Google's Gemini has introduced a limited-time free video generation feature, allowing all users to create 10 videos by August 4, marking the first time the feature has been opened to non-subscribers. This follows Google's release of the Veo model in 2024, Veo 2 in 2025, and Veo 3 in 2026, with Gemini starting internal testing of video generation in Q4 2025 and opening it to paid subscribers in January 2026. The free offer reflects a strategy of "pay first for verification, then acquire users for free," with the cost of 10 free videos estimated at $0.5 per user and a customer acquisition cost of $5 per person, compared to traditional advertising costs of $50 per person. Google's AI capital expenditure is expected to reach $120 billion in 2026, with 15% ($18 billion) allocated to video generation model development. Gemini's subscription revenue is projected at $1.2 billion in 2025, increasing by 85% year-on-year to $2.2 billion in Q2 2026. The global AI video market is expected to reach $5 billion in 2026 and $50 billion by 2030.