OpenAI Is Breaking New Ground With Sora. Why the AI Video App Is Bad News for Meta.

Sora, a new iPhone app, uses OpenAI’s latest AI video model. (Dreamstime)

This article is from the weekly Barron’s Tech email newsletter. Sign up here to get it delivered directly to your inbox.

Slop Wars. Hi everyone. Let the AI-generated user video battle begin.

Last week, when Meta Platforms introduced the concept of user-generated AI short videos in a feed called Vibes, my initial reaction was negative. Upon checking it out, I found the videos to be simplistic computer animations with music soundtracks, which didn’t seem necessary.

My perspective shifted when I tried OpenAI’s version. The AI start-up released a new iOS app called Sora on Tuesday afternoon, powered by its latest flagship AI video generation model Sora 2. The app is a near carbon copy of TikTok but exclusively features user-generated AI videos. OpenAI has made the app invite-only to manage growth with each new user receiving a handful of invites to share with friends.

Sora has taken the internet by storm, becoming the most talked-about social app in years. By this afternoon, it was the No. 3 most downloaded app on Apple’s App store.

With a simple text prompt, users can create 10-second videos with accurate real-world physics and fully synchronized dialogue, singing, and sound effects. We’re seeing the future of social media roll out in front of our eyes.

Upon opening Sora, users record themselves saying a few numbers and turning their head left and right. The app can then create videos using the user’s likeness. With permission, that likeness can then be used by friends in their own videos. The app has been overwhelmed with demand with users seeing messages about capacity constraints when trying to make new videos.

I asked the app to make a video using my likeness to talk about Nvidia . Sora created a video of me delivering a coherent message with words I have never said. It was indistinguishable from reality. As scary as that sounds, OpenAI has built in protections. I have to approve any usage of my likeness.

“This feels to many of us like the ‘ChatGPT for creativity’ moment, and it feels fun and new,” OpenAI CEO Sam Altman said in a blog post. “There is something great about making it really easy and fast to go from idea to result, and the new social dynamics that emerge.”

The innovation here is how Sora lowers the barrier to storytelling. Any user can become a screenwriter and create realistic, captivating videos. There is no limit to one’s imagination. Already, I’ve seen numerous entertaining memes and skits on my feed using the likeness of Altman and the Pokémon character Pikachu (although I don’t expect that to last).

It is so engaging that I worry about a dystopian future where we spend an inordinate amount of time mindlessly swiping through 10-second videos on our smartphones for AI-optimized dopamine hits.

The video-generation technology will only get better from here, with improved detail and fidelity. I can foresee a compelling AI-optimized feed stealing attention from real-world videos on Instagram and TikTok. If it captures a significant share of the attention economy—even 20% or 30%—there would be negative ramifications for viewership and ad revenue on those two apps.

Sora’s advancement and breakout success validates Meta CEO Mark Zuckerberg’s paranoia about the need to invest in AI or risk disruption. Traditional market leadership means much less when AI advancements can quickly upend business models and competitive positioning.

Meta will fight back. Zuckerberg has vowed to spend at least $600 billion on investments through 2028; Meta says it can’t afford to be left behind in advanced AI technology.

But I’m not sure money alone will be enough to catch up with OpenAI’s AI research prowess. Sora shows that OpenAI is already way ahead in video-generation abilities.

Investors seem to be taking notice. Meta shares are down about 2% since the Sora app was released Tuesday afternoon, materially underperforming the general market.

Meanwhile, this latest AI trend is more good news for Nvidia. AI generated video doesn’t come cheap. As the world moves toward more AI video generation it will require more AI computing capacity, which primarily runs on Nvidia AI chips.

This Week in Barron’s Tech

• Anthropic’s Claude Sonnet 4.5 Is a Game-Changer
• OpenAI Says Users Can Now Shop for Etsy Products on ChatGPT
• AppLovin Stock Rides Wave of Price-Target Hikes to a Record High
• CoreWeave Stock Climbs on Meta Deal. Why Two Is Better Than One.
• ASML Builds Machines that Make AI Chips. Why It’s Missing Out on the Boom.

Write to Tae Kim at tae.kim@barrons.com or follow him on X at @firstadopter.