Case Study: Baidu AI Cloud achieves 80% lower cold data storage costs with IBM Tape Storage and IBM Storage Scale

Baidu enables technology innovation with an ultra-scalable data platform

Baidu AI Cloud, part of Baidu, was facing explosive data growth from AI, autonomous driving, IoT, and cloud services, with storage needs tripling since 2019 to about 100 exabytes. Its legacy disk-based cold storage was too space-intensive, costly, and inefficient, so the company turned to IBM for a more scalable solution using IBM TS4500 Tape Libraries and IBM Storage Scale.

IBM implemented a tiered cold data storage architecture with tape for cold data and existing SSD/HDD tiers for hot and warm data, along with automation to speed deployment and expansion. The results included storing more than 2 EB of cold data on 14 IBM TS4500 Tape Libraries, over 80% lower operational costs, more than 90% power savings, and 12 nines of reliability, helping Baidu AI Cloud scale rapidly while improving sustainability and compliance readiness.

Baidu AI Cloud

Miao Yu

Senior Manager of the Cloud Storage Department