Viking Takes First Steps Toward Sanding Automation
by Aaron Porter23 October 20255 min read
In the eye of the discerning boatbuilder or buyer, beauty is more than skin deep. Not just the glossy surface finish determines the quality of a paint job; the underlying structure must be free of flaws and waves as well. Even in the age of high-quality molded composites, the perceived fairness of a boat hull or deckhouse remains directly proportional to the builder’s investment in longboarding and block sanding time and expertise. The flawless finish that signifies high quality, almost always comes at the cost of countless hours of physically demanding hand sanding. We don’t call them torture boards for nothing.
In the eye of the discerning boatbuilder or buyer, beauty is more than skin deep. Not just the glossy surface finish determines the quality of a paint job; the underlying structure must be free of flaws and waves as well. Even in the age of high-quality molded composites, the perceived fairness of a boat hull or deckhouse remains directly proportional to the builder’s investment in longboarding and block sanding time and expertise. The flawless finish that signifies high quality, almost always comes at the cost of countless hours of physically demanding hand sanding. We don’t call them torture boards for nothing.
Last year, Viking Yachts (New Gretna, New Jersey) turned to automation, artificial intelligence (AI), and robotics specialist Viam (New York City) to help reduce the thousands of hours of manual fiberglass sanding required for its deserved reputation for superb finishes. Conversations started in November 2024, and by July 2025 the first custom robotic block sander was installed at Viking and early-stage testing began, starting with a center console hardtop.
“Fairing is one of the most labor-intensive and critical parts of the build process,” said Daniel Thompson-Rhodes, Industrial Engineer at Viking Yachts and the project lead. “If we can maintain Viking’s standard of finish while reducing physical strain and reallocating our skilled techs to other areas, that’s a win. This isn’t about replacing people—it’s about evolving the process to support them and the long-term quality of the product.”
That rapid deployment was partially enabled by Viam’s deep experience in developing machine learning and AI technology for niche applications. It helps that company founder and CEO Eliot Horowitz is a passionate fisherman and boater. Indeed, Viam’s work on data-driven capabilities for sportfishing, and marine mammal protection first brought him into contact with Viking.
Viking had previously pursued automation for some build tasks in the shop and learned through experience how programmed repeatable robotics like those in the automotive sector wouldn’t work for boatbuilding applications where tasks and variables change frequently. “We needed an adaptive robotic system—one that could respond to different parts as we put them in front of it,” said Winston Zeberlein, an advanced manufacturing engineer at Viking.
Viam designed its sanding tool by learning how the best technicians at Viking did the work by hand. That meant avoiding the random orbital sanders that are good for final finish, but less so at fairing flaws out of the molded surface. The Viam tool is essentially a block sander on a robotic arm with a 63″ (1.6m) reach. It requires a skilled sanding technician to operate it, running the machine’s scanners, changing sandpaper, applying dye to the subject surface, identifying areas in need of sanding, directing whether to sand more or less, and approving when a task is complete.
Contrary to the hype that AI is completely automating production tasks, this is not a technology you can walk away from and find a perfectly faired surface when you return from lunch. “People should be thinking of them as tools first, robots second,” Horowitz said. “You can instill some knowledge into it and have it repeat tasks.”
With that in mind, the attainable goal was not to have a robot that would produce Viking-quality finished surfaces unattended, but rather to create a tool that automates as much of the unpleasant task of block sanding as possible for the skilled technician operating it. As anyone who has subjected their body to days of longboarding can attest, that’s a laudable ambition.
“The most annoying part of the job is the first couple of passes removing the most material. It’s also the bulk of the manual work,” Horowitz said. That’s what the robot is best suited to. Finish fairing will be left to technicians sanding by hand.
In late July, Viking had begun early-stage testing, starting with a center console hardtop. The robot is permanently bolted to a 42″ (1.1m) pedestal; the hardtops on metal carts are rolled to it and secured in position for sanding. Then the robot performs a 3D scan precisely capturing the part’s position, surface features, and dimensions. At the direction of an operator, the robotic sanding block engages with the part, fairing the surface based on the scan geometry. Zeberlein explained that technicians wipe blue dye on the surface to help the unit’s optical scanner and the human operator to see the low spots that remain after each pass, and to sand judiciously until the blue is gone.
When we spoke in July, Viking said early testing suggested the Viam robot could significantly reduce finishing labor on a hardtop while maintaining rigid quality standards. Next steps include more R&D and dialing in the process on hardtops. Long-term goals include expanding the system to larger parts, including hull sides and the eventual use in Viking’s mold shop where fair surfaces must adhere to strict tolerances.
Last year, Viking Yachts (New Gretna, New Jersey) turned to automation, artificial intelligence (AI), and robotics specialist Viam (New York City) to help reduce the thousands of hours of manual fiberglass sanding required for its deserved reputation for superb finishes. Conversations started in November 2024, and by July 2025 the first custom robotic block sander was installed at Viking and early-stage testing began, starting with a center console hardtop.
“Fairing is one of the most labor-intensive and critical parts of the build process,” said Daniel Thompson-Rhodes, Industrial Engineer at Viking Yachts and the project lead. “If we can maintain Viking’s standard of finish while reducing physical strain and reallocating our skilled techs to other areas, that’s a win. This isn’t about replacing people—it’s about evolving the process to support them and the long-term quality of the product.”
That rapid deployment was partially enabled by Viam’s deep experience in developing machine learning and AI technology for niche applications. It helps that company founder and CEO Eliot Horowitz is a passionate fisherman and boater. Indeed, Viam’s work on data-driven capabilities for sportfishing, and marine mammal protection first brought him into contact with Viking.
Viking had previously pursued automation for some build tasks in the shop and learned through experience how programmed repeatable robotics like those in the automotive sector wouldn’t work for boatbuilding applications where tasks and variables change frequently. “We needed an adaptive robotic system—one that could respond to different parts as we put them in front of it,” said Winston Zeberlein, an advanced manufacturing engineer at Viking.
Viam designed its sanding tool by learning how the best technicians at Viking did the work by hand. That meant avoiding the random orbital sanders that are good for final finish, but less so at fairing flaws out of the molded surface. The Viam tool is essentially a block sander on a robotic arm with a 63″ (1.6m) reach. It requires a skilled sanding technician to operate it, running the machine’s scanners, changing sandpaper, applying dye to the subject surface, identifying areas in need of sanding, directing whether to sand more or less, and approving when a task is complete.
Contrary to the hype that AI is completely automating production tasks, this is not a technology you can walk away from and find a perfectly faired surface when you return from lunch. “People should be thinking of them as tools first, robots second,” Horowitz said. “You can instill some knowledge into it and have it repeat tasks.”
With that in mind, the attainable goal was not to have a robot that would produce Viking-quality finished surfaces unattended, but rather to create a tool that automates as much of the unpleasant task of block sanding as possible for the skilled technician operating it. As anyone who has subjected their body to days of longboarding can attest, that’s a laudable ambition.
“The most annoying part of the job is the first couple of passes removing the most material. It’s also the bulk of the manual work,” Horowitz said. That’s what the robot is best suited to. Finish fairing will be left to technicians sanding by hand.
In late July, Viking had begun early-stage testing, starting with a center console hardtop. The robot is permanently bolted to a 42″ (1.1m) pedestal; the hardtops on metal carts are rolled to it and secured in position for sanding. Then the robot performs a 3D scan precisely capturing the part’s position, surface features, and dimensions. At the direction of an operator, the robotic sanding block engages with the part, fairing the surface based on the scan geometry. Zeberlein explained that technicians wipe blue dye on the surface to help the unit’s optical scanner and the human operator to see the low spots that remain after each pass, and to sand judiciously until the blue is gone.
When we spoke in July, Viking said early testing suggested the Viam robot could significantly reduce finishing labor on a hardtop while maintaining rigid quality standards. Next steps include more R&D and dialing in the process on hardtops. Long-term goals include expanding the system to larger parts, including hull sides and the eventual use in Viking’s mold shop where fair surfaces must adhere to strict tolerances.