AI Capex vs AI Opex: Stop Paying Agents to Re-solve the Same Problem
njyx · x · 2026-09-20
Steven Willmott of Steampunk AI proposes a simple mental model for agent architecture: split AI spend into "capex" (upfront investment in a reusable asset) and "opex" (recurring runtime cost) — not accounting terms, but a lens on where your agent money actually goes.
Using a data-processing example (an agent named Herman that aggregates multi-source data into recommendations), he argues many organizations are effectively paying AI to solve the same problem over and over.
He isn't arguing for fewer agents, but urges developers to examine which parts of a workflow should be baked into reusable assets versus which genuinely need runtime reasoning.
More from coding & agent
- Da7em Bench: independent AI benchmark scores models on 200 real client tasks across 12 areas — airesearch12 · 2026-09-20
- Using Codex + GPT-6 Astra to plan art installations: rebuild the wall in Blender, skip recalculation — perilli · 2026-09-20
- What does an AI engineer's day actually look like? A learner asks Reddit — tech_kie · 2026-09-20
- Open-source local pipeline generates full radio dramas from one click — fflluuxxuuss · 2026-09-20
- OpenClaw Hits 390K GitHub Stars as Open-Source Computer-Use Agents Proliferate — Aiden_Tech_Ai · 2026-09-20
- Gemini CLI PR fixes folder trust not persisting when running in podman/docker sandbox — 21vedansh · 2026-09-20