Google has been approaching major Hollywood studios including Walt Disney Co, Warner Bros Discovery, Universal Pictures and Sony Pictures Entertainment to secure licensing agreements for its AI models. The talks reportedly involve Google offering multi‑billion‑dollar payouts in exchange for the right to use movies, TV shows and related scripts as training data for its Gemini family of models. Sources say the search giant has discussed deals worth up to $5 billion for a multi‑year library access.

The studios, wary of relinquishing control over valuable intellectual property, are negotiating for high fees and usage restrictions. While Google sees the data as essential to improve video generation, scene understanding and multimodal reasoning, the studios argue that AI training could undermine their future monetization and want safeguards against unauthorized reproductions.

Google needs high‑quality audiovisual data more than the studios need AI assistance. Unlike text, where web scraping can yield massive corpora, video training requires curated, licensed content that is expensive to produce. If studios can command steep prices, Google’s AI development costs will rise sharply, potentially slowing its rollout of generative video tools and putting pressure on its cloud AI revenue targets.

This dynamic could set a precedent for how AI firms acquire media training data. A successful licensing model would increase the cost barrier for entrants, benefitting incumbent studios and large tech firms with deep pockets, while squeezing smaller AI startups that cannot afford such deals. It may also spur legislative action clarifying the rights of creators when their work is used to train AI systems.

Studios win if they secure lucrative licensing fees and retain control over how their content is used, gaining a new revenue stream without relinquishing IP rights. Google wins only if it can secure the data at a price that still allows profitable AI services; otherwise the deal becomes a cost center. Losers include Google’s competitors that lack similar cash reserves, AI startups that cannot afford licensing fees, and ultimately consumers if the increased costs are passed onto higher‑priced AI products or slower innovation.

Watch for whether Google finalizes any of the reported deals and the specific financial terms, especially any caps on usage or duration. Also monitor antitrust regulators’ reactions to large payments between a tech giant and content owners, and any legislative proposals that could alter the fair‑use framework for AI training. The timeline for Gemini’s video generation features will be a key indicator of how quickly Google can overcome the data hurdle.