
AI budgets and consumption-based pricing

Measuring the impact of AI on software engineering – with Laura Tacho
Topic 9 of 14
Published Jul 23, 2025 · Clip 47:07–52:03 (4:56)
Laura raises the new measurement challenge of consumption-based AI pricing: figuring out which developers and use cases gain the most from token allocations, and whether senior or junior engineers should get bigger buckets. The host draws a parallel to 2000-2010, when companies paid $3,000-$8,000 per developer per year for Visual Studio licenses, docs, and early software. Laura compares today's tool sprawl to the Kubernetes hype cycle, predicts stabilization within a year or two, and forecasts $1,200-$2,000 per month for an autonomous agent within 18 months, as companies trade headcount growth for AI spend.