A lot goes into a single basketball shot. Tiny, difficult-to-spot shifts in a player’s elbow or wrist placement can be the difference between a swish and an airball. A new lab in Toronto is using tech to get those measurements and analysis down to a science.
Peripheral Labs has teamed up with the Quantum Sports and Learning Association (QSLA) to develop what it claims is North America’s first biomechanics basketball shooting lab outside of the NBA.
“There is something about being in Canada I think you don’t really get anywhere else.”
Kelvin Cui, Peripheral Labs
The two organizations have partnered to bring Peripheral’s spatial intelligence technology to competitive basketball coaching at QSLA’s facility. The goal is twofold: to help students at QSLA-affiliated schools like Uchenna Academy improve their shooting, and to serve as a testbed for Peripheral to bolster and demonstrate the efficacy of its sports tech.
Last week, BetaKit toured Peripheral’s courtside set-up, played with its tech, and sat down with co-founder and CEO Kelvin Cui to unpack what this partnership means, why he moved from self-driving to sports tech, and the rationale behind Peripheral’s return from San Francisco to build in Canada when so many other ambitious founders are doing the opposite.
From LLMs to LRMs
To date, Peripheral has been trialling its tech and gathering the data it needs to refine its proprietary AI model on an adhoc basis, spending weeks planning and entire days setting up and taking down their cameras at various locations just to capture a few hours of footage.
“We’ve never had the opportunity to really run multiple reps of our system in the same location,” Cui said in an exclusive interview. “This facility now gives us that.”
The QSLA court in question is located on Dupont St., on the north end of Toronto’s Dovercourt Village in a large factory building-turned-basketball-and-educational-hub that used to make the gears to open and close the retractable roof of the Rogers Centre (then the Skydome).
Peripheral has recently lined the court, which Cui described as its new research and development lab, with 36 off-the-shelf cameras and moved its headquarters and nearly 20-person team to an office upstairs.
The company’s hardware is not its secret sauce: that distinction belongs to its AI. Peripheral is developing a large reconstruction model (LRM), a type of foundation model capable of quickly converting two-dimensional images captured by cameras into three-dimensional (3D) video.
RELATED: Canadian tech looks to poach H-1B visa castaways as its own “ambitious founders” flee
Peripheral claims its AI can generate photorealistic volumetric data and 3D video that can be viewed from any angle to help evaluate basketball players’ shooting mechanics in minute detail. It measures their form, trajectory, and release via its 65-point tracking system. The company also sees potential in performance improvement, injury prevention, and other applications across a wide variety of sports.
Navigating that video with an Xbox controller felt like using the instant replay function in an NBA 2K game. This is no mistake: the startup also hopes to turn live sports broadcasts into interactive, immersive, video game-like 3D experiences, particularly for younger fans amid waning viewership. Peripheral also sees potential to help broadcasters ensure that they “never miss a moment.”
Peripheral is less than two years old, but the startup’s tech has already caught the attention of the NBA. The basketball league is testing the tech via its Launchpad program for its automated officiating team, for use-cases like confirming which player touched the ball last before it went out of bounds.
An “arena-scale” biomechanics lab
While biomechanics-based player tracking already exists in baseball for pitchers and batters, Cui said gathering this data and making it actionable in basketball—which has far more player movement and variation—has proven not only technically tough but also expensive and time-consuming to date. The Peripheral CEO said basketball coaches have typically mounted a camera and then manually reviewed it for insights.
Dave McNee, managing director of QSLA and principal at Uchenna Academy said the school, which caters largely to young basketball players and supports professionals in the offseason, previously relied on such a system.
In an interview with BetaKit, McNee said Uchenna Academy hopes to give its coaches a leg up by providing them with “a perspective that no one else sees [or] has” by working with Peripheral. While the Peripheral partnership is still a “testbed” at this point, he said, “some of the data that they’ve provided to us, it just blows us away.
Cui claimed that QSLA marks the continent’s first “arena-scale” biomechanics shooting lab, adding that while NBA teams can measure players’ shooting from specific locations while practicing, Peripheral can capture these metrics across the entire court in practices and games.
RELATED: Waabi claims largest-ever Canadian tech fundraise as it hauls in $750-million USD Series C
“That’s a huge unlock that we’re expecting is going to be like the next Moneyball moment in terms of arena flow sports,” Cui said.
Peripheral’s goal over the next year is to get as many teams and facilities using its tech as possible, across basketball and other arena sports. It has done deployments for hockey, tennis, badminton, and volleyball, and it continues to work up to stadium-scale games like football and soccer.
From self-driving to sports tech
Cui is a competitive badminton player who grew up cheering for the Vancouver Canucks. He and his co-founder, Peripheral CTO Mustafa Khan, previously studied robotics engineering together at the University of Toronto (U of T), where they spent time building self-driving cars with U of T Formula Racing and aUToronto and won several trophies.
Khan, who played competitive basketball when he was younger, was also a researcher at Huawei and the Vector Institute, while Cui spent time interning at Cisco and AMD, working at Toronto pilotless plane startup Ribbit, and building chassis systems at Tesla.
Peripheral’s neural rendering and real-time 3D LRM is built on breakthroughs in robotics and autonomous vehicle perception. Cui said they decided to apply it to sports rather than self-driving because they wanted to build their first business in a market where they had the chance to build a “default-alive” company rather than a “billion-dollar research bet” in a field where it needed to play catch-up against much better-capitalized incumbents.
They moved to San Francisco and launched Peripheral in October 2024, and have since raised $3.85 million USD to date, including a $3.6-million seed round in May 2025 led by Khosla Ventures with support from Daybreak Capital, Entrepreneurs First, and Transpose Platform.
“Some of the data that they’ve provided to us, it just blows us away.”
Dave McNee, QSLA
They moved Peripheral back to Toronto last year to take advantage of their existing networks and the research talent coming out of U of T and the University of Waterloo, and apply what they learned in Silicon Valley here to their home turf by building an ambitious company from here and helping more Canadian athletes and sports teams win.
Cui is thrilled to be back in Canada, working from an office above a basketball court that offers a clear view of the cityscape and will give him the chance to spend his summers in Toronto.
“San Francisco is gorgeous,” Cui said. “Incredible city, right? But there is something about being in Canada I think you don’t really get anywhere else.”
Feature image courtesy Peripheral Labs.