AI content storage is cheaper than regeneration until 2040
To Store or To Regenerate? A Cost Model for AI-Generated Content at Scale
Performance
Summary
Storing AI-generated pictures and content uses a lot of space and energy as they grow in number. The authors studied when it might be cheaper to recreate images on demand rather than storing them forever. They found that simply regenerating images from prompts is more expensive than storing them for now and the next couple decades. However, saving a smaller, intermediate version of the image and regenerating from that can cut costs significantly right now. This approach balances storage and computation to make retrieval cheaper and faster.
What this means in practice
- •For cloud infrastructure teams: Reduce storage and compute costs for delivering large volumes of AI-generated images by storing compact intermediate data and regenerating artifacts on demand.
- •For content delivery engineers: Optimize caching layers and regeneration strategies in image serving pipelines to balance latency and operational costs using intermediate representation storage.
Authors
Yunjia Zheng, Zirui Wang, Haoran Ni, Tingfeng Lan, Zhaoyuan Su, Yue Cheng, Juncheng Yang
Abstract
AI-generated content is becoming a rapidly growing class of digital artifacts. Because these artifacts accumulate over time, their exponential growth creates a substantial storage, energy, and infrastructure cost for operators and society. At the same time, GPU compute cost continues to fall rapidly with each hardware generation. This divergence raises a fundamental question: when does on-demand regeneration become cheaper than persistent storage? This paper develops a cost model for comparing persistent storage and on-demand regeneration for AI-generated artifacts. The model accounts for corpus growth, HDD and tape price trends, drive replacement, electricity, request skew, caching, generator FLOPs, and future GPU price-performance improvements. For image generation, our analysis shows that prompt-based regeneration does not become cheaper than storage until around 2040, because every cache miss must still rerun the full prompt-to-artifact generation pipeline. We observe that widely used diffusion-based generation models operate in latent space, creating an alternative point in the cost tradeoff: instead of storing the final artifact or only the prompt, operators can store a compact intermediate representation (IR) and perform cheap on-demand decoding. Our analysis shows that caching combined with IR-based regeneration substantially reduces both storage and compute cost, making it at least 2x cheaper than both full-object storage and prompt-based regeneration even today. On a production image trace with 2.07 billion requests, the same conclusion holds: prompt-based regeneration is over 100x more expensive than storage, while IR-based regeneration reduces total cost to roughly half that of full-object storage while preserving interactive miss latency.