Historical Behavior

Google's layout in the AI video generation field began with the release of the Veo model in 2024, followed by Veo 2 in 2025, and an upgrade to Veo 3 in 2026. Gemini started internal testing of the video generation feature in Q4 2025 and officially opened it to paid subscribers in January 2026. This limited-time free offer is the first time it has been opened to all users, reflecting its strategy of "pay first for verification, then acquire users for free." Compared to OpenAI's Sora (released in 2024, still limited to internal testing in 2026), Google's "rapid commercialization" path is more aggressive.

Capital Path

The flow of funds is clear: Google's AI capital expenditure is expected to reach $120 billion in 2026, with 15% ($18 billion) allocated to the development of video generation models. Gemini's subscription revenue is projected to be $1.2 billion in 2025, increasing by 85% year-on-year to $2.2 billion in Q2 2026. The limited-time free strategy essentially exchanges "R&D costs for user growth"—the cost of 10 free videos is about $0.5 per user, but the customer acquisition cost is only $5 per person (traditional advertising costs $50 per person). Funding is shifting from "advertising customer acquisition" to "product customer acquisition," reflecting its strategy of "product as marketing."

Comparison + Industry Position

Similar cases include: In 2023, Midjourney opened free image generation, increasing users from 1 million to 15 million, but switched to paid subscriptions in 2024 (monthly fee of $10-60). In 2025, Runway offered limited-time free video generation, with a conversion rate of 15% (15% of users converted to paid). Gemini is currently in a "user growth phase"—compared to Sora (closed testing) and Runway (monthly fee of $12-95), Gemini's pricing is more competitive (Pro $20/month, Ultra $200/month). Its industry position is in the "early explosion of AI video generation"—the global AI video market is expected to reach $5 billion in 2026 and $50 billion by 2030.

Structural Judgment

This essentially represents a technological replacement leading to a reconstruction of content production. The traditional video production process (script → shooting → editing → post-production) requires a team of 5-10 people, takes 2-4 weeks, and costs $10,000-$100,000. AI video generation compresses the process to "prompt → generate → fine-tune," taking 10-30 minutes and costing $0.1-$1. This explains why Google is offering a limited-time free trial—not out of charity, but to "accelerate market education." As users become accustomed to AI video generation, the traditional video production market is expected to shrink by over 70%. This drives the "democratization of content"—high-cost video production (advertising, movie trailers, educational content) will become accessible to individual creators.

ABAB News · Cognitive Laws

Limited-time free offers are accelerators of market education
The essence of AI video is the zeroing out of time costs
The marginal cost of content production determines industry boundaries