Hidden Disruptions Inside AWS: Four Second-Order Effects of the AI Reorganization

This analysis is part of Amazon’s AI Business Model Pivot, a deep dive by The Business Engineer.

The reorganization’s public narrative is about “focusing on agents.” But beneath the surface, four disruptions reveal deeper strategic shifts—these are the things not in the press release.

1. Knowledge Base / Quick Suite Duplication

Internal teams built overlapping retrieval technology independently. Bedrock Knowledge Bases and Quick Suite used similar RAG technology targeting different customer groups. In mid-2025, Knowledge Bases moved under a Quick Suite executive (Jigar Thakkar, ex-Microsoft Teams) to unify stacks. The signal: teams racing to build similar AI capabilities suggests coordination gaps at scale.

2. SageMaker Consolidation Setback

In 2024, AWS announced plans to fold Glue, EMR, and Athena under the SageMaker brand—creating a unified data/ML platform. The March 2025 reorg separated data and ML teams into different organizations, forcing integration work to restart. Agents took priority over data platform. This creates an opening for Databricks and Snowflake.

3. The DeSantis Signal

Peter DeSantis moved out of AWS to lead Amazon’s Nova models, custom chips, and quantum computing—reporting directly to Andy Jassy. AI research was elevated from an AWS product to an Amazon-level corporate priority, mirroring Google’s Gemini structure.

4. Anthropic Positioning

Amazon invested $8B+ in Anthropic, generating $12.8B in TTM paper gains. the economics of AI compute infrastructure — /”>Project Rainier was built specifically for Anthropic. This is the deepest AI partnership in industry history—positioning AWS as the “Switzerland of AI infrastructure” while building proprietary capabilities.

Read the full analysis on The Business Engineer →