The Book

Understanding AI Economics

Cost Drivers, Value Creation and Pricing Strategies. A practical guide to the economics that determine whether AI products create sustainable value.

Front cover of Understanding AI Economics
By Miklos Kremser

AI changes what software costs—and what customers value.

Every software company is racing toward AI. Far fewer have stopped to ask whether the economics actually work. Traditional SaaS relied on near-zero marginal costs; AI introduces a real cost for every prompt, generated token and agent step. Yet the cost of intelligence is also falling dramatically. Understanding AI Economics shows how to navigate that contradiction—and turn capability into sustainable economic value.

Drawing on 23 years of pricing strategy, analytics and financial-modeling experience, Miklos Kremser connects AI architecture, unit economics, value creation, packaging, productivity and financial governance.

Buy on Amazon ↗
Inside the book

Eight chapters. One economic framework.

Explore the major ideas and selected topics covered across the book. Expand any chapter below to see more.

CHAPTER 1The Prediction Economy and the Fall of AI Marginal Costs
  • When Prediction Became Cheap
  • What Prediction Actually Means
  • Reframing Business Problems as Prediction Problems
  • Complements and Substitutes — The Ripple Effect
  • Prediction Isn’t Judgment
CHAPTER 2The SaaS Inversion — Why AI Breaks Traditional SaaS Margins
  • The Anatomy of the Margin Collapse
  • A Tale of Two Software Products
  • When Your Best Customers Become Your Worst Economics
  • Three Companies, Three Outcomes
  • The Metered AI Cost Stack
  • The Hidden Operating Cost
CHAPTER 3Agentic Economics — The Cost and Value of Autonomous AI
  • The Anatomy of an Agent Loop
  • Three Ways to Build It
  • Cost Control Frameworks for Agentic Systems
  • Evaluating the Unit Economics of Agents
  • Case Study: Dental Revenue Cycle Management and Claims Processing
CHAPTER 4Productivity, Automation, and the Jevons Paradox in AI
  • Price Elasticity of Demand
  • The Three Phases of Labor Reallocation
  • Automation vs. Augmentation
  • Local Productivity ≠ System Productivity
  • Designing an Organization for the Jevons Era
  • The Jevons Applicability Test
CHAPTER 5Tokenomics and Architectural Cost Levers
  • The Four Stages of an Inference Request
  • Input Tokens vs. Output Tokens
  • Memory Bandwidth vs. Compute
  • Context Windows and the KV Cache Tax
  • Self-Hosting vs. Commercial APIs
  • Prompt Caching, Model Routing and Cascading
  • Putting It All Together: The Agentic AI Cost Estimator
CHAPTER 6Packaging and Pricing AI — From Seats to Outcomes
  • The Monetization Mismatch
  • The Value Metric
  • Four Monetization Paradigms
  • Pricing Strategy Decision Framework
  • Designing an Abstract Credit System
  • Pricing AI Features Across Tiers
  • Van Westendorp, MaxDiff, and Conjoint Analysis
  • Needs-Based Segmentation and MaxDiff
CHAPTER 7Operating AI at Scale
  • The AI FinOps Lifecycle
  • Multi-Tenant Cost Allocation: Showback vs. Chargeback
  • API Key Governance and the AI Gateway
  • Practical Cost Controls and Guardrails
  • Latency, Throughput, and Cost
  • Hardware, Model and Infrastructure Choices
CHAPTER 8The Horizon of AI Economics
  • The Moving Target of AI Unit Economics
  • Autonomous Agents and Iterative Reasoning
  • Synthetic Data Generation
  • Next-Generation Model Architectures
  • Where the Cost Curve Is Headed
  • Return to Jevons
  • A Message to the Next Generation
Who it is for

For leaders making consequential AI decisions.

Written for pricing professionals, product and business leaders, consultants and anyone tasked with understanding AI costs or translating AI capabilities into sustainable value. No machine-learning engineering background required.

Explore the companion cost estimator →