The Robot Sculptors of Italy

Bloomberg

09-07-2025

Businessweek
Two studio owners whose clients include Jeff Koons and Maya Lin dominate robot-chiseled art. One is spending big to preserve the human touch.
By Matthew Hart
Photographs by Alecio Ferrari and Lyndon French
July 9, 2025 at 6:00 AM EDT
Sai Baba lay on his back in the cold sunlight on the mountain, his expression untroubled by the four-and-a-half-ton steel robot drilling at his head. The machine buzzed away in a mist of water and atomized stone. Its arm could move up and down and in and out. It could twist, tilt and spin its tool. Altogether it could work along nine axes—an impressive range of motion. Through the glass walls of the factory behind me, I saw other robots toiling, but this was the one Giacomo Massari had brought me up to look at.
Sai Baba was an Indian sage who died in 1918 and is still revered today. The followers who were paying Massari $600,000 to carve his likeness wanted something big. They were getting it. It had taken Massari's team a month to locate the 100-ton block of top-grade marble needed for the job and another month to define the block with diamond saws, coax it from the surrounding rock, and get it to the robot without breaking it. When finished, Massari said, the 15-ton statue would be the biggest robot-carved sculpture ever. It would take the machine four months.
'At that point it will be 95% done,' Massari said. 'The final 5% will be the hand-finishing.
'How long for that?'
He shrugged. 'Three months?'
We stood there contemplating the robot. Back and forth and back and forth went the diamond-studded carving tool. It never paused. It never took a break. By contrast the hand-finishers would stop for lunch and bathroom breaks and insist on halting work altogether at the end of the day. Maybe even go home and sleep. Humans, what can you say.
We were high above Carrara, the world's biggest and most famous source of marble. Hundreds of quarries have worked these slopes in the Apuan Alps above the Tuscan coast since Roman times. Carvers prized the creamy marble of Carrara for its granular fineness, a quality that helped produce an illusion known as morbidezza —the softness of living flesh. Gods and heroes and rearing horses have poured from their chisels. Smirking cherubs by the zillion. When he was 22 years old, Michelangelo ransacked the hills in search of the perfect block—on the heights above us we could see the scar of the quarry where he found it. Two years later he produced the Pietà, the work that ignited the High Renaissance.
Many of the most famous images of Western art were made from the rock of this great metamorphic uplift. Rome itself, the Eternal City, was clothed in Carrara stone. It's as if the entire project of the West, its courts and capitols, its idols and its art, came streaming from these hills.
Today the great meme factory is short of carvers. Their numbers have fallen sharply. Studios run by the same families for generations have closed. Many artists embrace the robots, praising their speed and accuracy. The machines save money too, taking half the time to perform the laborious roughing-out stage for which sculptors would otherwise often hire specialist artisans. Despite some artists' enthusiasm, though, they tend to hide their robot use behind nondisclosure agreements. Carrara is knee-deep in NDAs, and I asked Massari why. 'Artists,' he said with a tender smile, as if we were talking about his kids. 'They don't want to destroy the idea that they are still chiseling with a hammer.'
As a segment of the marble business, sculpture is dwarfed by the industrial side, which slices slabs by the millions of tons each year. Robots help these companies mill countertops and shower stalls for markets around the world. But fine art sculpture is big business too, worth billions of dollars a year.
The first robot sculptor appeared in Carrara in 2005. Now there are about 30, and the total worldwide is around 100. Two men play outsize roles in this rapidly evolving business. One is Massari, the more evangelistic of the two. His corporate mothership, publicly traded Litix SpA, trumpets Massari's vision of the future on the first page of a slick marketing brochure. 'We Don't Need Another Michelangelo: In Italy, It's Robots' Turn to Sculpt,' proclaims the newspaper headline he reproduced from a New York Times piece on his company.
The other man is a bluff Midwesterner named Jim Durham. In Carrara, they'd known him for decades. He often bought stone there for his thriving fabrication business in America. Still, the Italians must not have grasped what truly drove him. Otherwise, they wouldn't have been so surprised when he completed a yearslong stalk and, last Oct. 29, snapped up Franco Cervietti, the most respected carving atelier in Italy. Some in the Italian marble business were merely stunned; others, horrified. 'An American!' a third-generation stone trader gasped when I asked him about it.
They should have seen Durham coming, because back at his $20 million stone fabrication plant in Sun Prairie, Wisconsin, he was executing pieces for some of the world's top artists: names like Maya Lin and Martin Puryear. He was the biggest producer of fine art stone sculpture in America, and now, with his Franco Cervietti purchase, the world. To the Italians, this coup may have looked like Rome falling to the Goths—in would come the robots, and out would go the hallowed traditions of the last great hand-carving house in Italy. But that wasn't what Durham had in mind at all.
Litix creates its robots at a wholly owned subsidiary called Robotor. The business occupies one end of a sprawling stone yard in the industrial heart of Carrara. Massari took me on a drive past huge blocks of marble to a gray, barnlike structure. Inside was an immaculate assembly plant and a busy staff. About a dozen robots stood around in different stages of construction. Most had come from Kuka AG, a Chinese-owned German manufacturer.
Massari pays $70,000 for a basic robot, the kind you might find snapping windshields into cars on an assembly line. He transforms it into a sculpting machine he can sell for north of $200,000 by tearing it apart and reassembling it with extra joints. Then he adds computing power. 'Kuka has a language, and we write our own code to go inside that language,' Massari explained. It's the code that turns the robot into a sculptor.
A robot starts with a block known as the resting model. The machine carves by following instructions called toolpaths, which map the carving tool's successive routes through the stone. These instructions change every time the machine removes a layer of stone, because the removal effectively creates a new resting model, and the route from that to the finished piece will therefore be different.
Big-name artists may have their own staffs to write toolpaths; for those who don't, Massari offers a subscription service that takes on the coding for them. The client uploads a 3D file showing the desired finished piece and marries it with another 3D file, this one describing the block of stone selected for the job. Robotor's system does the rest. The simplest version of this software goes for $5,000 a year, and progressively more powerful iterations are available for $10,000 and $15,000. 'We try to make it easy for the client,' Massari said. 'They make the design, we do everything else.'
Human vs. Machine (and Human)
Time required to sculpt a reproduction of "Amore e Psiche" by Antonio Canova at Robotor
This makes the process sound simple, and sometimes it is. A lot of Sai Baba, for example, was his flowing robe—a breeze for the robot. But the software's origin and evolution were industrial, not aesthetic. It understands 'Turn left,' not 'Fix those lips!' Moreover, a carving robot doesn't experience the hand-brain feedback loop that neuroscience calls 'embodied cognition,' in which our brains tell our hands what to do and the hands help the brain to think about it. All this was running through my mind as we pushed through a door into a studio and met the toolpath from hell.
The robot stood to one side. It seemed to be considering the figure posed before it—for sure an eyeful. Soaring above us, her 8-foot height increased by the pedestal she stood on, was a white marble ballerina. Her expression was so softly carved you couldn't tell if she was in the throes of ecstasy or sound asleep. It didn't really matter, because the girl herself wasn't the main event. The real star was her skirt—a tempest of lace that foamed around her slender body. It was hard to imagine it had ever been a block of stone.
'Wow,' I said.
'I know,' said Massari.
'How long did that take to carve?'
'Twenty months.'
And still not finished. I could see little lumpy bits of stone clotted up high in the folds. We studied the statue for a minute. ' Jeff Koons?' I asked. The truth was, I'd recognized it immediately. A pink version is all over the internet. Massari didn't say a word. He let me gape for a bit before ushering me from the studio. I supposed the work was covered by an NDA, but why the secrecy? Koons is pretty much the General Motors of machine-assisted art—he'd be more likely to make a documentary about the robots than conceal them. The answer may have had more to do with the proprietary value of the toolpath, which could have taken years to code. Koons is, after all, one of the most successful artists in history, and there could be any number of purely business reasons for shielding his projects with NDAs.
When I spoke with Koons about the ballerina later, he acknowledged the work as his. The one at Robotor was part of a planned run of 12—four each in pink, black and white. He explained that he'd taken the form of the ballerinas from Meissen figurines —German porcelain statuettes that have had an avid following since the 18th century—and scaled it up. Interestingly, he discovered that Meissen craftsmen created the lacey skirts by dipping actual lace into porcelain slip. Then, when the pieces were fired in the kiln, the lace burned away, leaving only its ceramic memory. When Koons' workshop made CT scans of the original figurines, the staff spotted hollows where the lace had been. Later, when his coders wrote the toolpaths for his own ballerinas, he said, they had to eliminate the hollow spaces—presumably so the robots wouldn't try to drill them out.
The hard truth is that carving robots are literal to a fault. If a human makes a mistake, the robot won't catch it. That happened to Keara McMartin, who runs a studio in Pietrasanta, the town near Carrara where most of the sculpting has traditionally happened. She'd sent out a job—a larger-than-life head—to a local robot lab. But the lab forgot to swap out the drill bit used for roughing and replace it with the one for finishing, which was 4 millimeters shorter than the roughing bit. The robot just kept carving—4 millimeters deeper than it should have. By the time someone noticed the shrinking head, it was ruined. Making matters worse, the stone was a beautiful dark-green Chinese marble that turned out to be irreplaceable. The quarry had mined the last of it.
McMartin came to Carrara 45 years ago, after art school in Maryland. She told me she misses the social life she found back then—the bars teeming with young sculptors, the swaggering specialists known as ornatisti. 'Even at work,' she said, 'there was always a huge party when a job was finished. The new generation, sometimes they just send you a digital file. They may not even come to the studio.'
Yet the old artisanal community was withering even then. 'The work can be sporadic,' McMartin said, 'sometimes with long periods of unemployment.' It was also hard and poorly paid, which discouraged some artists from passing on their skills. 'There were master craftsmen in Pietrasanta who would deny their own children access to their studios,' she said. The robots' arrival simply sped up a decline that was already underway.
Then the inevitable happened: Where there had been too many carvers, now there were too few.
Robot sculpture often needs hand-finishing. The Koons ballerina, after 20 months with the robot, would probably go through a year of manual finishing, maybe longer. It would be exacting work—getting up into the lace to finish the holes and hollow out the flounces. Such work would be too delicate for a robot. With the number of carving studios reduced, the surviving ones, with their veteran artisans, have become increasingly valuable—none more so than the atelier of Franco Cervietti.
Cervietti had the carvers, and he had the cred. His was the Vatican's go-to studio for conserving and restoring pieces from its vast collection. When Milan Cathedral—festooned inside and out with one of the densest hoards of sculpture in the world—had a restoration job that demanded an artist's touch, Franco Cervietti was on the speed dial. A-listers like Koons sent pieces there. That was the studio Jim Durham bought, and the reason he bought it.
Durham grew up in Fairfax, Virginia, the son of a military officer attached to the Pentagon and a mother who'd trained as a nurse but ended up doing a different kind of triage, sorting and prioritizing the daily mail at the White House. At college he majored in chemistry but concluded that life in a lab wasn't for him. He'd had summer jobs at stone yards in the DC area, cutting flagstone and loading crushed marble into the backs of station wagons, and he found that he liked the work—the mineral smell, the materiality. Drawn to the artistic possibilities of the medium, he signed up for a night course in sculpture at George Washington University.
'My goal was to carve a perfect rose, with the petals so thin the light could pass through them,' Durham told me. He worked with a piece of gray Vermont marble and a mallet bought for him by his father. Every few days he would buy a fresh tea rose to use as his model. Over the weeks, the rose emerged from the stone. But when he got down to the tiniest details, he couldn't find chisels small enough for the work. 'I was working days at a stone yard in Falls Church at the time,' Durham recalled, 'and I mentioned the problem to one of the German masons who ran the place. There were four of them—gnarled old guys who knew a ton about stone. This one grabbed an empty Pepsi bottle and smashed it on the floor. Those old soft-drink bottles were really chunky, and the glass shards were thick. 'There's your chisel,' he told me, and he was right. Some of the shards tapered to a sort of diamond point, and if you held one between your thumb and forefinger you could scrape out stone with it.'
When Durham's wife enrolled in a graduate program at the University of Chicago, he packed away his mallet, bought a book called Who's Who in the Stone Business, and started looking for work. At a quarry in South Elgin, Illinois, now a Chicago suburb but then still rural, the owner drove Durham around the property then asked if he thought he could run it. 'I said sure. That was on a Thursday. On Monday I had 50 union guys waiting for me when I got there and heavy equipment I knew absolutely zero about.'
He learned fast. 'The demand was crazy,' he said. So crazy that in a single year he was able to run up the price of a ton of limestone from $18 to $96. Another quarry snapped him up, this one in Wisconsin. Two years later he teamed up with his biggest customer and bought out the stone fabricator that owned the quarry. He was 25.
Durham ran Madison Block & Stone for eight years, fabricating countless tons of masonry and custom stone and developing a presence in architectural work. By the time he went solo and formed Quarra Stone Co., his private company, he felt increasingly drawn to sculptural fabrication. He brought in master carvers from Germany so he could compete for complex jobs. His confidence grew, and in 2017 he took his most ambitious shot—bidding on the Eisenhower memorial in Washington, DC. The design was by Frank Gehry. Durham threw himself into the pitch, visiting the Eisenhower library in Abilene, Kansas, and constructing a full-scale model of Gehry's design that included actual Kansas prairie grasses. The bid failed when the jury decided Durham didn't have the fine art chops. Crushed, he made a vow to himself: 'That is never going to happen to me again.'
In March, while I was visiting Pietrasanta, the machines at Durham's Wisconsin plant were finishing a pair of granite reliefs by Maya Lin—a 7,000-square-foot installation for the outside of the new JPMorgan Chase & Co. headquarters on Park Avenue in Midtown Manhattan. Some 30 blocks to the north, Charles Ray's monumental granite Two horses stands at the entrance to the Metropolitan Museum of Art's contemporary galleries. Durham's robots carved that too, just as they've split apart huge blocks and boulders and pixelated them with bright epoxy dots for Sarah Sze, a MacArthur Fellow and veteran of both the Venice and Whitney biennales. Clearly Durham had the fine art chops that others once thought he lacked. He himself had never been among the doubters, otherwise he wouldn't have had the confidence to pursue the famous Cervietti atelier and nail it for $7 million.
I met Durham for lunch in a comfortable bar at his hotel in Pietrasanta, where he now spends two months a year. He's a tall, thickset man, his hair cut short and neatly combed. He ordered a mocktail. Beside him was Martin Foot, the artistic director at Durham's Italian spinoff company, Quarra Italia Srl. Wiry and gray-haired, in jeans and a sweater, Foot asked for red wine and, when he got it, raised his glass.
'A real stone carver,' he intoned, 'is a robot that runs on this.'
Foot first picked up a chisel when he was 14, in his uncle's stone-restoration studio in Liverpool, England. Five decades later he's still carving, making sculpture in the classical tradition at his studio in Pietrasanta. After Durham lost the Eisenhower bid, it was Foot he hired to advise him. In a tidy stroke of fate, the man who'd recommended Foot was Cervietti. 'I didn't need any help with robots,' said Durham, who'd already had them at his American plant for years. 'Nobody can compete with us in robot fabrication.' But there's only so far a robot will go. 'People think if you feed it enough data, a robot can make a perfect replica of anything.'
One time, as an experiment, Durham scanned a tabletop in his conference room. The resulting 3-D file comprised 20 million data points, and he loaded it into one of his machines. 'The software took one look at the 20 million points and reduced them to three,' he said. 'Three points is all you need for a flat surface, so that's what I was going to get. But it wouldn't be the surface I'd scanned, with all the little nicks and bumps that made it different from any other surface.'
'And that's what robots do to sculpture,' Foot interjected. 'I see it. The art is starting to look the same. The robot doesn't take the art up. The art comes down to meet the robot.'
'You can get a pretty good finish with a robot,' Durham said. 'Where you quit is when you reach the place where the human can do a job better. Everybody makes a choice about where to make that switch. If you let the robot go as far as it can, you finish the job faster, but you get a sort of Ikea version of a sculpture. Some people with big robot operations might think they can make art faster and cheaper. I say, knock yourself out. I'm not Ikea.'
At the studio in Pietrasanta, still called Franco Cervietti in deference to the master, Durham has a single milling machine, with two more on order for delivery this year. There are no robots as such, though advanced milling machines are robots of a kind—automated and programmable and capable of executing intricate tasks along multiple axes. I met Foot at the studio one morning, in an industrial suburb of the town. The place looked like a warehouse, an impression that carried inside, where plaster models of famous sculpture loomed from the clutter. Foot was directing a wholesale rearrangement of this impressive stash—the largest, he told me, in the world. Such models are essential to making copies of existing sculpture. We took off on an eye-popping tour, shelf after shelf and room after room. Bernini. Verrocchio. Rodin. Durham planned to photograph every model—there were more than 2,500—to create a proprietary database. This treasury would be an important part of his new Italian business, allowing him to fill orders for reproductions of famous sculpture. It was one of the reasons he'd bought the atelier. Another reason was hard at work upstairs.
The satisfying smell of cut stone filled the air as I followed Foot to a sunny workshop. We stopped to watch a finisher named Simone Fortini work on a copy of Antonio Canova's Reclining Naiad, an 1814 sculpture of a river nymph sprawled on a couch. Fanned out on a table and pinned to an easel nearby were dozens of photographs of the original. A commercial gallery in Florence had ordered the copy, Foot explained, and sent them a model produced by another studio in Pietrasanta.
'I thought the model was off,' Foot said, 'and when we milled the piece we left some extra stone in that problem area so we could do that whole part by hand.'
Fortini was trying to recapture the languid grace of the original. I was struck by how raptly he studied the photographs, and by the tenderness of his expression. His take-home pay would average around $3,000 a month. He was 47 and had been doing this for 30 years. Through Foot, I asked if he came from a carving family and, when he said no, why he'd taken it up.
The question stopped him. Light streamed through the windows of the long room as he considered it. Carving is a grueling trade. Stone chips nick the skin. A misstruck chisel can cause injury. Fingertips bleed from polishing for days on end. Every force a carver exerts on the stone is a force the stone will send back into his hand. They suffer this every day, and finally Fortini said why.
'For the passion of the stone,' he said, in a voice not much louder than a murmur. His face was powdered in marble dust. 'For the love of the stone.'
I caught Foot watching me. He raised his eyebrows to make sure I hadn't missed the point. The love of the stone—that's what Durham had bought from Cervietti. For $7 million, he now owned the most distinguished stone-carving house in the world.
Often when I spoke to Durham, we ended up speculating about what sculptors call the zero surface. The zero surface is the point where the robot stops carving. It has followed its track through the stone and is done. Only an artist can go past the zero surface into the illimitable data cloud of the human soul.
Durham's arc, from a suburban stone yard in Maryland to the most storied carving atelier in Italy, sped eagerly toward that point, where art is perfect and sublimely certain solely because the maker put a chisel one way and not another. 'The unexplained human choice,' he called it.
'Could a robot make an unexplained choice?' I asked the last time we spoke, and Durham chuckled.
'Not yet.'