How AWS-powered Next Gen Stats changed the NFL forever

This Sunday will see the Seattle Seahawks face off against the New England Patriots in Super Bowl LX.

The game will also mark the conclusion of the tenth football season featuring Next Gen Stats, the analytics system that delivers detailed data about every game to coaches and broadcasters through a partnership with Amazon Web Services. Next Gen Stats began in 2015, when the National Football League deployed RFID chips in player shoulder pads and even in the football itself, enabling the league to capture location data multiple times per second through sensors installed throughout stadiums.

It has since become a mainstay of football broadcasts and training sessions, delivering granular insights to a sport that previously could track only a fraction of the complex movements of 22 players and the ball across the field.  

“Next Gen Stats is part of the vernacular now,” says Julie Souza, AWS’s global head of sports. 

Bringing data to the gridiron

Behind the scenes, dozens of machine learning models—the same kinds of systems AWS offers to process business data—translate the raw numbers generated by the sensors into understandable stats in real time. With the recent addition of 4K cameras to NFL venues, the system can now capture not just player position on the field but the precise position of shoulders, elbows, knees, and hands, generating 29 data points per player 60 times per second. That data is processed by in-stadium AWS servers in roughly 700 milliseconds, then sent to the cloud to feed machine learning models that run in under 100 milliseconds. The result is analytics delivered to broadcasters within about a second, shorter than the NFL’s typical broadcast delay.

Announcers are equipped with dashboards that surface key stats, along with AI systems that allow them to ask natural-language questions based on new and historical data, Souza says, such as, “When was the last time this particular play happened, or that you know, this metric was achieved?” The data is also increasingly used to inform player coaching and off-the-field training, as well as rule changes designed to make the game safer.

AWS helped the NFL run thousands of simulated football seasons that informed the Dynamic Kickoff rule, introduced in the 2025 season. The change helped boost returned kickoffs while reducing the play’s historically elevated concussion rate. “What’s amazing about that is everything that we had modeled for them is what has panned out from the results,” Souza says.

Analytic dashboards also help teams identify players at risk of injury, allowing coaching and training staffs to intervene before injuries occur. Those changes in play and training led to roughly 700 fewer missed games by players last season, she says.

More detailed stats can also help newer fans, including international audiences and younger viewers, understand the game more quickly. Richer player data has enabled new types of broadcasts as well, including animated versions of real games that appeal to families with children, and Amazon Prime Video’s Prime Vision with Next Gen Stats stream of Thursday Night Football. Features tested in the Prime Vision stream, such as highlighting players likely to blitz the quarterback, have since made their way into the main broadcast.

“You can do all of these different versions of broadcast to serve different and specific audiences, but it’s all coming from that same set of data,” says Souza.

A different kind of bowl game

Next Gen Stats data is also used in the NFL’s annual Big Data Bowl, an analytics competition that invites contestants to develop new use cases for the league’s vast trove of data, and in some cases leads to jobs with the NFL or individual teams. Souza, who has served as a judge in the competition, says new judging criteria are being added to evaluate how proposed analytics could be conveyed to fans during a broadcast. The shift reflects a broader recognition that even as sports become more driven by data, storytelling remains central.

“Everything we’re talking about right now is the science—the science, and the engineering, and the analytics, and the rigor, and the math,” she says. “It only matters if the art is there, and the art is the storytelling.” 

View the full article