Amazon just rolled out an AI-powered Video Generator that turns a single product image into 6-to-9-second video ads in about five minutes. It creates four 720p variations at no extra charge, complete with customizable text and branding. Why it matters: It lowers the barrier for e-commerce video ads, though access is currently limited to select U.S. advertisers.
Alibaba Cloud just released over 100 open-source Qwen 2.5 models, with sizes ranging from 0.5 to 72 billion parameters and sharper math and coding skills. The company also unveiled a new text-to-video AI model under its Tongyi Wanxiang family, alongside an upgraded Qwen2-VL vision model. It marks one of the biggest open-source AI drops from a major tech giant to date.
Kling AI just launched its Kling 1.5 model, bringing 1080p HD generation to professional mode. The update also adds a new Motion Brush feature to give creators direct control over visual motion, making AI video significantly more steerable.
Runway just partnered with Lionsgate to train a custom AI model on the studio's proprietary catalog. The tool generates cinematic footage designed to help filmmakers iterate during pre-production and post-production. Why it matters: Studios are officially moving from copyright disputes to actively licensing their archives to train tailor-made generative video models.
California Governor Gavin Newsom just signed two new bills, AB 1836 and AB 2602, making it illegal to use AI-generated digital replicas of performers without explicit consent. The rules protect both living actors and deceased individuals, effectively turning SAG-AFTRA's recent strike victories into statewide law. Why it matters: It sets the strictest legal guardrails in the nation against Hollywood replacing human talent with unauthorized digital clones.
Snap just announced its own AI video generator at the Snap Partner Summit, powered by its in-house foundational video models. Select creators can now generate clips from text prompts on the web, with image prompts coming soon. All generated videos include watermarks, but Snap hasn't shared any actual output examples yet.
Luma AI just launched the Dream Machine API, giving developers direct programmatic access to build video generation into their own creative apps. In a spicy coincidence, rival Runway announced its own developer API on the exact same day. The generative video battle is officially moving straight into developer workflows.
Google and Tel Aviv University researchers just unveiled GameNGen, a neural model that simulates the 1993 classic Doom in real time at over 20 frames per second using Stable Diffusion. In blind tests, human raters struggled to distinguish the generated video from the real game. It points to a future where games aren't programmed line by line, but hallucinated on the fly by neural networks.
Blackmagic Design just officially released DaVinci Resolve 19 alongside its iPad update. The release packs in over 100 upgrades and new AI features, including IntelliTrack AI tracking, text-based speaker editing, and Fairlight AI audio panning to video. It brings automated heavy lifting straight to your video editing timeline.
D-ID just launched AI Video Translate, a new feature powered by its proprietary Rosetta-1 model that translates videos across 30 languages while cloning the speaker's voice and adjusting lip movements to match. Currently available to subscribers with a one-month trial, it handles single-speaker clips from 10 seconds to 5 minutes to help creators localize content on a budget.
a16z just released the third edition of its Top 100 Gen AI Consumer Apps, revealing that nearly 30% of companies on the list are newcomers. Video and music generators saw massive surges, with Suno jumping from #36 to #5, while Perplexity and Claude cemented top-five spots on web. ByteDance quietly dominated with six apps across web and mobile rankings.
AI filmmaking platform LTX Studio just dropped its waitlist and opened to the general public, rolling out several notable creative tools. Users can now upload gestures to give characters human-like motion capture or turn rough scribbles into video scenes. It brings full-pipeline cinematic AI generation to anyone without waiting.
AI startup Hotshot just released a text-to-video model that generates up to 10 seconds of 720p footage without watermarks. Remarkably trained in four months by a four-person team using thousands of GPUs, it aims to rival big players in video generation. The free web tier offers two generations per day.
Nvidia just introduced its first on-device small language model, Nemotron-4 4B Instruct, to its ACE digital human suite at Gamescom. Distilled from a 15B model, it is optimized to run locally on RTX GPUs with minimal memory and fast response times. Why it matters: Gamers can now talk to dynamic NPCs in games like Mecha BREAK entirely on local hardware, without relying on cloud servers.
Luma AI just launched Dream Machine 1.5, upgrading its public video generator just two months after debut. The standout improvement is clean in-video text rendering alongside smarter prompt comprehension for non-English inputs. It generates a 5-second clip in about two minutes, giving creators a practical way to animate titles and dynamic logos.
Stability AI just appointed entertainment technology veteran Hanno Basse as its new Chief Technology Officer. Basse brings over 30 years of deep industry experience, previously serving as CTO at Digital Domain, Microsoft Azure Media and Entertainment, and 20th Century Fox. The hire signals Stability AI is doubling down on integrating generative AI directly into professional studio production workflows.
Hedra just dropped Character-1.5 alongside a new Stylize tool for its AI avatar generator. The updated model brings noticeably more natural head movements, facial expressions, and tighter lip-syncing across photos, illustrations, and figurines. Stylize lets creators quickly swap outfits, tweak backgrounds, or transform characters into anime and Lego styles.
Tavus just launched its Conversational Video Interface, enabling developers to build digital twins that can see, hear, and talk in real time with under one second of latency. The setup is fully white-labeled and modular, letting developers swap out their own LLM and TTS components. It basically brings ultra-low-latency, face-to-face video AI agents to any app.
Runway just launched image-to-video for Gen-3 Alpha Turbo, generating video clips seven times faster and for half the price of the original model while maintaining comparable quality. It is now open across all plans, with free users also getting trial access.
Rendora just launched a web platform that turns written scripts into fully animated 3D avatar videos in seconds. Users can pick characters and environments to generate footage complete with realistic lighting and cinematic angles, all fully customizable afterwards. It basically cuts out the need for real actors or complex 3D animation pipelines.
