The Catch-Up Layer
Rapid Recap, Catch Up to Live, AI vertical cropping — every AI feature that shipped in sports streaming compresses time. The recommendation layer everyone spent a decade predicting never arrived. That is not a failure of the technology; it is a truth about live.
The AI that actually shipped in sports streaming answers “I’m late.” It does not answer “what should I watch.”
Run the inventory. Amazon’s NBA coverage — year one of an eleven-year deal — ships Rapid Recap, a roughly two-minute AI-generated catch-up that plays a late tuner everything they missed and then hands them into the live feed, alongside Key Moments, real-time AI highlight curation. ESPN’s app ships Catch Up to Live. Peacock spent this spring beta-testing AI that crops a live NBA broadcast into vertical video in real time. Different companies, different leagues, different stacks — the same product. Every deployed AI feature in a major sports app compresses time.
Now run the other inventory, the one from a decade of conference keynotes: the recommendation layer. The discovery engine that would learn your teams, your players, your tolerance for a blowout, and serve you the right game at the right moment. It mostly never arrived. The closest thing at scale is ESPN’s SC For You, an AI-personalized SportsCenter that launched in beta in August 2025 — and even that personalizes on the teams you already told it you follow, not on anything it discovered about you.
I don’t read that as a failure of the technology. I read it as the technology telling us something true about the product.
Catching up is a bounded problem with a clear reward. The system knows which game you opened, knows what you missed, knows where now is. Solve it and the payoff is immediate and legible: you are back in the room while it still counts. Recommendation on live sports, by contrast, fights the calendar. The game is on when it is on. A fan does not need an algorithm to discover that their team plays tonight — the schedule made that decision months ago, and the fan organized their evening around it. The product of live sports is synchrony, being in the moment with everyone else who cares, which is the whole reason it holds attention that nothing else can. An engine that suggests other games is solving a problem the fan does not have. An engine that gets them back to now is solving the only one they do.
The industry is starting to say this out loud. Adweek came back from Cannes last month with the temperature check in its headline: sports fans want better experiences, not better algorithms. Personalization earns its keep when it simplifies — when it removes the distance between opening the app and being inside the game — not when it curates.
A decade of product roadmaps promised Netflix for sports. Fans kept asking for a faster road back to live. The features that shipped are the ones that listened.