AI's Physical Constraints: How AI Rewired the Data Center
Why AI capacity stopped behaving like elastic compute and started depending on physical infrastructure, power, and place.
By Hien Phan
July 2nd, 2026
Receive the latest technical articles and release notes in your inbox.
Why AI capacity stopped behaving like elastic compute and started depending on physical infrastructure, power, and place.
By Hien Phan
July 2nd, 2026
Great models aren't enough for Physical AI. Real deployments are gated by regulation, safety, operations, and the data your machines produce.
By Hien Phan
June 18th, 2026
Most search stacks run four systems to answer one question. You don't need any of them. Build production hybrid search in Postgres with pg_textsearch for BM25, pgvectorscale for vector similarity, and Reciprocal Rank Fusion to combine them. One query. One database.
By Erin Mikail Staples
April 20th, 2026
Agentic AI requires forkable state: comparing Replit's object storage approach and Tiger's block storage architecture for fast, safe database experimentation.
By Mike Freedman
December 22nd, 2025
Why MongoDB is an architectural dead-end: bolt-on features, operational burden, and declining performance at scale. Postgres compounds value instead.
By Jose Sahad
November 26th, 2025
Tiger Data’s Jacky Liang argues that vector search alone isn't sufficient for many AI applications since it provides similarity when users need exact relevance.
By Jacky Liang
August 13th, 2025
Three Tiger Data engineers reveal the unvarnished truth about Model Context Protocol (MCP) implementation, exposing critical security concerns and practical challenges the AI hype cycle isn't discussing.
By Jacky Liang
July 24th, 2025
Learn what makes Tiger Data an open, composable platform that unifies OLTP + analytics with 10M+ row/sec performance, ACID compliance, and seamless ecosystem integration—without vendor lock-in.
By Jose Sahad
July 24th, 2025
Tiger Data’s Jacky Liang shows why AI is search and why the industry is heading to hybrid search, which involves lexical and vector approaches working together.
By Jacky Liang
July 10th, 2025