Google and Kuaishou dominate free AI video tier as OpenAI exits

The collapse of the 'two-step' pipeline — where video and audio were generated separately — into a single multimodal pass is now the industry standard for 2026. With OpenAI’s strategic retreat from the consumer video market, Google and Kuaishou are aggressively using generous free tiers to lock in creator loyalty and train models on massive user datasets. For B2B strategists, the immediate implication is a move toward 'SlopTok' social feeds and automated ad-variant generation at near-zero marginal cost. Watch for the September 24 decommissioning of the Sora API as a catalyst for a massive migration of enterprise video workloads to Google Vertex AI and Kling’s Omni One architecture.

The AI video generator market is projected to reach $1.04 billion in 2026, growing at a 22.4% CAGR, per The Business Research Company in June 2026. This growth is increasingly concentrated in vertical (9:16) formats, which now account for 59% of all AI-generated clips as of early 2026, up from just 31% two years prior. Enterprise interest has pivoted toward 'World Models' that simulate physical interactions, a development highlighted by Luma Labs’ September 2025 release of the Ray3 reasoning model, which introduced Adobe Firefly integration.

Simultaneously, the competitive landscape has undergone a major consolidation. OpenAI verified the shutdown of its Sora brand in March 2026, citing unsustainable compute costs and a pivot toward robotics, according to reports from Digen in April 2026. This retreat left a vacuum for ByteDance’s Seedance 2.0 and Google’s Veo 3.1, the latter of which was updated in January 2026 to support native vertical generation for YouTube Shorts. Research from Intel Market Research in March 2026 suggests businesses are now achieving up to 65% cost reductions by replacing traditional product demonstrations with these generative suites.