Peripheral Labs

Market position

Peripheral positions itself as making volumetric video affordable and practical enough for broader adoption by leagues, teams, broadcasters and rights holders, in contrast to legacy volumetric capture systems requiring dense camera arrays. TechCrunch identifies Arcturus Studios as a competing startup in volumetric capture for sports. Investors frame the company as both a sports and an entertainment play, with Entrepreneurs First partner Joe Ros arguing that its sports deployments build a data, technology and deployment moat, and Inovia partner Karamdeep Nijjar describing real-time volumetric video plus biomechanics at scale as something no one else has cracked.

The company attributes its edge to applying autonomous-vehicle robotics and machine learning experience to neural rendering, allowing it to use off-the-shelf cameras with its own sensor packaging, sharply reduce the number of cameras needed, and scale from small practice enclosures to large soccer and football stadiums. It also claims real-time performance and combined volumetric video plus biomechanics output from a single platform, and states install cost reductions exceeding 90 percent relative to legacy volumetric capture.

Technology

The system applies robotics perception and 3D vision techniques from autonomous driving to sports video. A large reconstruction model performs neural rendering — described by the CTO as taking sensor inputs and learning a photorealistic 3D representation of the world — converting standard multi-camera feeds into navigable 3D scenes in near real time. It uses off-the-shelf cameras combined with the company's own depth-capturing sensor stack, similar in concept to self-driving car sensors. Reported camera requirements range from as few as 32 up to at least 36 for basketball and hockey arena installations, against over 100 for prior approaches; the company says the approach cuts install costs by more than 90 percent versus legacy volumetric capture. Additional engineering priorities include reducing latency and increasing reconstruction resolution.

Go-to-market

The company sells directly to North American leagues and teams, working through proofs-of-concept and evaluations of temporary or permanent installations, and has used accelerator-style validation channels such as the NBA Launchpad program and grassroots partnerships (QSLA) to demonstrate its technology. Social distribution via team accounts, such as Toronto Raptors clips, has provided public exposure. A browser-based consumer replay product is planned as a direct-to-fan channel.

Sports leagues, teams, stadiums, broadcasters, rights holders and media companies, plus coaching and athlete-development programs; the ultimate audience is sports fans, with an explicit focus on younger viewers who grew up with video games such as NBA 2K, NHL and FIFA.

Geography

Headquartered in Toronto, Canada, after relocating from San Francisco in 2025; commercial activity is focused on North American leagues, teams and stadiums. Investors are based in Toronto, Montreal and the United States.

History

The company was founded in 2024 and relocated from San Francisco to Toronto in 2025. In May 2025 it announced a partnership with the Quantum Sports and Learning Association to build what it describes as North America's first biomechanics basketball shooting lab in Toronto. It emerged from stealth in December 2025 alongside a $3.6M seed round led by Khosla Ventures (dated by one source to May 2025). In 2026 it joined the NBA Launchpad cohort and demonstrated its system to NBA and WNBA officials at Summer League, and on August 17, 2026 announced an additional $8.7M seed round co-led by Deloitte Ventures and Inovia Capital, bringing total disclosed funding to $12.5M.

Risks & controversies

Sources note that live sports viewing has declined for some sports, particularly among Gen Z, which is the demand backdrop the company targets, and that investors are often hesitant to back sports-related startups. The company has not publicly named any league or team customers, describing engagements as conversations, proofs-of-concept and evaluations. Camera-count requirements and install complexity remain an acknowledged constraint the team is working to reduce, along with system latency and reconstruction resolution. Reported figures are inconsistent in places: camera minimums are cited as 32 in one source and at least 36 in another, and the initial seed round is dated to December 2025 by most sources but to May 2025 by one.