How AWS Harnesses ML for Real-Time Super Bowl Analysis

Ever wonder how AI tracks and measures data at the Super Bowl?

Well, it’s not just tallying touchdowns and yards anymore. Modern analytics capture player location, speed, acceleration and dozens of advanced metrics on every single play.

At the NFL's annual championship game, Amazon Web Services (AWS) powers a vast computational system that transforms every player movement into immediate insight through machine learning and real-time analytics.

Every game of the NFL season – including the Super Bowl – generates millions of positional data points from Radio-Frequency Identification (RFID) chips embedded in player pads and footballs.

More than 20 ultra-wideband receivers track each of the 22 players 10 times a second and the ball itself at 25 times per second.

The system's core, Next Gen Stats (NGS), converts that torrent of telemetry into decision-grade analytics in under a second.

“Football, for 100-plus years, has been a box score game: you've got yards, you've got touchdowns, you've got tackles,” says Mike Band, Senior Manager of Research and Analytics with NFL's Next Gen Stats in conversation with AWS.

But just as the game has grown, so too has the tech that surrounds it.

Machine learning transforms football analytics

Back in 2018, when the NFL and AWS expanded a partnership that shifted tracking from a tactical tool into a strategic one.

That's when the league introduced completion probability – an ML metric built with Amazon SageMaker using XGBoost models.

The model considers factors such as pressure, receiver separation and throw depth to calculate a single percentage score for each pass attempt.

“That became our entry point into machine learning,” Mike says.

This has helped the NFL turn raw tracking into a real-time product.

By deploying AWS' infrastructure – spanning SageMaker, AWS Lambda and Amazon QuickSight – the NFL's analytics backbone has drastically evolved.

At the Super Bowl, these tools provide broadcasters, coaches, and analysts with feedback in under a second, enabling both visual and statistical decoding of each play for a global audience.

Mike estimates that Next Gen Stats now produces between 500 and 1,000 stats per play.

AWS says that keeping the system responsive depends on the provider's infrastructure to ingest the feed, run the models, return results within seconds for teams and broadcasters and store the wider data trove for deeper analysis.

AI-driven insights improve player safety

At the league level, this capability extends beyond entertainment.

Next Gen Stats models inform player safety policies, officiating reviews and even rule changes.

For instance, the dynamic kickoff introduced in the 2024 season stemmed from NGS analysis that quantified the risks of high-speed collisions.

“The season before, we were showing Next Gen Stats animations of the space and relative speeds of the players and that analysis became a critical part of the rules change,” says Mike Lopez, Senior Director of NFL Football Data and Analytics.

Powered by data, the redesign delivered a 35% drop in lower-extremity injuries, while driving more returns and a more dynamic on-field product.

Two seasons of data show the dynamic kickoff is working: the 2025 return rate jumped to 75% from 32% in 2024 and, even with 1,157 more plays, lower-extremity injuries dropped 35% whilst concussion rates remain below the old kickoff format.

Real-time processing powers decision-making

AWS’s impact now stretches far beyond the server room. New optical tracking systems – deployed in every NFL stadium – use 4K cameras to capture full three-dimensional skeletal movement, generating precise digital models of each player.

“The explosion in the volume of data can be daunting,” says Dashiell Flynn, AWS' Principal Sports Consultant. “But once folks wrap their heads around it, the ideas start flowing very quickly.”

Each stadium is equipped with on-site AWS servers that process data in roughly 700 milliseconds.

That data is then transmitted to the cloud, where machine learning models execute in under 100 milliseconds and deliver insights back to the production team – keeping the entire capture-to-analysis loop under one second.

For AWS, the Super Bowl has evolved into a yearly proving ground for real-time, enterprise-scale AI.

A decade into the partnership, the NFL’s data infrastructure now resembles less a set of servers and dashboards and more a living neural network – one that learns, predicts and adapts with every play.