Icon’s plans to 3D-print its way out of the housing crisis

Jason Ballard took the stage at Dell Hall, a 2,440-seat auditorium in Austin, and for a second seemed intimidated by the crowd. 

“Howdy!” he said, wearing a big white cowboy hat and dark jeans. “I can’t believe the room is this full,” he added, chuckling nervously. 

Ballard, the CEO of construction technology company Icon, doesn’t move a room with fast talk and soaring rhetoric like the stereotypical startup founder. Instead, Ballard prefers to outline his out-of-the-box thesis — that 3D-printed homes can solve the housing crisis — with the slow, methodical pacing of a lay preacher. 

Speaking in March at the largest real estate event of South by Southwest, the lay preacher had been thrust onstage at a megachurch. 

Ballard warmed up as he moved on to more familiar territory. 

“We’ve been building with sticks and bricks for about a thousand years,” he said, a graphic projected behind him showing the skyrocketing expenses of building a home. “We’re pretty good at it, and that’s the problem.”

Icon is arguably the best-funded 3D-printing construction company in the world, but its work has largely been confined to small projects in Texas so far. Outside the buzz of the venture-funded innovation industry, it can be hard to spot the moment when promise becomes business — when a startup shows it can succeed, and then when it actually does.

Maybe it is now. Ballard says the company has refined its technology to a point where it is ready to hit the gas. It is releasing a next-generation printer, capable of building 24,000-square-foot buildings up to 27 feet tall, and a suite of tech tools to speed up adoption by architects.

Still, success is far from guaranteed. Most startups flame out, and even unicorns die. (Icon reached a $2 billion valuation in 2022). Proptech in particular has seen its share of high-flying promises followed by flameouts. Icon says it has reached a pivotal point where it’s cheaper to build walls with its printer than by traditional methods, but if its calculations are even a little off, that could threaten adoption — developers are sharply sensitive to any cost increases right now.  

Friends with benefits

For all the aw-shucks charm, Ballard is a sharp operator.

As of October 2023, the company had raised $443 million and employed nearly 500 people, according to PitchBook Data. Icon’s most significant haul came in early 2022, when it raised $392 million at a $2 billion post-money valuation. Norwest Venture Partners and Tiger Global Management led that round, which included investments from Lennar’s tech and innovation arm, LenX; Bjarke Ingels Group; and Fifth Wall, the country’s largest real estate-focused venture capital firm. 

In all, according to PitchBook, 65 firms have put their money behind Ballard’s conviction that 3D printing can, at the very least, prove a profitable alternative to traditional construction techniques. 

Babe Ruth only got a hit in about three in every 10 at-bats; the best venture capitalists similarly swing and miss far more often than they get a hit. That’s why investors spread their bets, and why a startup is not guaranteed to succeed just because a big-name VC backs it. Unicorns can go bust particularly when there’s a physical product.  

But in several cases, Icon has worked on developments with its financial backers.  

One such collaboration is the biggest project Icon has actually delivered yet. Lennar and Bjarke Ingels are working with Icon to print 100 houses in Georgetown, a suburb in Austin’s northern reaches. The homes that are currently available range between $470,000 and $579,000, according to Lennar’s website. A dozen have already sold. 

“We almost won’t get out of bed for less than a hundred homes anymore.”
Jason Ballard

While 100 homes are just a drop in the bucket for a volume homebuilder like Lennar, the project has, thus far, hit Icon’s dual goals of affordability and aesthetics. The Ingels designs undulate, as it is just as easy to print a curve as it is to print a straight line. 

Stuart Miller, Lennar’s executive chair, told the New Yorker he was “somewhat dismissive” of 3D-printing technology. “Innovation is an expensive exercise with unproven returns,” he said.

But Eric Feder, LenX’s president, thought Ballard had “the kind of founder vibe — a charisma rooted in relentless energy — that, though familiar in Silicon Valley, was uncommon in the construction industry,” according to the same article. Construction has proven more resistant to reinvention than other industries — the Egyptian pyramids contain plaster — and few startups in the sector have made it as far as Icon. 

Another major investor, NASA, isn’t interested in building the average cul-de-sac. Icon won a $57 million contract to help develop construction technology that would work on the moon, using materials harvested there so astronauts wouldn’t have to lug it up themselves. NASA also invested in Icon in a round that ended in January 2021. 

Of course, NASA has not printed a moon base just yet, but the plan isn’t total science fiction. The agency planned to test an Icon printer in a simulated space environment in February, and at the Johnson Space Center in Houston, four astronauts are spending a year inside a 3D-printed structure meant to simulate life on Mars. 

None of these projects is big enough to make a dent in Icon’s ultimate ambition of printing our way out of the housing crisis. But the company is moving closer to walking the walk. 

“We almost won’t get out of bed for less than a hundred homes anymore,” Ballard told the New Yorker. “This is a problem at scale, and so we need to be working at scale.”

Your Next Architect: AI 

While it improves its printing technology, Icon is also working on technology to spread its uses to more architects. At South by Southwest, the company announced a design chatbot, Vitruvius, that creates floor plans and renderings based on text prompts. 

Ballard has long enjoyed AI art programs like Dall-E, according to the New Yorker. His new system combines some of the more whimsical aspects of AI art, encouraging prompts like “Ice Palace,” with the nuts and bolts of traditional architectural work.

When prompted to create plans for an Arctic igloo complex, Vitruvius responded with a few prompts that a bespoke architect might ask a client. “Could you tell me more about your family? How many people will be living in this Arctic Igloo Complex?”

Given the input of “two adults and four children,” Vitruvius understood.

“Great, a family of six,” it wrote. “Would you prefer each child to have their own bedroom, or are shared bedrooms an option you’d consider?”

It then asked about open floor plans, sustainability goals and aesthetic requirements.

Within seconds, Vitruvius spat out three options, each with its own renderings, design and floor plans ranging from 1,500 square feet to 2,200 square feet. The largest was like Frank Lloyd Wright’s take on an igloo, bright and angular with big hexagonal windows. The renderings showed a kitchen and bedrooms decorated with fluffy white cushions and Scandinavian wood floors.

The renderings were slightly fuzzy in the way lots of AI-generated art can be, but they included surprising details. In one child’s bedroom, the fluffy pillows turned cotton-candy pink and purple. Another child’s bedroom was decorated with no fewer than eight stuffed penguins.

A request for a writer’s retreat in the woods got an equally detailed and quick response, but the results didn’t do quite as well capturing bespoke requests. Certain textures and details persisted between the three options, like a buckhorn chandelier and stone fireplace. Vitruvius did not include the requested bay window, but each floor plan did include a library. 

Like most custom works, the homes looked expensive.