Multiple reports suggest that live sports viewing has declined for certain sports, especially among Gen Z. To solve this, leagues and broadcasters are trying to make sports more engaging for fans with different kinds of viewing experiences, stats, and analysis.
One way to do this is using volumetric video generation that lets users view the play from various angles, giving an inside-the-video-game experience. The core technology uses numerous cameras to capture the footage in 3D for everyone to look at it from various viewpoints. Canada-based Peripheral Labs wants to make this technology affordable for leagues and teams so it can reach more broadcasters and fans.
Peripheral Labs was founded by Kelvin Cui and Mustafa Khan in 2024. Both have worked on driverless cars for the University of Toronto’s team, winning several trophies. Khan has worked as a researcher at Huawei, and Cui has experience working on chassis systems as a software engineer at Tesla.
“Both Mustafa and I are huge sports fans. He has been a massive Arsenal fan, and I grew up watching the Vancouver Canucks since I was seven. When Mustafa showed me his research about 3D reconstruction, my brain said it would be cool to watch hockey like this [in a free-flowing, multi-angle way]. This is how we started on Peripheral Labs,” Cui said in a call with TechCrunch.
The company said the idea of volumetric generation isn’t new. But with the new AI models and advances in computer vision, its founders are confident the technology is ready for the masses.
The duo is using their experience with self-driving cars to apply concepts of robotics perception and 3D vision for the 3D reconstruction of video in sports. This system can reduce the camera requirement from over 100 to as few as 32, helping decrease cost and operational overhead, according to Cui and Khan. The startup aims to keep the hardware cost as minimal as possible for teams and broadcasters and sign multi-year contracts for its platform.
The software platform will bring biomechanical data of players and stats for teams and leagues using its own sensor stack, which is similar to the sensors on self-driving cars that capture the scene with depth. It will enable new ways to control the viewing of the play for broadcasters and fans using photorealistic 3D reconstruction technology. For instance, if fans wanted to track only the player with the ball, they could do that. They can also freeze a moment in-game to see different angles for a foul or a critical moment in play.
Techcrunch event Join the Disrupt 2026 Waitlist Add yourself to the Disrupt 2026 waitlist to be first in line when Early Bird tickets drop. Past Disrupts have brought Google Cloud, Netflix, Microsoft, Box, Phia, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, and Vinod Khosla to the stages — part of 250+ industry leaders driving 200+ sessions built to fuel your growth and sharpen your edge. Plus, meet the hundreds of startups innovating across every sector. Join the Disrupt 2026 Waitlist Add yourself to the Disrupt 2026 waitlist to be first in line when Early Bird tickets drop. Past Disrupts have brought Google Cloud, Netflix, Microsoft, Box, Phia, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, and Vinod Khosla to the stages — part of 250+ industry leaders driving 200+ sessions built to fuel your growth and sharpen your edge. Plus, meet the hundreds of startups innovating across every sector. San Francisco | WAITLIST NOW
“While we work with off-the-shelf cameras, the way we package it with our experience in robotics and ML is what gives us an edge both in terms of platforms and also scaling from small practice enclosures to big soccer and football stadiums,” Cui said.
On the software side, the platform said it can observe different joints, including finger movements of players, to measure flexion. For instance, in the video above of two people playing football (soccer), the system measures flexion of knees and ankles. This could give coaches more ideas about body positioning and the flexibility of a player, and help them improve.
... continue reading