xAI, spearheaded by Elon Musk, recently announced Grok users will soon gain the ability to generate AI-driven videos, incorporating sound, directly within the application. This upcoming functionality, branded as Imagine, empowers users to produce both static images and dynamic videos from textual prompts, as well as transform existing images into video content.
Download the @Grok app and subscribe to try out our video generation beta and Valentine/Ani
— Elon Musk (@elonmusk) July 28, 2025
Currently, access to Imagine appears to be restricted to a select group of xAI employees and chosen influencers. The official @Grok X account posted a statement confirming the phased rollout: “Video generation is coming to Grok via our Imagine feature, powered by Aurora. Create instant videos with sound from text prompts. Download the standalone Grok app, subscribe to SuperGrok, and join the waitlist for early access in October.” This indicates a tiered access system, with wider availability anticipated in the near future.
Excited to confirm: Video generation is coming to Grok via our Imagine feature, powered by Aurora. Create instant videos with sound from text prompts. Download the standalone Grok app, subscribe to SuperGrok, and join the waitlist for early access in October. Let’s make some…
— Grok (@grok) July 29, 2025
Access to the Imagine feature is tied to a premium subscription tier. SuperGrok, the paid tier required for early access, carries a monthly fee of $30. This subscription model suggests xAI’s intention to monetize the advanced AI video generation capabilities.
Prior to the formal announcement, Elon Musk alluded to the impending feature, teasing the “bringing back Vine, but in AI form.” This statement references the once-popular social media platform, Vine, known for its short-form video content, which was discontinued by Twitter in 2017. The reference suggests a potential focus on brevity and shareability for Grok’s AI-generated videos.
We’re bringing back Vine, but in AI form
— Elon Musk (@elonmusk) July 24, 2025
An xAI employee further clarified the connection to Vine by stating that Grok Imagine videos are capped at six seconds in length, mirroring the video duration that defined Vine. This limitation points toward a design philosophy centered on creating concise, easily digestible video content.
Based on shared material from xAI employees and users who have been granted early access, the Imagine feature operates through a text-prompt interface. Users input descriptions of their desired video content, and the app then generates multiple video options reflecting those descriptions. The user can then select their preferred video from the generated results.
Illustrative examples of the feature’s capabilities have been disseminated online. One xAI employee showcased an AI-generated video depicting a cat purring while traversing space. Another demonstrated an astronaut navigating an alien landscape. These examples serve to highlight the potential creative applications of the Imagine feature.
Beyond the conventional applications, an xAI employee indicated the existence of more provocative possibilities within the Imagine feature. Mati Roy stated on X that “Grok Imagine videos have a spicy mode that can do nudity.” Roy further elaborated that the technology is capable of producing realistic depictions of human forms, potentially leading to the creation of explicit content.
Prior to the unveiling of the Imagine feature, Grok had already faced scrutiny regarding its AI “companions.” These AI characters, designed for user interaction through text and speech, have been criticized for exhibiting vulgar and sexually explicit behavior. The addition of the new companion named “Valentine,” further intensified these concerns.
xAI is investing heavily in computational infrastructure to support its AI development efforts. The company is adding 110,000 graphics processing units (GPUs) to its supercomputer, Collosus, located in Memphis, Tennessee. These specialized circuits are designed to accelerate the generation of computer graphics, facilitating the development of more sophisticated AI models.
