Case Study: Freepik scales FLUX image generation to millions of requests per day with DataCrunch and WaveSpeed

How Freepik scaled FLUX media generation to millions of requests per day with DataCrunch and WaveSpeed

Freepik, a leading AI-powered creative suite, faced the immense challenge of scaling its FLUX AI image generation to handle millions of daily requests. They needed to maintain extremely low latency and high output quality while optimizing for cost-efficiency as their user base grew exponentially. To meet these demands, they partnered with vendor DataCrunch for its managed GPU infrastructure and high-velocity model serving with WaveSpeed.

The solution implemented by DataCrunch provided managed GPU orchestration, elastic scaling, and near-zero cold starts, enabling Freepik's infrastructure to seamlessly handle traffic spikes. This technical partnership, which included rigorous optimization to preserve output quality, allowed Freepik to generate over 60 million images per month. DataCrunch's infrastructure and services were fundamental in achieving this scale, providing significant cost savings and enabling higher generation quotas for users.

Freepik

Iván de Prado

Head of AI