Tesla AI Spending Cap

Six months ago, Tesla ranked employees by AI token consumption to push adoption. Starting today, it caps their spending at $200 a week.

That reversal, first reported by The Information, says more about where enterprise AI is heading than any model launch this year.

Tesla engineers were burning through thousands of dollars in tokens weekly. Uber, Meta, and Walmart have quietly imposed similar limits. The era of "use as much AI as you want" lasted roughly two quarters.

Here is the part worth watching: the cap exempts beta versions of xAI products, even though reports suggest Tesla engineers largely prefer other models. When cost controls pick winners, procurement policy starts shaping which AI actually gets used inside the enterprise.

The lesson for every operator: AI adoption was phase one. AI unit economics is phase two. Companies that never built cost visibility into their AI rollout are now discovering the meter was running the whole time.

The winners of this next phase will not be the teams using the most AI. They will be the teams that can tie every token to an outcome.

Does your organization actually know what it spends on AI per employee, per week?