Matt Lubbers says the genesis for his new startup was a visit to his friend Tim Clothier‘s house, where a living room view of Mount Rainier was partially obstructed. The problem? A mountain of trading cards from Clothier’s personal collection was in the way.
They weren’t just in the living room. The garage was full of boxes of cards stacked on top of more boxes. A longtime collector, Clothier numbers his lot at about 7 million cards. Separating and organizing them all by hand, he figured he could handle about 25,000 cards a week. He told his wife it would take about 15 years to sort them at that pace.
“I don’t think it was crazy for me to say, ‘What are you doing here?’” Lubbers told GeekWire.
“My friends, when they’re over, I’ll be sorting and they kind of run the other way,” Clothier said. “Matt’s very inquisitive and he started asking questions, and he said, ‘What do you think technology could do for you?’”
More than four years after that initial conversation, the startup co-founders are finding out what technology can do. Renton, Wash.-based Gradient is up and running, using custom robotics and artificial intelligence to help sort, analyze, list and sell sports trading cards, gaming cards, and more.
The goal is to grab a slice of the $15 billion U.S. trading card market, to help customers manage collections small and large, and to simply and quickly get a return on eBay for sometimes forgotten treasures.
Card geeks and engineers
The stealthy operation is located across the hall from the headquarters offices for Seattle Sounders FC at the soccer club’s Renton facilities — the Providence Swedish Performance Center & Clubhouse. Sounders majority owner Adrian Hanauer is an investor in Gradient, which has raised $6 million from mostly friends and family.
Clothier, the CEO, has known Hanauer since he was 15 years old. He spent 30 years at Pacific Coast Feather Co., the Hanauer family’s onetime pillow and blanket manufacturing business.
The sprawling Gradient space looks like any upstart tech company office with a few notable exceptions. There are boxes upon boxes full of trading cards everywhere, stacked near rows of rolling racks also containing boxes of cards — 10 million in all and room for three times that.
A close look at any open box or neat stack of cards reveals the faces of sports heroes past and present across baseball, football, basketball, hockey and more.
Around a few tables there are employees shuffling through some cards by hand. Others at computer stations digitally flip through card files or write the code that helps manage such work. The environment is a mix of card geeks and engineers.
And in one corner, the hum of eight robotic sorters can be heard, pulsing with little bursts of air and whirring as components move cards back and forth on a custom rigging apparatus that looks like something from a rock concert stage.
The system is the brainchild of Lubbers, the chief technology officer, who is a computer vision and AI expert who spent the past 15 years building complex systems and robots for autonomous vehicles and self-flying drones at ZF Group, Faraday Future, Voyage, Amazon Robotics and Zipline.
“We saw that there wasn’t much tech, at the time, in this industry. That’s what got us excited,” he said. “What if we could process cards extremely fast? What if we could reduce the amount of time someone, a customer or even expert, took to identify or price or list the card? That’s what we built.”
Up to 100,000 cards a day can be processed by the robots — and there is room to add more machines.
Lubbers is especially protective of what he’s built, and wasn’t ready yet for GeekWire to shoot photographs or video of the robots at work.
Under bright lights, the machines rapidly move cards to flatbed scanners to capture images of the card backs as cameras positioned overhead take photographs of the card fronts. Every single card is physically and digitally cataloged.
While it may sound like fast-moving robots could be a recipe for disaster when mixed with delicate and sometimes quite valuable paper cards, the system is impressive. From the shape of the 3D-printed trays in which the cards are picked and then dropped, to the buttery soft suction fingers that gently lift each card, there is great care taken to never mark or damage any card.
The collected images are instantly sent to a nearby server room where three custom supercomputers — utilizing a high-density configuration similar to NVIDIA’s H100 or H200 chips — house six GPUs each. These machines handle all AI model training and inference testing, crunching through 500,000 images a day to analyze and score cards against a database of 30 million variants.
Storing and managing a collection
Gradient joins an increasingly tech-heavy ecosystem where AI-powered platforms like Ludex, CollX, Card Boss and eBay’s own scan-to-list feature are already used by collectors to instantly grade and price cards with quick scans via mobile phone apps. Gradient’s closest industrial competitor is probably TCG machines, which makes a robotic sorter used by card shops to process thousands of cards an hour.
Gradient’s goal beyond demonstrating how quickly it can process and accurately assess many thousands of cards is also to prove that it can efficiently store them, find them easily via QR-coded boxes and trays, and move them on the collectors’ market.
The company is just getting started in attracting customers, but its largest so far has given Gradient more than 500,000 cards to process.
Different subscription price tiers attract different customers and collection sizes. Pay-as-you-go card scanning runs 40 cents per card. A premium level subscription is $9.99 per month for up to 10,000 cards; Pro is $29.99 per month for up to 30,000 cards; and Commercial is $99.99 per month for up to 100,000 cards. The levels include secure storage and other perks.
Customers gain access to a personal web portal where they can manage their collections and see images of their cards, read the card details, list them on eBay through the Gradient Collects store, and monitor active and sold listings. A customer can choose one card or “send all my cards to eBay” and Gradient’s system will generate such a request.
Gradient takes between 16% and 20% per sale, depending on the subscription level, with 13% or 14% of that covering the costs with eBay.
The startup, which employs 25 people, streams live auctions on eBay where hosts excitedly open packs of Pokémon cards from a makeshift in-house studio located behind piles of boxes. And the company is also building its own marketplace so it can give customers the option of listing with Gradient, eBay or both.
Like a kid opening a fresh pack of cards at the corner mini mart, the possibilities with Gradient seem pretty endless. Especially for the kid, or, let’s face it, the adult collector, who finally uncovers those attic shoeboxes stuffed with thousands of cards and doesn’t know where to start.
“Our job is to help you digitize and inform you what you have, and then you get to choose what you wanna do with it,” Lubbers said.
