is not simply released into the harbour.
At Caofeidian in March, the exercise was different in kind. Two competing robot models cleaned the same ship under observation by the Hebei Maritime Safety Administration and the Ministry of Transport's environmental protection centre, with water sampling and imaging before and after. The stated purpose was to build the data needed for third-party certification and, eventually, for rules.
Shihang Intelligent supplies the commercial version of the story. Its machine is shaped like an orca for drag, uses magnetic-coupled thrusters instead of a magnetic track, and the company says it has cleaned more than 1,000 vessels from service teams in over 20 ports. Its pitch is blunt: divers cannot keep up, and the hull coatings survive the robot better than they survive a brush.
The regulation is the business model
Biofouling is not only a fuel problem. Organisms carried on hulls are one of the main ways marine species cross oceans, which is why the IMO has been building a mandatory framework, with completion targeted for 2030.
Once rules bite, a clean hull stops being an efficiency choice and becomes a compliance requirement — with documentation. That is the part the industry is positioning for now: not just cleaning, but recording when a hull was cleaned, where, by what, and what was removed.
The economics, and the catch
The claimed savings are large: fuel-efficiency gains of 15–20% after robotic cleaning in tests at Zhoushan, annual fuel savings of US$750,000 to US$1 million for a large container ship, and payback within 6–12 months. Removal rates above 90% with paint damage below 5% are the lab figures the industry quotes.
Treat those as directional. They depend on how fouled the ship was, how fast it sails, what fuel it burns and what the alternative would have cost. A ship cleaned on schedule needs less cleaning; a ship that skipped two years is a different job.
The more durable advantage is that robotic cleaning can be sold as a service at the berth, priced per hull, with a data record attached. That turns a chore into a subscription — and gives ports a new revenue line that does not require new concrete.
Honest limitations
The Shenzhen efficiency figure comes from the port group and the maritime authority's account of a single trial, not a controlled comparison across a fleet. Shihang's "1,000 ships" and "50 times a diver" are company claims; no independent audit is published. The fuel-saving and payback ranges come from industry and naval-engineering coverage that cites laboratory and single-port tests, and they assume a level of fouling that may not match a well-maintained ship. Water sampling during cleaning does not prove that all removed material is captured. And the regulatory timeline is a plan, not a rule: the IMO's 2030 target could slip, which would weaken the compliance case that underpins current investment.
What readers can do now
- If you charter or operate ships, ask for before-and-after hull performance data, not cleaning hours — fuel curves are the only honest measure.
- If you run a port, the service model is worth studying: berth-time cleaning needs no new land and monetises an existing asset.
- If you follow ocean policy, watch the IMO's biofouling text. It will decide whether this becomes a compliance market or stays a fuel-saving nicety.