Amazon Web Services has launched AWS Elemental Inference, an AI tool that automatically reformats live video into vertical clips for social media in near real-time, aiming to help broadcasters capture viral moments as they happen. The move gives traditional media companies a new weapon in the battle for audience attention on mobile platforms.
The short-order cook: Instead of waiting for post-production, Elemental Inference applies AI during the video encoding process. This allows the service to re-crop a horizontal broadcast into a 9:16 vertical format with just 6- to 10-second latency. It’s built on an agentic AI model that operates autonomously, acting like a short-order cook that can spot key moments and create highlight clips on the fly without human intervention.
The vertical imperative: The tool addresses a major headache for media giants, whose efforts to reach audiences on platforms like TikTok and Instagram have been hampered by the slow, manual slog of editing. Early adopters include Fox Sports and NBCUniversal, with Fox noting that nearly 90% of its digital audience consumes content vertically.
Slashing the turnaround: For Fox, an early development partner, the AI tool has transformed that workflow. In an interview with Deadline, the company said it has slashed the turnaround time for creating a vertical highlight clip from over 45 minutes to less than 15, a dramatic increase in speed and content volume.
By automating the creation of vertical video, AWS is giving broadcasters a chance to compete with native social content on its own turf—speed. The "process once, optimize everywhere" model, available on a pay-as-you-go basis, turns a defensive scramble for relevance into a proactive content strategy. For a closer look at the announcement, AWS published its own in-depth overview on its news blog.
