OpenAI and Google Impose New Daily Limits for Sora, Nano Banana AI Tools Amid Demand Surge

Why It Matters

The caps expose scaling challenges for AI services and push users toward paid subscriptions, affecting revenue models and infrastructure planning.

Key Takeaways

• OpenAI caps free Sora videos at six daily
• Google restricts Nano Banana Pro to two free images daily
• Paid subscribers keep unlimited or higher limits
• Limits reflect GPU compute and energy pressures
• Older Nano Banana still allows 100 images per day

Pulse Analysis

The release of OpenAI’s Sora video generator and Google’s Nano Banana Pro image model sparked a wave of user experimentation that quickly outpaced the capacity of existing data‑center hardware. Within weeks, social media feeds were flooded with AI‑crafted clips and pictures, driving unprecedented traffic to the free tiers of both platforms. This surge exposed a mismatch between viral demand and the finite GPU resources allocated for real‑time generation, prompting the companies to act before service quality deteriorated.

GPU clusters powering generative models consume significant power and generate heat, making them expensive to scale on short notice. OpenAI’s internal memo cited "melting GPUs" as a hyperbolic warning that the current load was pushing hardware toward thermal limits, while Google’s variable caps on Gemini 3 Pro reflect a similar need to balance load across shared infrastructure. By throttling free usage, both firms preserve capacity for paying customers, protect hardware longevity, and avoid spikes in electricity costs that could erode profit margins.

The new limits also signal a strategic shift toward monetizing high‑cost AI services. With free tiers now constrained, power users are more likely to upgrade to ChatGPT Plus, Pro, or Google’s AI subscription plans, generating recurring revenue that can fund next‑generation hardware upgrades. Competitors watching the rollout may pre‑emptively adjust their own usage policies to avoid similar bottlenecks. In the longer term, the industry may see tighter integration of specialized AI accelerators and more transparent pricing models as compute scarcity becomes a central business consideration.