Headline: Google slashes AI video costs with Veo 3.1 Lite — a potential win for crypto builders as OpenAI’s Sora shuts down
Google this week quietly redrew the economics of AI video for developers. Veo 3.1 Lite, delivered via the Gemini API, arrives at under half the price of Veo 3.1 Fast and brings generated video costs into ranges that are finally practical for high-volume use.
What’s new
- Veo 3.1 Lite supports Text-to-Video and Image-to-Video in both landscape (16:9) and portrait (9:16) formats, at 720p and 1080p.
- Video length options are 4, 6, or 8 seconds, with pricing that scales by duration.
- Google says Veo 3.1 Fast will also get a price cut on April 7, positioning three tiers so developers can choose based on need and budget.
Price perspective
- Previously, full Veo 3.1 cost roughly $0.40 per second of generated video with audio via the API. Veo 3.1 Fast ran about $0.15 per second.
- Veo 3.1 Lite brings the floor to $0.05 per second for 720p — that’s $3 per minute versus $24 per minute under the older $0.40/s pricing. This makes repetitive iteration and large-scale video features far more affordable for smaller teams and startups.
Hands-on notes
- Early testing showed impressive speed: an 8-second video generated in under one minute.
- Quality held up well against Veo 3.1 Fast, with only minor prompt-adherence issues (e.g., lettering glitches). The downgrade in quality is smaller than the jump from the original full Veo 3.1 to Fast.
Why this matters to crypto and Web3 builders
- Cost has been a critical bottleneck for integrating generative video into apps. At $0.05/s, developers can start to experiment with dynamic video NFTs, on-chain experiences that reference off-chain generative assets, in-app content personalization, and richer metaverse assets without catastrophic spend.
- Lowered API costs encourage more iteration, which is essential for token-gated content, creator tools, automated marketing creatives, and real-time media features in decentralized apps.
Market context and fallout
- Price pressure is reshaping the space. OpenAI’s Sora reportedly burned roughly $15 million per day and was shut down last week; OpenAI said it’s “pivoting to world simulation research to advance robotics.” The collapse also disrupted a reported $1 billion deal with Disney.
- Google’s Veo roadmap began with Veo 3 in May 2025, then Veo 3.1 in October as a contender to Sora 2. While quality has impressed, the cost of large-scale deployment remained prohibitive — until now.
Competition and specialization
- Chinese players have competed aggressively on price: Kuaishou’s Kling AI markets comparable generation at much lower rates than Google’s high-end tiers, and Tencent’s Hunyuan Video released an open-source model free to use.
- On the pro end, startups like Utopai (PAI) target long-form cinematic production with scene-level control and consistent characters. Those tools aren’t cheap (PAI’s current pricing: $100 for 10,000 credits), but they serve creators who want precise pipelines rather than raw generation.
Where Veo 3.1 Lite fits
- Veo 3.1 Lite isn’t trying to replace cinematic pipelines or to be the cheapest option globally. It’s a middle-tier infrastructure play: a way for developers to treat video as a standard product feature instead of an expensive novelty.
- If Google follows through with the April 7 price adjustment for Veo 3.1 Fast, the cost of building with Google’s video models will fall across the board in a single week — a meaningful shift for anyone building at scale.
Bottom line
For crypto and Web3 projects looking to add video—whether for NFTs, DAO media, immersive ads, or metaverse content—Veo 3.1 Lite makes experimentation and scale much more accessible. The competitive shakeout continues: high spend sank Sora, open and cheaper alternatives win market share, and Google’s tiered push could mainstream generated video as a routine app feature.
Read more AI-generated news on: undefined/news
