Google DeepMind's Veo 3.1 Brings Vertical Video and Better Control to AI Generation
Google DeepMind released Veo 3.1 today, adding vertical video generation and improved consistency controls to its AI video model. The update signals Google's intent to capture the mobile-first creator market where competitors have struggled to gain traction.
The vertical video capability is the headline feature. While previous versions of Veo defaulted to landscape formats, 3.1 natively generates 9:16 content—the aspect ratio that dominates TikTok, Instagram Reels, and YouTube Shorts. For creators targeting these platforms, it removes a friction point that made AI-generated video impractical for social content.
What's Actually New in Veo 3.1
Beyond vertical support, Google claims improvements across three dimensions: consistency, creativity, and control. The "consistency" piece matters most technically. Earlier video generation models suffered from temporal coherence issues—objects would morph between frames, physics would break mid-clip. DeepMind says 3.1 generates "lively, dynamic clips that feel natural and engaging," suggesting they've made progress on maintaining object and motion continuity.
The "control" improvements likely reference better prompt adherence. A persistent challenge in video generation has been getting models to follow specific compositional instructions while maintaining quality. More control means creators spend less time regenerating failed outputs.
The Competitive Landscape
This release lands amid an intensifying race in AI video generation. OpenAI's Sora captured headlines in early 2024 but has rolled out slowly, with limited access frustrating potential users. Runway continues iterating on Gen-3, positioning itself as the professional-grade option. Kling from Kuaishou has gained traction in Asia with competitive quality at lower costs.
Google's advantage is distribution. Veo integrates with the broader Google ecosystem—YouTube, Google Cloud, and potentially Gemini. If a creator can generate vertical video directly within tools they already use, switching costs to competitors increase substantially.
The vertical video focus also reveals strategic thinking about where AI video actually gets used. Most commercial applications today aren't Hollywood-style productions—they're social clips, ads, and short-form content. Optimizing for that use case is pragmatic rather than glamorous.
What This Means for Creators
The practical implications are straightforward: creators targeting mobile platforms now have a first-party Google tool that speaks their format natively. No upscaling, no cropping, no workarounds. Combined with better consistency, the gap between "AI-generated" and "usable in production" continues to narrow.
For the broader industry, Veo 3.1 represents another step toward commoditization. When multiple companies offer high-quality video generation, differentiation shifts to integration, cost, and specialized capabilities. Google is betting that its ecosystem advantages will matter more than any single technical edge.
The takeaway: Google isn't trying to win on pure video quality alone. They're building the most convenient path from idea to published content—and vertical video support suggests they understand where that content actually lives.