Token Maxing comes to an end: AI Revenue, CRO Comp Bubbles, and Hyperscaler Economics

Doom and Quinn discuss signs that “token maxing” is peaking as AI token spend has surged (Ramp data cited as 13x higher than January 2025) while finance teams begin tightening controls and accounting (a proposed AI COGS line, reclassifying credits, departmental allocations, and new AI margin metrics). They react to headlines including reported $100M CRO packages at frontier AI labs, group quotas, and concerns about accountability and churn risk in usage-based models without committed contracts; Uber’s COO questioning AI ROI after burning a 2026 budget in four months; Microsoft canceling some cloud code subscriptions; and Amazon scrapping an internal AI leaderboard amid soaring costs. They explore customer optimization (e.g., cutting cloud spend 40% while increasing AI usage), model aggregation/exclusivity dynamics, and why hyperscalers may profit more from tokens than raw GPU IaaS, highlighting Amazon/Google advantages in energy planning and custom silicon versus Microsoft’s internal demand and Nvidia-reseller “neo clouds.”

00:00 AI Token Spend Surge

00:54 Week Kickoff and Headlines

04:23 Sales Comp and Quota Debate

06:44 Contracts vs Usage Churn

11:05 Transactional vs Relational Selling

15:13 AI Tools Flatten GTM Orgs

19:14 FinOps Playbook for Tokens

20:35 Leaderboards and Budget Blowups

21:53 OpenRouter and Model Switching

26:16 Hyperscaler Token Economics

30:52 Hype Cycle and ROI Reality Check

34:59 Everyday ROI and Lightbulb Phase

38:02 Wrap Up and Next Week