Rubrik (RBRK) Cloud Resilience Summit 2026 summary

AI-era risk and resilience

• Global cloud infrastructure spending exceeds $140B per quarter, with AI the fastest-growing segment; most organizations have yet to prove recovery under real cybersecurity conditions.

• Zurich defines resilience as adapting to disruption, responding quickly and recovering confidently; its CISO views incidents as a matter of when, not if.

• Zurich is mapping system dependencies and planning recovery for network configurations, permissions, LLMs and connectivity, alongside testing restores before production recovery.

• AWS cited mean time to exploit falling from 2.3 years in 2018 to five days in 2024 and 20 hours in 2026; agentic AI also brings novel, less predictable failure modes.

S3 and AI data foundations

• AWS reported over 500 trillion S3 objects across hundreds of exabytes; examples include Indeed’s 100-PB-plus data lake and BMW Group’s 20-PB cloud data hub.

• S3 has evolved into an analytics and AI data platform; customers use Iceberg and S3 Tables, while S3 Vectors had over 250,000 indices and 50B-plus vectors ingested in six months.

• Iceberg enables schema evolution, row-level updates and multi-engine access; Indeed previously spent 1,000+ developer hours annually managing self-managed Iceberg infrastructure.

• AWS said S3 offers 11 nines of durability, but this does not prevent accidental or malicious deletion; resilience requires protecting both data and Iceberg metadata.

Data protection announcements

• Rubrik Cloud Protection for AI data foundations launched this past quarter, supporting Iceberg on AWS Glue Catalog and Amazon S3 Tables, plus Google BigQuery, with immutable copies outside the blast radius.

• BigQuery protection launched in October, covering datasets, tables, views, routines and row-level security policies, with immutable backups, long-term retention and restore options across projects or regions.

• Google Cloud Storage protection was announced, with immutable bucket backups, policy assignment by project, bucket or label, lower-cost archiving, and recovery at bucket, folder or file level.

• BigQuery Time Travel generally covers seven days; Time Travel and Fail-safe Recovery together cover around 14 days, while compliance retention may require months or years.