Offshore wind's most stubborn cost line is not the steel in the tower or the copper in the export cable. It is the rope-access technician who climbs a 100-metre blade to repair the erosion that rain, sand and salt spend years carving into it. In 2025 and 2026, robots started doing that job from a commercial vessel. The question is no longer whether they can. It is whether the industry can scale them faster than its turbines get bigger.
Scale already exists. Aerones has serviced 25,000 turbines and 75,000 blades across 35 countries, for customers representing more than half the world's installed wind capacity.
The next layer is resident robotics. Bubble Robotics raised $5 million to keep inspection robots at sea for up to six months, attacking the vessel-and-crew line that accounts for 80–90% of offshore inspection cost.
The shift matters because offshore maintenance is the industry's least scalable cost. Blades get longer, farms move further from shore, and the weather windows for sending people up a turbine keep shrinking. A robot flown to the blade by drone and finished in minutes changes what a maintenance campaign can be. It also changes what an operator is willing to underwrite.
Fastest robotic blade repair
Clobotics' SPARROW repaired a single blade at Ørsted's Horns Rev 2 in Denmark, and up to eight blades in one day under suitable conditions. · Clobotics, 2026
Reported cost reduction
The company reports that robotic leading-edge repair costs 50–75% less than conventional methods and runs up to ten times faster. · Clobotics, 2026
Aerones robotic fleet reach
75,000 blades across 35 countries since 2018, for customers such as NextEra, GE, Vestas, Enel and Siemens Gamesa. · Aerones, 2026
Growing: robots that repair, not just inspect
Inspection was the first beachhead. Clobotics' IBIS drone inspects all four sides of a turbine's three blades in roughly 15 minutes, detects defects down to a millimetre, and has flown more than 180,000 blade inspections across 40-plus countries. Drones can already check an operating offshore turbine without shutting it down. As we wrote in August, autonomous drones had already inspected an operating offshore turbine, and the repair work was still done by people.
Repair is the harder problem, and it is the one that moved in 2026. SPARROW is a remotely operated robot delivered to the blade by a heavy-lift drone. Once attached, it sands and cleans the surface, then applies a new leading-edge protection layer. At Ørsted's Horns Rev 2 wind farm in Denmark, the system completed a leading-edge repair in 27 minutes on a single blade and up to eight blades in one working day. The company reports a 50–75% cost reduction against conventional methods, at up to ten times the speed. In an earlier 2026 campaign the same team worked across 29 blades in six days, with a fastest single repair of 38 minutes.
The company's own history shows the arc. Sparrow existed as a rough concept in 2022; by versions 6 and 6.1 the robot body had stabilised and the engineering focus shifted to making offshore deployment safe and repeatable on a working vessel, with containerised robot storage, battery handling and a dedicated operator station behind plexiglass. That is the unglamorous part that turns a demonstration into a service business. The first offshore robotic leading-edge repair was completed in 2025 in one hour and 40 minutes; the same task now takes 27.
Aerones attacks the same problem from the inspection side and has pushed into repair. The Latvian company says its robots have serviced 25,000 turbines and 75,000 blades across 35 countries since 2018, for customers that represent more than half the world's installed wind capacity, among them NextEra, GE, Vestas, Enel and Siemens Gamesa. In August it added robotic UV curing for the most severe categories of leading-edge erosion, extending a repair robot's reach into damage that previously meant a longer intervention. TIME and Statista ranked it the ninth-largest green-tech company of 2026. Its Crawler Gen 3 inspects blade interiors, the same niche Clobotics targets with its KIWI robot.
Between them, Clobotics and Aerones define the commercial repair layer: robots that arrive with the maintenance vessel, work inside the same weather window, and leave behind a digital record of what was repaired and when. The record is not incidental. A repair that is documented becomes a data point an operator can use for warranty claims, insurance and the next maintenance cycle.
Falling: rope access and the scheduled shutdown
What shrinks is the traditional repair economy. Manual blade work is a craft: twelve-hour shifts in protective suits, fibreglass dust, heat, and a rope. It is skilled, flexible and expensive, and its output depends on the person doing it. Two crews can inspect the same blade and produce different findings.
The cost structure explains why robotics is moving in. In offshore operations, vessels and crews account for 80–90% of inspection cost and can run up to $100,000 a day, according to Bubble Robotics. Every hour of vessel time spent on a blade is a fixed cost that a faster robot amortises across more turbines per campaign. The scheduled shutdown compounds it. Leading-edge erosion is driven by rain, dust, salt and insects hitting the blade at speed over years of operation, and the standard response has been to stop the machine, send up a team and work through a narrow weather window. An inspection that stops a 15-megawatt turbine costs far more in lost generation than the same stop on the 5-megawatt machines it replaced, because fewer, larger units concentrate the revenue risk on each blade.
Rope teams are unlikely to disappear. The likelier outcome is a hybrid: robots handle the repetitive, high-frequency leading-edge work and the routine inspections, while people handle the complex structural repairs the machines flag. That division is already visible in how the vendors describe their own products, and in the fact that most robotic systems still carry an operator, a technician and a crew lead on the vessel.
New: the resident-robotics bet
The frontier beyond the maintenance vessel is the resident model: robots that stay on site instead of travelling with a crew. Bubble Robotics, founded in 2025 by Jean Crosetti and Patricia Apostol, raised a $5 million pre-seed round in April 2026 led by Episode 1 Ventures and Asterion Ventures, with Norrsken Evolve and several offshore-energy operators participating.
Its BubbleDock is a surface platform that deploys and retrieves subsea robots capable of staying at sea for up to six months, drawing on onboard energy generation. The robots combine autonomous docking, navigation and optical, sonar and environmental sensing to build millimetre-scale maps of offshore wind foundations, pipelines and ports, with AI turning the data into inspection reports, digital twins and predictive-maintenance signals. Bubble says it has more than $4 million in signed letters of intent and deployments planned across offshore wind, maritime security and subsea infrastructure. The idea is to replace episodic, weather-dependent vessel visits with continuous coverage, removing the vessel and crew rather than the sensor.
The same logic shows up in public research. Sandia National Laboratories, with International Climbing Machines and Dolphitech, has built ARROW(e) (Assessment Robot for Resilient Optimized Wind energy), a crawler that deploys from the turbine nacelle, suction-cups to the vertical blade surface and runs phased-array ultrasonic inspection through the full thickness of the composite, finding subsurface damage a camera cannot see. In the Netherlands, TNO, TU Delft and Robohouse are developing a robot that installs strain sensors inside a blade. And ARPA-E has opened a programme, Harnessing Autonomy for Energy Challenges Offshore, with a funding opportunity expected in 2026 and aimed squarely at autonomy for offshore energy.
None of this is yet a solved market. A 2021 review of AI and robotics in the offshore wind sector, published on arXiv, identified certification of autonomous platforms for safety compliance and the reliability of resident operations as the two barriers standing between field trials and standard practice. Five years on, the commercial records and the funded prototypes suggest the first barrier is being negotiated in the field rather than the laboratory.
Where the two models stand
| Parameter | Robotic repair | Manual rope access |
|---|---|---|
| Turbine status | ✔ Can work on a stopped unit inside the window | ✗ Full shutdown for the campaign |
| Time per blade | ✔ 27 minutes, reported | ✗ Hours to days |
| Cost | ✔ 50–75% lower, reported | ✗ Crew and vessel heavy |
| Subsurface data | ◐ Ultrasonic crawlers still maturing | ✗ Visual inspection, operator-dependent |
| Safety exposure | ✔ No work at height | ✗ Rope access at height |
The comparison is not close on cost or speed. What manual work still wins is judgement: the ability to improvise when the damage is worse than the inspection suggested. That is the gap the resident and ultrasonic systems are trying to close, by giving the robot richer data before it starts.
Signals to watch
Whether Ørsted extends robotic leading-edge repair beyond Horns Rev 2 to its wider offshore fleet, the clearest proof the model travels
Certification of autonomous platforms for safety compliance, the barrier named in the sector's own literature and the gate to uncrewed resident operations
Bubble Robotics converting its $4 million of letters of intent into paid, long-duration deployments
Whether ARPA-E's offshore-autonomy programme funds a cohort of resident-robotics projects in 2026, which would pull public capital into the same thesis
