The Information Machine

The substitution math is the part to internalize. Tasks that used to need a junior analyst for several hours, converting…

SemiAnalysis Twitter · SemiAnalysis (@SemiAnalysis_) · 2026-06-27

SemiAnalysis reports that agentic AI workloads reduce effective Opus 4.7 costs to $0.99 per million tokens — far below sticker price — by exploiting 300:1 input-to-output ratios and above-90% cache hit rates, fundamentally restructuring the unit economics of professional services.

Open original ↗

Appears in

Extraction

Topics: agentic-aitoken-economicsprofessional-services-disruptionprompt-cachingclaude

Claims

  • Agentic workloads run at approximately 300:1 input-to-output token ratios.
  • Cache hit rates above 90% reduce effective Opus 4.7 cost to about $0.99 per million tokens against a $5/$25 sticker price.
  • Tasks that previously required analyst hours now complete in minutes for a few dollars of tokens.
  • This shift represents a fundamental change in the unit economics of professional services, not a marginal efficiency gain.

Key quotes

The blended Opus 4.7 cost we observe is about $0.99 per million against $5/$25 sticker, mostly because agentic workloads run 300:1 input-to-output ratios and cache hit rates above 90% pull the effective price down.
Thats a real change in the unit economics of professional services, not a 10% efficiency gain.