Exclusive: How Major League Baseball Teams Are Using AI Analytics

At a half-empty ballpark during a late-season game, there’s a moment when you start to notice things you didn’t notice before. Between batters, a pitching coach is constantly checking an iPad. A scout in the press box is staring at a laptop screen rather than the field. Instead of chewing sunflower seeds while leaning on the dugout rail, a manager is engrossed in a private, passionate discussion with someone who is holding a tablet. Despite its long history, baseball is subtly evolving into something else.

The majority of fans are unaware of the extent to which Major League Baseball teams are utilizing AI analytics, and the divide between teams that have made the commitment and those that are still on the fence is widening every season. As most people are aware, it began with Moneyball, the Oakland Athletics’ well-known experiment in the early 2000s to replace traditional gut feeling with objective data. Decades earlier, Bill James had established the foundation for statistics. The book was written by Michael Lewis. Brad Pitt gave it a stylish appearance. However, the current state of affairs makes those early analytics initiatives seem almost antiquated. Teams were debating on-base percentage back then. In order to detect stress fractures in mechanics before the pitcher experiences any pain, machine learning models are currently fed seventeen distinct biomechanical variables on a pitcher’s delivery.

It’s important to take a moment to recognize how much raw data a single baseball game generates, and the infrastructure supporting all of this is astounding. About seven terabytes of data are produced per game by Statcast, the tracking system that Major League Baseball implemented in all 30 ballparks in 2015. Each and every pitch. each swing. The initial action of each fielder on a fly ball. The precise spin rate of a curveball, the hard-hit liner’s launch angle, and the bat’s exit velocity at the exact millisecond of contact. 7 terabytes. for each game. No human analytics team, regardless of size, could effectively handle that volume in real time. At this point, the machines take control and things start to get really interesting.

The Miami Marlins have emerged as a quiet case study of how a smaller-market team is utilizing AI to catch up to larger-budget teams. The organization records high-frame-rate footage of every pitch, swing, and sprint from both their major league roster and their minor league affiliates using vendors like Hawk-Eye, KinaTrax, and Trackman. According to the team’s vice president of baseball systems, Brian Chase, they use this information to determine players’ current and potential future selves. That’s more than just a polite business expression. It describes a truly different approach to roster construction, where a 22-year-old prospect in the minor leagues is monitored not only by his performance but also by whether his mechanics are trending toward injury or durability. Development projections for three to five years. The models are generating that.

Naturally, the financial stakes are highest when it comes to injuries. MLB teams frequently spend $20 to $30 million, and occasionally much more, on individual player contracts. Additionally, for the majority of baseball’s history, injuries were handled reactively, meaning that you had to wait until something broke before attempting to fix it. It turns out that biomechanical stress builds up in ways that are not visible to the unaided eye. Over the course of a season, a pitcher’s arm angle may change almost imperceptibly, weeks before it becomes an issue. These days, AI systems are able to identify those deviations in real time, highlighting red flags that a human coach would miss from the dugout. Although it’s still unclear if this will result in quantifiably longer careers and fewer Tommy John surgeries on a large scale, the early signs are positive enough for teams to continue investing.

Subtle but cumulative changes have also occurred in the decision-making process on the field itself. These days, during games, managers refer to win-probability models. Before making pitching adjustments in the seventh inning, they examine matchup data, which includes the batter’s performance against left-handed pitchers with high spin rates, night games, and ballparks with particular dimensions. It does not take the place of human judgment. The top teams appear to have a clear understanding of that. It does this by providing a stronger basis for human judgment. Real-time probabilistic analysis in conjunction with an experienced manager’s intuition is most likely superior to either one acting alone.

Additionally, some of the most commercially intriguing AI research is taking place in the fan experience dimension. MLB’s increased collaboration with Adobe, which was revealed prior to the 2026 season, offers insight into the league’s future business goals. In order to customize content for fans at scale, the agreement makes use of Adobe’s generative AI tools, such as GenStudio for Performance Marketing and their Firefly creative platform. offers in real time while playing games. creative variations that are consistent with the brand across digital platforms. An LLM optimizer makes sure the league appears correctly in chatbots and AI-driven search interfaces. It sounds abstract until you understand what it really means: the league is attempting to use AI to make each fan’s digital connection to baseball feel more unique and less generic, rather than just to help teams win games.

Then there’s the robo-umpire controversy, which has sparked more discussion in baseball circles than nearly anything else in recent memory. After years of testing in minor leagues, the Automated Ball-Strike Challenge System will be implemented league-wide in 2026. It tracks each pitch with an accuracy of one-fifth of an inch using twelve synchronized Hawk-Eye cameras. Although it seems right to keep human umpires on the field for the time being, players can contest dubious ball-and-strike calls, and the AI system will make a decision in about fifteen seconds. 69% of fans expressed a preference for the system over conventional human-only officiating during spring training testing. It’s difficult to overlook that substantial figure.

It felt like a proof-of-concept moment when the Oakland Ballers became the first professional sports team to have an AI system handle every in-game strategic decision, from lineup construction to defensive positioning to substitutions, during the September 2025 game at Oakland’s Raimondi Park. They prevailed. The experiment opened a door that is difficult to close, whether that was due to the AI or not. For years, MLB teams have been utilizing AI in less obvious ways. There is a different kind of visibility and pressure on teams to determine how much they trust the machine when it appears in the dugout in real time and makes calls that coaches can see and react to.

It’s still unclear if the teams making the biggest investments in AI analytics are creating long-term advantages or just raising the league’s overall standard, which would require everyone to keep up in order to remain competitive. Eventually, it happened with Moneyball; within ten years, what seemed like a secret weapon in Oakland became standard procedure throughout the sport. Predictive scouting models, biomechanical tracking systems, and the current generation of machine learning tools all have a good chance of doing the same thing. Everyone catches up. The edge disappears. The next edge is then discovered by someone.

The data infrastructure itself is more difficult to envision going back. Seven terabytes of Statcast data, fifteen million data points per game, and high-frame-rate biomechanical video passing through the minor league system are all things that will never go away. Baseball has developed a real-time nervous system that continuously feeds information upward through all organizational levels. The sport is now permanently altered in ways that are largely undetectable to the casual fan watching from the stands, regardless of what happens with the competitive dynamics. The scout gazing at a laptop screen rather than the field and the iPad in the dugout are no longer curiosities. They are the foundation of contemporary baseball.