Change the calculus of viral video

The NBA recently posted that its 2026 playoffs are “delivering their highest viewership through the Conference Semifinals in 29 years, with an average of 4.5 million viewers per game across ABC, ESPN, NBC/Peacock and Amazon Prime Video.” That is designed to sound impressive. Then you realize the viral video of a fan eating spaghetti directly behind legendary NBA commentator Mike Breen gained an estimated 20–50 million aggregate views and hundreds of thousands of likes across platforms within 24–48 hours.

This snapshot underscores just how much the rules of content are evolving. It’s no longer about watching an event, movie, or show start to finish. It’s what’s happening on the sidelines or in a specific moment—a reaction, a witty line, one small nugget that catches attention. 

In the world of traditional media though, these moments often languish in an archive somewhere, never to be seen or heard from again. This is a lost opportunity. 

CONSUMER DEMAND IS NECESSITATING CHANGE

Today, content is everywhere and it is active, in a state of perpetual motion. Viewer attention spans are short, with viewers often preferring reels, TikToks, and highlights over watching a full-length episode, movie, or game. You see it in ratings and box office challenges while the popularity of social content surges.

At the same time, archives have never been worth more. Streaming wars, brand storytelling, and sports highlights are driving value up. Additionally, organizations outside of sports and entertainment are seeking access to original, licensed, or proprietary material to train AI models (so much is now generated that training data is in short supply). In this landscape, simply owning content isn’t enough. You must know exactly what you have to use it for your own means or to sell it to someone else. Therein lies the problem. 

The most valuable content libraries in the world are the slowest to activate. Content owners have no idea of the entirety of their archives, let alone a good way to quickly call it up and share it.

As such, content owners aren’t being out-edited by other studios. They’re getting beaten by their own fans on TikTok and YouTube, using phones and a free editing app. Fans don’t have stores of assets, but they do have infinite velocity, and this trumps the content owners who can’t find the moments they need fast enough to make an impact, even if they are just trying to find a clip from the game last night. 

Why is this happening? Content owners have relied on manual tagging systems that were not built for a world where instant is everything. To maintain relevancy and leverage, it’s not just the clip from the game, show, or film that recently aired. It requires searching and pulling from decades’ worth of material in their possession. Media companies, networks, sports leagues, studios—the big players—need new tools and a fresh perspective.

EASE + ACCESS = OPPORTUNITY REALIZED

Something very specific happened when ChatGPT launched and it had nothing to do with the technology. The models had been there. Researchers knew what they could do. What changed was that any person could type something into a user interface and watch it respond. That moment of discovery made its potential real.

Video hasn’t had that moment yet, at least not on a similar scale. Content owners lack systems built around retrieving specific instances. That capability could change the calculus of content. 

Consider this. Someone asks a question in their own words. An agent searches massive archives, across multiple modalities: video, images, audio, text, and documents at once. This surfaces short, useable snippets of the most relevant content that could be rapidly deployed for audience engagement. 

Let’s build on the NBA example. What if someone used AI to search for “funny moments of fans eating or drinking behind commentators or players at a sports event?” All the clips in the NBA’s (or the network’s or team’s) library could be surfaced in seconds, rather than the hours someone would typically spend watching footage to find the right thing. 

Take it a step further. The creator asks their system to edit pulled content in different ways to make something new, interesting, and entertaining. The same archive then yields multiple stories, thereby compounding the content’s value.

Organizations could function at the cadence audiences have been waiting for, changing the relationship between an archive and audience.

A FORMULA FOR FUTURE ENGAGEMENT

The reality is this: There is more video being created right now than at any point in human history. Some of the most resonant video is sitting in archives with no AI video search, no video metadata automation, and no way to find what’s inside those archives, at the speed the audience expects. The math of who gets to tell which story changes when finding becomes free. This can impact studios, brands, sports rights holders, and independent makers. New kinds of stories become possible. Old kinds of stories get told by people who couldn’t tell them before. 

Those who figure this out will outpace the ones still waiting for the perfect tagging system. They are the ones who will define the next era of content.

Jae Lee is cofounder and CEO of TwelveLabs.