TAHO Labs

TAHO Labs

Insights

Market and industry analysis on AI infrastructure, compute economics, and the forces driving demand.

Series I

The Compute Paradox

Every efficiency gain in AI compute drives more demand, not less. Eighteen articles on the structural forces behind the paradox and what it means for organizations building at scale.

Start here

Wirth's Law × Jevons Paradox in Modern AI Compute

Morgan A. Gebhardt7 min readIndustry Analysis

  1. The 5% Problem

    Justin Gelinas9 min readInfrastructure Economics

  2. The Fit Thesis

    Todd Smith6 min readInfrastructure Economics

  3. The Workload Changed Shape

    Michal Ashby7 min readIndustry Analysis

  4. Why We Build Here

    Todd Smith4 min readProduct

  5. The Energy Waitlist

    Morgan A. Gebhardt8 min readInfrastructure Economics

  6. Heterogeneity Is the Moat

    Todd Smith7 min readAI Infrastructure

  7. CFO-Legible Infrastructure

    Todd Smith8 min readInfrastructure Economics

  8. Change the Unit of Work

    Justin Gelinas8 min readAI Infrastructure

  9. Why Buying More Capacity Is Wrong

    Morgan A. Gebhardt8 min readInfrastructure Economics

  10. Reclaiming Idle Capacity

    Justin Gelinas5 min readAI Infrastructure

  11. Recovering Capacity Inside the Workload

    Justin Gelinas7 min readAI Infrastructure

  12. The Scheduler Remains with the Execution Fabric

    Jason Schultz7 min readIndustry Analysis

  13. Measuring Utilization the Right Way

    Jason Schultz7 min readIndustry Analysis

  14. Why Hasn't Utilization Been Solved Already

    Jason Schultz5 min readIndustry Analysis

  15. Running Work Where Data Is Allowed to Live

    Justin Gelinas6 min readIndustry Analysis

  16. The Agentic CPU Tax

    Todd Smith6 min readIndustry Analysis

  17. What Competing on Execution Looks Like

    Todd Smith5 min readIndustry Analysis

More insights to come.