Chips & Hardware
Glimpse speeds up CT scanning to catch hardware defects earlier
Glimpse, a startup that boosts throughput on CT scanners to catch manufacturing defects, is launching a new product called Explore for aerospace, automotive, electronics and medical-device customers.
Recalls are costly for manufacturers across industries. The one that sidelined every Chevy Bolt built between 2017 and 2022 — triggered by improperly installed parts inside battery cells at an LG factory — ultimately cost LG $1.9 billion. Volkswagen, Stellantis, Hyundai, Toyota and others have all suffered through their own EV recalls.
“The majority of those major quality issues could have been caught with CT scanning. And the reason why they weren’t is because it’s way too slow to do any kind of statistical sampling,” Eric Moch, co-founder and CEO of Glimpse, told TechCrunch.
To catch more defects, Glimpse has built software that lets battery manufacturers inspect cells 10 to 30 times faster by significantly boosting throughput on their existing CT scanners. Using the software, manufacturers could scan tens of thousands of cells a day, up from a current handful, Moch said. A customer “super scanner” that Glimpse is developing with partners could eventually complete a scan in a second or two, he added.
Batteries aren’t the only parts prone to costly defects, so Glimpse is launching a new product, Explore, the company told TechCrunch exclusively. It will let customers in industries such as aerospace, automotive, electronics and medical devices use Glimpse’s image-processing software to speed up quality-control checks on CT scanners in their own labs and development environments.
Glimpse currently has over 100 customers, including Anker, Lucid and the U.S. Navy, spanning both buyers and sellers of batteries, and generates revenue in the “mid seven-figure” range, Moch said.
To wring more speed out of traditionally slow CT scanners, Glimpse wrote its own image-processing pipeline, built around deep-learning algorithms, which compresses and cleans up scans. Much of the initial processing happens on edge computers deployed at a customer’s site before the data reaches the cloud. The result, the company says, is clear images that load in seconds rather than the minutes or hours scans traditionally take.
Glimpse’s dashboard also shows customers part measurements they specify — in batteries, for instance, the distance between layers in a cell — and flags and highlights deviations from the norm in the scanned images. The underlying deep-learning models retrain on customer input as they go.
Customers without their own CT scanners can ship parts to Glimpse, which runs its own scanner in the Boston area; the company is opening a second Scan On Demand center in the San Francisco Bay Area. “You can spend a couple hundred bucks or a few thousand dollars on Glimpse, and you don’t have to buy a very expensive piece of equipment,” Moch said, adding that customers can later buy their own scanner and run Glimpse’s software on top of it.
Why it matters
By making CT scanning fast and cheap enough for routine use rather than spot sampling, Glimpse is betting it can catch the kind of defects that have triggered billion-dollar recalls before parts ever reach a finished product.