Case Study: Decart achieves high-performance, cost-efficient AI training with Backblaze B2 Overdrive

Decart cuts storage costs by 75% with Backblaze

Decart, an AI research lab, faced a significant challenge scaling its high-performance multi-cloud training infrastructure. The startup was burning through free credits from a traditional cloud provider and needed a storage solution that could cost-effectively handle petabyte-scale datasets to support its growth. Decart chose Backblaze B2 Cloud Storage to meet its needs for reliably high performance, GPU interoperability, and efficiency.

By implementing Backblaze B2, Decart was able to transfer petabyte-scale datasets with no issues and achieved a 75% reduction in AI operation costs due to Backblaze's pricing being one-fifth that of traditional providers. The free egress policy was a game-changer, allowing Decart to train models on multiple GPU clusters simultaneously with zero additional cost for data transfer. This solution from Backblaze unlocked significant budget and provided the stable, high-performance foundation necessary for Decart to disrupt the AI market.

Decart