Helping engineers see inside their products to solve design challenges
If you’re a neurologist and you suspect there’s an irregularity in your patient’s brain, like a possible aneurysm or tumor, you don’t saw the patient’s head in half to look. You schedule them for a CT scan or an MRI.
But think back half a century to the early 1970s, when MRI hadn’t been invented yet and CT machines were new, expensive, and rare. In that era, unless you worked at one of the handful of hospitals that owned a CT scanner, you probably would have ended up ordering a riskier and more invasive test such as an angiogram or a biopsy.
Product designers and manufacturers today are in precisely the same position as neurologists back then. They often need to know whether their prototypes or production samples contain internal defects — and an industrial CT scanner is a powerful tool for looking inside objects without destroying them. But these machines are so expensive and complicated that only giant, deep-pocketed companies can afford to own and use them. That leaves most quality-control labs with more primitive options, like cutting their products in half with 1960s-era band saws.
Starting today, all of that changes. Lumafield, a Cambridge, MA-based startup that DCVC has backed since 2020, is entering the commercial market and announcing the availability of Neptune, the world’s first low-cost industrial CT scanner, and Voyager, a companion browser-based CT analysis system. Just as companies like GE and Siemens transformed healthcare by building medical scanners every hospital could afford, Lumafield aims to revolutionize manufacturing by giving literal X‑ray vision to every designer and engineer.
“This changes everything about the way manufacturers think about inspecting their parts,” says Eduardo Torrealba, Lumafield’s CEO and co-founder. “People will often ask me, ‘Who’s your biggest competitor?’ And I always say it’s a person with a saw. That’s what people do to understand the things that they’ve made: they cut things in half, they polish them, they take pictures under a microscope. The only people who have access to CT scanning either have millions of dollars to buy a CT scanner, or they mail their parts off to somebody and wait weeks to get scans back. And even then, they don’t get access to the raw data to be able to manipulate it and really understand what they’re looking at. We give them all of that for $36,000 per year.”
CT, which stands for computed tomography (from the Greek tomo=slice and graph=write), came to the world of design and manufacturing the long way around. The originator of the idea, an American neurologist named William Oldendorf, believed that if he could beam X‑rays through an object from all angles, he could mathematically transform the density readings into a series of two-dimensional slices of the object. He proved that his prototype worked by scanning objects he found around his house, including a record player, an alarm clock, and a toy train.
Geoffrey Hounsfield, an electrical engineer working at EMI, the British tech conglomerate now known as one of the “big four” record companies, figured out how to computerize Oldendorf’s process for medical scanning. (That breakthrough earned him the 1979 Nobel Prize in medicine.) Decades later, in the mid-2000s, companies such as Nikon and Carl Zeiss rediscovered Oldendorf’s idea and built the first CT machines for scanning parts rather than people. Using more math, today’s machines can combine hundreds of 2‑D slices into detailed 3‑D reconstructions: real-world reflections of the idealized CAD drawings used in product design.
Lumafield has brought industrial CT within the reach of more product-testing labs by reimagining almost every step in computed tomography. Scanners from Zeiss and other established players depend on large, expensive flat-panel X‑ray detectors. But Neptune, which can scan any product up to the size of a paint can, uses an entirely new, low-cost system for capturing X‑rays and transforming them into digital signals. On the software side, meanwhile, Lumafield moved the computing-intensive image reconstruction process from specialized desktop computers to secure cloud servers. And it built an intuitive browser-based tool, Voyager, that non-experts can use to interrogate the 3D scans and zero in on potential product defects.
“All of this stuff has been reevaluated from scratch, from a fundamental physics level, to come up with the least expensive, easiest-to-use version of a CT scanner that we could possibly make with the technology that exists,” Torrealba says. “Sometimes we joke about how, with the hardware, the reconstruction pipeline, and the Web application, we’ve done three startups’ worth of engineering.”
That hard work is already paying off. Lumafield has quietly built a customer base of dozens of companies who are using CT data in design decisions, including 3D printer maker Desktop Metal, fitness tracking company WHOOP, and consumer products giants such as L’Oréal, OXO, Saucony, and TrekBicycle. One L’Oréal packaging-quality director, for example, says Lumafield’s scans saved his team four months of design and testing work by identifying the source of a stubborn leak in a bottle cap. (The culprit was a tiny, 100-micron indentation — invisible to traditional inspection techniques — caused by unexpected interference between the bottleneck and the cap.)
Lumafield’s founders come from the worlds of consumer electronics, additive manufacturing, battery manufacturing, and design software, which means they’re acutely aware of the shortcomings of existing diagnostic tools in manufacturing. DCVC led a series A round in 2020, but the company remained in stealth, allowing it to perfect its products free of distractions or hype. “We didn’t want people to know about us until we’d actually shipped multiple products to customers and they were solving multimillion-dollar problems,” Torrealba says. “Now real companies are using our technology to make real decisions.”
Here at DCVC, we seek to fund Deep Tech companies with expert founders who are tackling hard science and engineering problems that only they know how to solve and that, once solved, could transform old industries and create new ones. Lumafield and its experienced cofounders — Andreas Bastian, Kevin Cedrone, Ric Fulop, Scott Johnston, and Eduardo Torrealba — qualify in all respects.
We also see the industrial CT as a classic example of the innovator’s dilemma narrative so compellingly described by Harvard Business School scholar Clayton Christensen. Existing makers of industrial CT scanners, reluctant to cannibalize their own sales, have no incentive to develop cheaper or more accessible versions of their tools. Lumafield envisioned technology that performs scans at a drastically lower price, and, as a result, it’s creating a market much broader than the big scanner makers ever dreamed of.
“We’re bringing the cost down so much, it’s possible to use this technology to understand any object — it doesn’t matter if it’s a complex electronics assembly or a single injection-molded piece,” Torrealba says. That means product builders can rethink the way they design things, and use CT in situations they never considered before — with consequences we can only begin to predict. “If we can get manufacturing right, we have a better chance of getting all these other things right — climate change, transportation, medical devices, communication, all this stuff,” Torrealba says. “But right now, one part of manufacturing is stuck in the past, and that’s understanding and inspecting the things that you’re building. These new tools make it possible for every engineer to have access to this powerful technology.”