After racing to the top of the charts at launch, OpenAI’s Sora video app is showing signs of fatigue. Market intelligence from Appfigures indicates sequential declines in both downloads and consumer spending as the initial buzz around AI-generated social video cools.
Appfigures data shows downloads fell 32% month over month, then slid another 45% to about 1.2 million installs in the latest period. Lifetime installs across iOS and Android have reached roughly 9.6 million, with total in-app consumer spending of $1.4 million, $1.1 million of which comes from the U.S. Recent spending totaled $367,000, down from a prior peak of $540,000. In the U.S., Sora hovers near No. 101 among overall free iOS apps (No. 7 in Photo & Video), and about No. 181 on Google Play.
Early Spike and the Gravitational Pull of Novelty
Sora’s debut had all the ingredients of a viral hit: a coveted invite-only rollout, eye-popping AI video quality from the Sora 2 model, and an easy prompt-to-clip workflow. The app crossed 100,000 installs on day one despite access limits, and cleared the million-install mark faster than ChatGPT’s mobile app did at launch.
But generative media apps often face a steep novelty curve. Once users test the magic trick—turning a prompt into a cinematic snippet—habit formation gets harder. Social creation apps typically fight to keep weekly retention above the low-20% range, according to industry analyses from data.ai and Sensor Tower, and that challenge compounds when the product is iOS-first, invite-gated, and meters creation via paid credits.
Competition and the Mounting Content Constraints
Competitive pressure arrived quickly. Google’s Gemini app, including on-device Gemini Nano features, has tightened its grip on Android and pushed deeper into mobile workflows. Meta’s expanding set of AI creation tools, including Vibes-style video experiments, has given creators alternative playgrounds with built-in distribution across Facebook, Instagram, and Reels.
Sora’s content policy shifts also matter. Early on, the app’s permissive approach to using well-known characters helped fuel viral remixes, but it drew heavy criticism from rights holders. OpenAI pivoted from an opt-out to an opt-in licensing stance and added more restrictions, curbing some of the recognizable IP that supercharges sharing. A high-profile deal with Disney opened limited, approved use of its characters; however, Appfigures’ download and spending trends suggest that single-licensor content hasn’t meaningfully reversed the slowdown.
Business Model Pressure and the Cost of AI Video
The economics of AI video are unforgiving. Generating high-fidelity clips requires expensive inference, so Sora leans on paid credits and subscriptions to manage compute load. With consumer spending down 32% from the peak and recent monthly revenue at $367,000, the app faces a balancing act: keep creation fast and cheap enough to drive engagement without eroding margins.
Distribution headwinds compound the issue. Falling out of the Top 100 overall on the U.S. App Store reduces organic visibility, forcing greater reliance on paid acquisition and creator partnerships. For social apps, chart position and network effects are tightly linked: fewer new users means fewer remixes, which means less fresh content to bring the next cohort in.
What Could Reverse the Slide for Sora’s AI Video App
There is a path back. More opt-in licensing deals from major studios and game publishers would give creators safe, remixable worlds and characters—fuel for the virality loop without legal overhang. Stronger consent controls for likeness, default watermarking, and clearer provenance signals could improve trust for both users and partners while reducing moderation overhead.
On product, Sora could lean into collaboration and distribution: deeper integrations for exporting to TikTok, Instagram Reels, and YouTube Shorts; better editing timelines; asset libraries; and templates that lower the barrier to repeat creation. Android parity and a frictionless web editor would widen the funnel. Longer term, lightweight, on-device generation for short clips—paired with cloud for higher fidelity—could cut costs and keep sessions snappy.
For now, the numbers don’t spell doom, but they do signal a reality check. Sora proved that high-quality AI video can captivate mainstream users. To sustain that attention, OpenAI must turn spectacle into habit, secure the right content rails, and make creation feel less like a demo and more like a daily tool.