At around 3 am in 2018, the battery in Smartex’s only prototype exploded. At 7 am, co-founders Gilberto (pronounced “Jeel-berto”) Loureiro, Antonio Rocha, and Paulo Ribeiro were supposed to be in Hong Kong demoing that prototype for Calvin Klein executives. Founded earlier that year, Smartex was developing a computer vision system to detect textile manufacturing defects before they ruin a production run.
Smartex’s solution, however, set fire to the Shenzhen workshops of HAX, SOSV’s hardware startup program, warranting a visit from the local fire department. The team had only a few hours to make another prototype that, hopefully, wouldn’t explode in front of Calvin Klein’s executives. They succeeded and pulled off the demo (barely).
Today, Smartex’s customers produce fabrics for Calvin Klein as well as fashion brands like Gucci, Balenciaga, and Tommy Hilfiger. Smartex has raised over $27 million in venture funding and has installed more than 800 of its defect detection systems in factories around the world. The Portugal-based company aims to create the waste-free, automated textile factory of the future.
Fashion is a $2.4 trillion industry accounting for more than 2% of global GDP. However, the industry is plagued by waste. During textile production, about 5% of fabric rolls are lost to defects caused by poor-quality fibers, unmaintained machines, calibration issues, and human errors. Defects force production to stop, and defective rolls of fabric must be thrown away.
By Smartex’s calculation, a typical 5% defect rate costs textile manufacturers $33 billion annually and worsens fashion’s significant environmental footprint. Research by the Ellen MacArthur Foundation, a circular economy advocate, suggests that textile production is responsible for 2.4% of global greenhouse gas emissions and 4% of annual freshwater withdrawals.
Long before launching Smartex, Gilberto became intimately familiar with textile defects and their costs. Born in 1994 in Barcelos, Portugal, Gilberto is the son of textile workers and spent his high school summers working as a textile inspector. In eight hour shifts, he’d stare at finished rolls of fabric, scanning for tiny defects. He promised himself he wouldn’t spend life working in textile factories.
Gilberto became the first person in his family to attend college. In 2012 at the University of Porto, he met Antonio, a fellow physics major, who also grew up in a working-class family. His mother, born in the Democratic Republic of Congo, owned a hairdressing business, and his father worked for the government.
Gilberto and Antonio became roommates that winter. Not long after, Gilberto went to a board game party to resell beer at a markup (one of his many side hustles). There he met Paulo, a video game hacker and computer science major, who was the son of vegetable farmers—and like Gilberto did not want to work in his parents’ field.
Paulo moved in with Gilberto and Antonio—sort of. “Paulo lived with us without paying rent,” explains Gilberto. “He spent all his time in our apartment, just playing video games.”
“I had an apartment,” insists Paulo, “but I spent a lot more time in their place.” The reason: Gilberto and Antonio were good cooks.
In 2015, Carlos Lei Santos, a mutual friend of Gilberto and Antonio, launched a Silicon Valley startup called UpLink. Inspired by his success, Gilberto and Antonio wanted to launch a startup. The textile industry was a perfect target. It used sophisticated machinery but depended on manual, error-prone processes, which can lead to defects. Gilberto knew from firsthand experience as a textile inspector that reducing defects would be a huge win for the industry and anyone who figured out how.
In factories around the world, a device called the circular knitting machine turns yarn into fabric. If Smartex could retrofit these machines with a system to spot defects before they became visible to human inspectors and stop production automatically for adjustments, factories would reduce wasted energy, water, dye, and yarn.
Gilberto understood the industry, and Antonio, fortuitously, had specialized in the physics of electronics. Since Paulo essentially lived in their home, he ended up helping Gilberto and Antonio with a software problem one night. They asked him to join the team.
The group had the right skills to go after textile inspection. As Gilberto sums it up, “This guy [Antonio] is the hardware nerd and that guy [Paulo] is the software nerd and I’m just a textile business nerd…and the money guy. But I’m technical enough to know if they’re bullshitting me.”
Now, our customers are asking us for traceability and compliance. There are no other startups trying to solve similar problems.Gilberto Loureiro, CEO
After earning their master’s degrees, Gilberto, Antonio, and Paulo took jobs and spent nights and weekends working on their startup. Early in 2018, they had a breakthrough.
In a textile factory owned by a friend of Gilberto, the team retrofitted a circular knitting machine with high-resolution cameras and lights. The images of fabric went to a machine learning algorithm they trained to recognize visual cues associated with defects. The system automatically stopped the knitting machine when it produced a defect called a lycra dash, which manifests as a thin line or streak on the surface of the fabric and is almost invisible to the human eye.
Elated, Gilberto wanted to sell the technology to factories immediately. Antonio, however, thought they should raise venture capital. Their friend Carlos recommended they apply to SOSV’s HAX startup program. Despite Gilberto’s reservations, Antonio went ahead with the HAX application, and they were accepted.
The question for Gilberto was, should they abandon jobs, significant others, and local deals with textile factories to participate in the HAX program in Shenzhen? “If I knew then what I know today, I would have slapped myself in the face,” says Gil. “I was so naïve.” (Today, HAX has a facility in Newark, NJ, as well as Shenzhen.)
To convince Gilberto, Antonio and Paulo posed a question made famous by Jeff Bezos: when you’re 80 years old, will you regret not doing this? For Gilberto, the answer was yes. In November 2018, they incorporated Smartex and moved to Shenzhen.
Shenzhen is heaven for electronics nerds. Prototyping went fast because the Smartex team could fill bags of inexpensive cameras, resistors, and capacitors in the electronics markets. Suppliers were available 24/7, and HAX’s engineering experts were on hand to guide the team.
Smartex arrived at HAX with a conservative plan to expand in Portugal first and Europe second. The HAX team convinced them to think bigger. Their new mentality became, “F**k it, let’s conquer the world,” as Gilberto puts it.
Despite almost burning down HAX’s workshop, Smartex successfully trained their computer vision system to recognize all the major textile defects. Importantly, it could be retrofitted to any of the 300,000 circular knitting machines around the world. At HAX demo day in San Francisco, the founders demoed their prototype for 600 VCs. Nothing exploded. Smartex landed a $2.9 million seed round co-led by DCVC and Spider Capital in 2019.
As COVID-19 emerged in 2020, the team regrouped in Porto to focus on product development. Luckily for Smartex, the textile industry didn’t unravel during the pandemic.
Smartex sold direct to factories, and orders piled up so quickly that they struggled to keep up with demand. In November 2022, that momentum sealed a $24.7 million Series A led by Lightspeed Venture Partners and iPhone co-inventor Tony Fadell’s Build Collective. Two major textile players—clothing giant H&M Group and Fashion for Good—participated in the round.
According to Smartex, each of their defect detection systems saves an average of 264 hours of production time per year and pays for itself in 18 months (assuming each roll requires one hour of production, and each defect detected saves half a roll, on average). Using before-versus-after data from customers, Smartex calculates that it has prevented 491,000 kg of defective fabrics from being produced.
“The opportunity keeps growing,” says Gilberto. “Now, our customers are asking us for traceability and compliance. There are no other startups trying to solve similar problems.”