The week of March 23–29 will be remembered as the week the AI video landscape shifted beneath everyone’s feet. OpenAI stunned the industry by shutting down Sora on March 25, collapsing a billion-dollar Disney partnership in the process. Meanwhile, SAG-AFTRA floated a radical new proposal to tax studios for using AI-generated performers, Oscar-winning filmmakers premiered a landmark AI documentary, and the post-Sora market began rapidly reorganizing around new frontrunners.

OpenAI Pulls the Plug on Sora — and a Billion-Dollar Disney Deal

The biggest story of the week — and arguably of the year so far in AI video — landed on Tuesday, March 25, when OpenAI announced it was shutting down Sora, its flagship video generation platform, just six months after its splashy public launch. CEO Sam Altman made the call to kill the product, free up compute resources, and refocus the company’s efforts elsewhere. The financial reality behind the decision was staggering: Sora had been burning through roughly a million dollars per day in inference costs while generating only $2.1 million in total lifetime revenue. Its user base, which peaked at around one million, had collapsed to fewer than 500,000 active users.

The collateral damage was immediate and significant. Disney, which had committed $1 billion to a partnership that would have licensed more than 200 iconic characters for use in AI-generated short videos, learned of the shutdown less than an hour before the public announcement. The deal evaporated overnight. As TechCrunch noted, Sora’s demise may serve as a broader reality check for the entire AI video sector — a reminder that generating impressive demos and building a sustainable business are two very different things.

SAG-AFTRA Proposes the ‘Tilly Tax’ on AI Performers

On Saturday, March 28, Bloomberg and Fortune reported that SAG-AFTRA is bargaining for what insiders have dubbed the “Tilly tax” — named after Tilly Norwood, the controversial AI-generated actress who has become a lightning rod in the debate over synthetic performers. The proposal would levy a fee on studios for every AI-generated or “synthetic” character used in a production, with the goal of making digital performers cost roughly as much as hiring real human actors. Revenue from the tax would flow into the union’s healthcare and pension funds.

SAG-AFTRA Executive Director Duncan Crabtree-Ireland, speaking at an AFL-CIO workers’ summit in Washington, called collective bargaining “the fastest and most effective way for the regulation of AI technology.” Not everyone in the union is fully satisfied with the approach — AI task force member Brendan Bradley acknowledged it’s “under the category of the best bad idea we’ve got in 2026” — but with the current contract expiring in June, it represents the union’s most concrete strategy yet for addressing the rising tide of AI-generated talent.

SAG-AFTRA Backs Trump Administration’s AI Policy Framework

In a related development earlier in the week, SAG-AFTRA publicly endorsed the Trump administration’s national policy framework for artificial intelligence. The union expressed strong support for the framework’s call for Congress to pass federal legislation against digital replica abuse while maintaining First Amendment safeguards. SAG-AFTRA specifically urged lawmakers to move swiftly on the bipartisan NO FAKES Act, which would establish federal protections against nonconsensual AI replications of individuals’ likenesses and voices. The endorsement signals an unusual alliance between the entertainment union and the current administration on the specific issue of AI worker protections.

Oscar Winners Premiere ‘The AI Doc’

On Monday, March 24, news broke that a team of Oscar-winning filmmakers had completed “The AI Doc: Or How I Became an Apocalyptimist,” a documentary nearly three years in the making. Co-directed by Daniel Roher (who won for “Navalny”) and Charlie Tyrell, and co-produced by Daniel Kwan (co-director of “Everything Everywhere All at Once”), the film attempts to zoom out from the daily headlines to offer audiences a more lasting examination of what’s at stake as artificial intelligence rapidly evolves.

The documentary features on-camera interviews with more than 40 subjects, including OpenAI’s Sam Altman, Anthropic’s Daniela and Dario Amodei, and Google DeepMind’s Demis Hassabis, resulting in over 3,300 pages of transcripts. In a pointed creative choice, the filmmakers opted for a deliberately anti-digital visual approach, incorporating handmade illustrations and stop-motion animation rather than AI-generated imagery. The film is being distributed by Focus Features.

The Post-Sora Market Reshuffles

With Sora gone, the AI video generation market is rapidly reorganizing around three main contenders. Google’s Veo 3.1 has emerged as the new benchmark for quality, generating synchronized audio — including ambient sound, dialogue, and sound effects — directly alongside video in a single pass, with true 4K output at up to 60fps. Kuaishou’s Kling 2.0, which launched globally with “Standard” and “Master” quality tiers, is positioning itself as the cost-efficient leader, delivering comparable quality at roughly 40% of the per-second cost. And Runway’s Gen-4.5 continues to dominate professional advertising and narrative work thanks to its superior temporal consistency and motion control.

Meanwhile, the open-source ecosystem continues to mature. Models like Wan 2.2, HunyuanVideo, and Lightricks’ LTX-Video are now producing results that rival commercial platforms, with LTX-Video capable of rendering 30fps video faster than real-time on an H100 GPU. For independent filmmakers and studios looking to build proprietary pipelines, these open models increasingly represent a viable — and far cheaper — alternative to API-based services.

AI Filmmaking Goes Global

Higgsfield AI released the results of what it called the largest AI filmmaking competition ever held, with 8,752 submissions from 139 countries competing for $500,000 in cash prizes. Perhaps most notably, India led all nations with 1,805 submissions — nearly double the United States’ count of 1,041 — with Germany, France, Italy, Brazil, and the United Kingdom rounding out the top territories. The data underscores a point that’s becoming harder to ignore: AI filmmaking is not a Silicon Valley phenomenon. It’s a global creative movement, and much of its energy is coming from outside traditional Western production centers.

Separately, the Hollywood Reporter profiled an emerging AI filmmaker known as “Gossip Goblin,” whose bleak sci-fi short films — generated and assembled entirely using AI tools — have attracted significant attention on social media. The profile posed a provocative question: at a moment when Hollywood and Silicon Valley are still arguing over what AI filmmaking actually is, creators like Gossip Goblin suggest it may already have a recognizable aesthetic of its own.

What March 23–29 Tells Us About Where AI Video Is Heading

This was a week that clarified the AI video landscape in ways that months of incremental updates had not. Sora’s shutdown proved that even the most hyped product from the most well-funded company in the space can’t survive without a viable business model. The “Tilly tax” proposal showed that labor is no longer just reacting to AI — it’s trying to shape the economic rules of the game. And the global filmmaking competition data, the Oscar-winning documentary, and the rise of AI-native creators all point to the same conclusion: the creative use of AI video tools is accelerating faster than the industry’s ability to govern it. The question is no longer whether AI will transform filmmaking. It’s who will set the terms.