#ai-economics

Redoing client work after a model bug: the re-run costs $22, the review up to $6,400. How fixed-fee, hourly and per-item pricing split that bill, and the fix.

KPMG cut its auditor's fee 14% citing AI. An Indian IT chief executive says clients want the same work for 25 to 30% less. The sourced ledger and four answers.

Five earnings reports, opposite verdicts: Microsoft and Amazon rallied on AI revenue proof while Meta and Apple fell. Three 2026 capex guides: $500-525B.

Amazon's first $200B quarter: AWS up 37%, ads up 26% to $19.8B, and a $53.4B non-operating mark-to-market gain on its Anthropic stake, not operating profit.

Apple's record $109.4B June quarter still knocked ~6% off the shares after hours on soft guidance and chip supply limits. Siri AI is slated for this fall.

CNBC reports an internal OpenAI all-hands where the CFO said July's annualized run-rate topped all of Q2. What that comparison can and cannot prove.

Open-weight models now run 10-12x cheaper than frontier SaaS. A 2026 framework on the TCO crossover, vendor lock-in, and data sovereignty for agencies.

DeepSeek abandons its no-outside-capital stance in a ~$7.4B maiden round led by Tencent and CATL, valuing it near $59B and reshaping open-weight economics.

Forecast monthly LLM token spend by client tier, workflow, model mix. Downloadable forecasting model + reconciliation cadence for agency CFOs.

Why $/token is the wrong unit and $/successful-task is the right one. Formulas, worked examples across 6 task families, and a downloadable scoring template.

AI agent productivity statistics for 2026 from controlled trials, Fed and Census data, executive ROI surveys and company filings, each with its source and date.

Anthropic's Opus leads spend at $25.1M/month but Xiaomi dominates volume at 5.49T tokens. The cost-routing story behind OpenRouter's Q2 2026 data.

Q2 2026 LLM API pricing update — cost-per-token deltas for GPT-5.4, Opus 4.6, Sonnet 4.6, Gemini 3.1, Qwen 3.5/3.6 family, MiMo V2, and ultra-low tier.

10M token context window economics — when long context outperforms RAG, infrastructure costs, and real-world agency workflows that actually use it.

Q2 2026 efficient-frontier analysis — Pareto scatter plots mapping speed, quality, and cost across 20 frontier models. Identifies the dominant strategies.