For decades, sports fans have relied on post-match reviews and expert panels to break down all the action. But what if that analysis could appear while the match is still being played?
This is what SportVot, a Mumbai-based sports streaming startup, is leveraging.
Founded in 2019, the platform allows sports tournaments, especially those outside mainstream leagues, to stream matches live while also generating automated analytics, performance insights, and graphical overlays in near real time.
At its core, the platform aims to simplify live sports production for tournaments that often lack the budgets or infrastructure of major broadcasters.
“SportVot simplifies live video production for lower tiers of sports and underserved sports tournaments that don’t have access to mainstream media or big broadcast channels,” Sidhhant Agarwal, Co-founder and CEO of SportVot, tells AIM.
“Our goal is to help these tournaments get visibility while also giving them the same level of analytics and production quality that you typically see in elite sports broadcasting.”
Unlike traditional sports broadcasts that require expensive equipment, large production crews, and dedicated infrastructure, SportVot’s system relies heavily on cloud-based video production and automated processing pipelines.
“AI-driven approach to sports production and analytics brings efficiency and scalability to an industry that has traditionally been complex and expensive,” says David Sargeant, who has invested in the startup through his sports-focused investment management firm, sporteva.
Proprietary Tech
The platform’s architecture is designed around three major layers: video capture, video processing, and analytics.
The video capture module, developed in-house, handles ingestion of match footage and is built on a combination of Java and Python-based modules where most of the platform’s business logic resides.
Once captured, the video moves into the processing layer, where the live feed is enhanced with graphics, commentary overlays, advertisements, and scorecards. “We add graphics, commentary, sponsor visibility, and even multiple camera angles,” Agarwal explains.
This layer uses C++-based video processing libraries and computer vision tools to manipulate the stream before delivering it to viewers. The system then distributes the content through SportVot’s platform and other channels, enabling fans to watch matches with real-time overlays that resemble professional sports broadcasts.
The company’s most ambitious feature is its AI-driven analytics engine, which processes video footage to identify key moments and generate insights automatically.
SportVot, with a total team of 61 (nine in tech), combines in-house tools with external AI models, including technologies based on Google Gemini and Segment Anything Model (SAM), a foundational computer vision model by Meta AI. These models allow the system to detect actions in the game and convert visual signals into structured data. “We train them further using our own sports data to get the outcomes we need,” the co-founder adds.
The analytics engine identifies match events, segments footage into highlights, and generates textual outputs that can be turned into graphics and statistics for viewers. This data pipeline feeds into the video processing layer, which displays visual overlays such as player statistics or key match moments during the live stream.
However, Agarwal is cautious about overstating the platform’s reliance on AI. “Some things are machine learning, some are computer vision, and only some are truly generative AI. We try to avoid using the term loosely and instead focus on where the technology genuinely adds value.”
The Challenge of Live Analysis
Delivering analytics during a live sports match is technically demanding, primarily due to the time required for AI inferencing. While post-match analysis can process hours of footage after the game ends, live systems must work within seconds. “The biggest challenge is latency—how quickly you can process the footage and push the insights back to the viewer,” Agarwal notes.
To tackle this problem, SportVot relies heavily on cloud computing infrastructure, where most of the video processing and analytics workloads currently run. However, the company is increasingly experimenting with edge computing to reduce delays. “Right now, about 80–90% of what we do still runs on the cloud, but we are moving certain compute workloads to the edge to reduce latency.”
Processing some workloads closer to where the video is captured reduces the need to send large video files to distant servers for analysis. This hybrid approach allows the platform to deliver analytics within 20–30 seconds, particularly for features like replay systems and review tools.
One example is the Decision Review System used in professional cricket. Traditionally, DRS requires sophisticated camera setups and expensive ball-tracking systems that only top-tier tournaments can afford.
“The platform is trying to create a more accessible alternative. “We are trying to build a much more accessible review system so even smaller cricket tournaments can use technology to review decisions,” Agarwal remarks.
Democratising Sports
SportVot’s focus on underserved sports ecosystems has allowed it to capture a significant audience across regional tournaments and amateur leagues. One of the platform’s largest events, the 72nd Senior National Kabaddi Championship 2026, generated between 15–16 million viewers during the tournament.
Overall, the platform sees around 20–25 million views every month across its streaming channels. Content distribution is also multi-channel. While live matches are hosted on SportVot’s platform, highlights and video-on-demand content are often distributed through YouTube to reach wider audiences.
“Olympic sport sailing is challenging to broadcast or live stream, as everything unfolds across open waters with constantly changing wind and wave conditions. SportVot’s production, broadcast, and content support have helped us bring races to audiences in a far more engaging way,” add Udya Bhan Singh Yadav and Sarla Yadav, Directors, UDAY Sailing Foundation, dedicated to promoting sailing in India.
SportVot operates on a twin revenue model.
Sports federations, leagues, and tournament organisers pay the company to stream matches on a usage-based pricing system. On the consumer side, viewers can watch some content for free while premium matches are behind paywalls.
“The platform enables practical pathways for monetisation and commercialisation, which are essential for building a more sustainable sports ecosystem globally at all tiers," Sargeant adds.
The startup last raised ₹9.4 crore ($1.1 million) in February 2024 in a Pre-Series A round led by Omidyar Network India, with participation from Anicut Capital, StartupLynk, and Ventana Ventures, among others.
While India remains a core market, SportVot is increasingly looking overseas. The company has already begun building a presence in Australia and is exploring opportunities in West Europe, where amateur and semi-professional sports have strong participation but limited broadcast coverage.
“India is still an evolving sports ecosystem, but globally, there are markets where grassroots sports are already very structured,” Agarwal notes.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.