Google has released Gemini Omni 1.2 Flash, cutting the cost of its video generation API by 30 percent while extending video output from 10 to 15 seconds and improving prompt adherence.
The move targets the emerging enterprise market for generative video, where adoption has hinged on production economics rather than technical capability. Longer clips and lower per-transaction costs directly reduce API spending for developers embedding video generation into workflows—a metric that directly flows to Google Cloud's gross margin.
Runway, Pika, and OpenAI's Sora dominate consumer and creator tools, but enterprise adoption has lagged pricing. Google's strategy mirrors the playbook in other AI services: commoditize on cost, capture volume, defend with margin. At 30 percent lower cost per second of generated video, Gemini 1.2 Flash forces competitors to choose between cutting prices or ceding enterprise development deals.
The shift from 10-second to 15-second clips matters less for technical reasons than for business model efficiency. Longer outputs reduce per-clip API calls, lowering transaction overhead and improving unit economics for customers. This is the opposite of artificial constraint—it is economic rationalization.
Google is now competing explicitly on cost-per-second generated, not on demo capability. That shift signals the market is maturing past the hype phase and into the adoption phase, where operational efficiency determines winner and loser in cloud services.
