Industries / Transport and vehicles / Shipbuilding and repair / Boat building
Hulls
Hull work is choosing the material for the hull and deck and turning it into a watertight shell: fastening planks, riveting or welding metal, laying glass fabric and resin in a mold, or printing plastic.
7
Milestones achieved, from 1942 to 2019.
3
To watch for: announced, not yet confirmed outside.
3
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1942
Glass and polyester hull
Ray Greene, a boatbuilder in Toledo, Ohio, molded a small boat from glass fiber and polyester resin, a liquid plastic that hardens into a solid. He got the glass through work he did for Owens-Corning Fiberglas, also in Toledo, and a laboratory batch of polyester from American Cyanamid. Earlier plastic boats, including his own, used resins that had to be baked under heat or pressure. The historian Pete Smyth, who interviewed Greene in the 1970s, called it "the world's first fiberglass polyester boat" and gave the year as 1942; Practical Sailor found a few references to 1941 and nothing that put anyone ahead of him. By 1947 or 1948 Greene was building the 16-foot Rebel sailboat in fiberglass in series, and Good Old Boat counts more than 5,000 built.
Credit
Ray Greene; Owens-Corning Fiberglas and American Cyanamid for the materials
1945
Canoes from aircraft aluminum
Grumman, which had built aluminum aircraft for the US Navy, began making aluminum canoes in 1945. Light Metal Age says William Hoffman, the company's chief tool engineer, had the idea in 1944 after carrying a heavy wooden canoe in the Adirondacks, and that Alcoa supplied an alloy and sent Russell Bontecou, who had built his own aluminum prototype. Sheets were stretched over a mold and riveted. The 13-foot pilot model weighed 38 pounds, a little more than half the weight of the wood and canvas canoes of the time, and by 1947 Grumman was the largest canoe supplier in the United States. The Spokesman-Review reported in 1996 that the plant in Marathon, New York, was to close, with the market split among canoes of polyethylene, fiberglass and Kevlar.
Credit
William Hoffman and Russell Bontecou; Grumman; Alcoa
1970
Wood sealed in epoxy
Wood rots and swells when it takes up water. Meade and Jan Gougeon of Bay City, Michigan, learned from the pattern maker and boatbuilder Vic Carpenter to glue wood with epoxy, a two-part resin that needs little clamping. With chemists from Dow Chemical's epoxy laboratory nearby, they worked out a mix that could also be brushed on as a coating, so every piece of wood in a hull was both glued and sealed. In July 1970 they launched the trimaran Adagio, which the company calls "the first large, all epoxy bonded and sealed wooden boat built without the use of fasteners" and says is still sailing. In January 1971 they began packaging the epoxy for other builders as WEST SYSTEM, and in 1979 they published a manual of the method.
Credit
Meade, Jan and Joel Gougeon; Vic Carpenter; Dow Chemical's epoxy laboratory
1987
Blisters traced to water
In the early 1980s fiberglass hulls began to grow blisters below the waterline, between the gel coat (the pigmented outer layer of resin) and the laminate (the layers of glass and resin behind it). The US Coast Guard paid for a study in 1985 through the American Boat Builders and Repairers Association. Thomas Rockett and Vincent Rose, chemical engineers at the University of Rhode Island, found that water passes through the gel coat, dissolves water-soluble material left in the laminate and forms an acidic solution; osmosis then draws in more water until the surface lifts. Power & Motoryacht reported in 2008 that builders answered by using vinyl ester resin in the outer layers in place of cheaper general-purpose polyester, controlling shop temperature and humidity, and adding barrier coats, and that blistering on new boats had nearly stopped.
Lessona hull can leave the factory sound and still fail slowly in the water, and the cure was a costlier resin in the outer layers.
Credit
Thomas J. Rockett and Vincent Rose; US Coast Guard
1990
Resin drawn by vacuum
In open molding, workers roll or spray wet resin onto glass fabric in an open mold. On 20 February 1990 William Seemann received US patent 4,902,215 for another order of work: lay the fabric dry, cover it with a sealed plastic sheet and a mesh that spreads resin, then let a vacuum pull the resin through. One of the patent's examples is the hull of a 13-foot dinghy, 3/32 inch thick. The method is now known as resin infusion. The US Environmental Protection Agency's 2001 rule for boat plants lists resin infusion molding as one of three basic kinds of closed molding used in boat manufacturing.
Credit
William H. Seemann III
2001
Limits on styrene
On 22 August 2001 the US Environmental Protection Agency issued air rules for boat plants (40 CFR Part 63, Subpart VVVV). It had identified boat manufacturing as a major source of hazardous air pollutants such as styrene, which is given off by polyester resin and gel coat. The rule capped the content of such pollutants in resins and gel coats used in open molds, set different caps for sprayed and non-sprayed application, and gave existing plants three years to comply. EPA counted about 119 plants subject to the rule and estimated it would cut their emissions by 3,450 tons a year, about 35 percent below the 1997 level. It wrote that it encouraged closed molding "because of its low-emitting nature."
Credit
US Environmental Protection Agency
2019
A hull printed whole
The University of Maine's Advanced Structures and Composites Center printed a boat, 3Dirigo, in 72 hours on a large 3D printer, a machine that builds a part by laying down beads of melted plastic. The university gives its size as 25 feet (7.62 m) and 5,000 pounds (2,268 kg) and says Guinness World Records certified it as the largest 3D-printed boat and the largest 3D-printed object. In 2023 Al Seer Marine and Abu Dhabi Maritime took the boat record with an 11.98 m water taxi, by CompositesWorld's account: two hulls, each printed in 5.5 days on a machine from the Dutch firm CEAD, from material that was 67 percent recycled. Neither boat is a production model.
Credit
University of Maine Advanced Structures and Composites Center; Al Seer Marine, Abu Dhabi Maritime and CEAD
Watch for
A fast printed boat
The University of Maine launched 3Dirigo X, a 30-foot 3D-printed boat it says was built to handle offshore conditions at speeds approaching 40 knots, and showed it on the water at Hampden, Maine. The university says its printing rate has risen from about 20 pounds an hour to about 500. At the launch the US Defense Department's manufacturing technology director said the boat still had to be evaluated for "its ability to stand up over time, over stress."
Achieved whensea-trial and durability results are published or confirmed by a party outside the university and its funders, such as a classification society.
Claimed by
University of Maine, August 2026
Printed hulls in series
Dutch Boat Factory in Delft, which took over CEAD's boat-printing equipment in January 2026, says it prints complete hulls of up to 12 m in one piece. It has added a second printer for unmanned boats and large parts and plans a second 12 m line in early 2027 for series production. CEAD said in 2025 that it had printed a full-scale fishing boat in 88 hours from a plastic based on high-density polyethylene.
Achieved whena customer or a classification society reports printed hulls in regular service and says how many were delivered.
Claimed by
Dutch Boat Factory, November 2026
A robot block sander
Fairing is sanding a molded part until its surface has no waves, and it takes thousands of hours of hand work on a large yacht. Viking Yachts of New Gretna, New Jersey, installed a robot from the software firm Viam in July 2025: a sanding block on an arm with a 63-inch (1.6 m) reach that scans the part and sands where an operator directs. Testing began on the hardtop of a center-console boat. Professional BoatBuilder reported Viking as saying early tests suggested a significant cut in finishing labor; no figures were given, and final fairing stays with people.
Achieved whenViking or an outside party publishes hours saved and finish quality on production parts, including hull sides.
Claimed by
Viking Yachts and Viam, October 2025
Open
01
Robot lays dry fabric
A robot arm setup that lays pre-cut pieces of dry glass fabric into a small female mold, such as a dinghy mold or a one-meter section of a hull, and smooths them into the corners so the stack is ready for resin infusion. No resin is used in the test.
The briefCredit · unclaimed02
Old hulls, cut and tested
A public table of strength, stiffness and glass content measured on small test pieces cut from fiberglass hulls that are being dismantled, with each hull's build year, builder and the place on the hull the piece came from.
The briefCredit · unclaimed03
Infusion, predicted and filmed
A public set of filmed infusion trials on flat and ribbed panels under a clear cover, each with its layup, the liquid's thickness, the vacuum level and the fill time, beside what an open simulator predicted for the same panel.
The briefCredit · unclaimed
Open 01 · the brief
Robot lays dry fabric
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A robot arm setup that lays pre-cut pieces of dry glass fabric into a small female mold, such as a dinghy mold or a one-meter section of a hull, and smooths them into the corners so the stack is ready for resin infusion. No resin is used in the test.
Why it is still open
The European Commission's summary of the DrapeBot project (2021 to 2024) says almost all marine structures are made by draping, which is placing soft fabric on a shaped mold, that most draping is done by hand, and that only 5 percent of aerospace composite parts use any automated draping. The project built its cells around carbon fabric and a person working beside a large robot. Sea Ray's RIMFIRE robots, described by CompositesWorld in 2005, sprayed chopped glass and binder into hull molds, which is a different material from woven or stitched fabric, and tight corners were still laid by hand. The nearest marine work is Leibniz University Hannover's robot-mounted draping head, which its AutoBLADE project (2021 to 2022) tried on molds for tidal turbine blades about 2.5 m long, made from carbon fabric, by CompositesWorld's account. Searches of the web, GitHub and Hugging Face in November 2026 found draping simulators but no filmed, counted trial of a robot laying glass fabric in a boat mold.
Built with what exists
One or two low-cost robot arms and a camera; an open robot model that can be taught from a few demonstrations, such as NVIDIA GR00T 2 (see the Note "Five demonstrations per task"); stitched glass fabric cut to pattern; a secondhand dinghy mold or a plywood mold; gloves, long sleeves and a dust mask for handling glass fabric.
First step
Borrow a dinghy mold, cut three layers of fabric to fit, and film one hand layup to set the marks and to define what counts as a wrinkle.
Done looks like
Twenty filmed attempts at a five-layer stack, with a published tally of layers placed within 10 mm of their marks and free of wrinkles, and the time per layer beside a person's time on the same mold.
Open 02 · the brief
Old hulls, cut and tested
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public table of strength, stiffness and glass content measured on small test pieces cut from fiberglass hulls that are being dismantled, with each hull's build year, builder and the place on the hull the piece came from.
Why it is still open
A 2019 report for the International Maritime Organization says fiberglass boats were given a life expectancy of 30 to 50 years and that many older ones are still in service. France's take-back scheme dismantled 3,093 boats in 2025 at an average age of 42 years, by its own count, so the material is on hand. The nearest study found, from the University of Rijeka, tested new test pieces sunk in the sea for 6 and 12 months. A web search in November 2026 found no public measurements from recreational hulls after decades of use.
Built with what exists
Offcuts from a licensed dismantling yard; a wet tile saw, dust extraction and a respirator for cutting; a three-point bending test on a benchtop test frame at a college or makerspace; a college laboratory furnace to burn off the resin and weigh the glass; Python for the tables.
First step
Ask one dismantling yard for offcuts from three hulls whose build years are known, and test them.
Done looks like
Test pieces from at least 30 hulls built 30 or more years ago, five from each hull, with results, photos and method published, and each hull compared with a newly made panel of the same glass content.
Open 03 · the brief
Infusion, predicted and filmed
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public set of filmed infusion trials on flat and ribbed panels under a clear cover, each with its layup, the liquid's thickness, the vacuum level and the fill time, beside what an open simulator predicted for the same panel.
Why it is still open
Regulators favor closed molding (see 2001 above), and a yard that misjudges where resin will flow can leave dry fabric in a hull. Open simulators exist, among them RTMsim, published in the Journal of Open Source Software in 2023. Brigham Young University published videos of ten resin infusion experiments in 2025, with the flow front and voids traced in each, but they are carbon fiber preforms in resin transfer molding, a closed rigid mold, not glass under a vacuum bag. Searches of GitHub and Hugging Face in November 2026 found no public set of filmed fills of boat-style layups, such as thick glass stacks over grooved foam, set beside a simulator's predictions.
Built with what exists
A sheet of glass or thick acrylic as the mold face, a vacuum bag and a small vacuum pump; a harmless stand-in liquid such as corn syrup thinned to the thickness of infusion resin, so that no resin fumes are involved; a phone camera; OpenCV or a segmentation model to trace the moving edge of the liquid; RTMsim.
First step
Fill one 500 mm square stack of four glass layers with thinned syrup under a bag on a glass table, and film it from below.
Done looks like
Twenty filmed fills across at least four layups, each with measured and predicted fill time and a map of where film and prediction differ, with all video, inputs and code published.