The Pulse: AI token spending out of control – what’s next?
Token spend is spiraling out of control. Here's how 15 tech companies are fighting back.
Why it matters
As AI inference costs become a material P&L line item for enterprises, token spending patterns and vendor capacity constraints are forcing a strategic reckoning on model deployment, optimization, and cost governance — a board-level concern that separates serious AI adopters from the rest.
The key facts
10 to know15 tech companies analyzed for token spending trends
Rapid growth in token consumption documented
AI vendor capacity constraints vs. demand
Meta morale deterioration noted
Cost governance emerging as operational priority
15 tech companies surveyed on token spend growth
AI vendors reporting demand outpacing supply
Meta experiencing plummeting morale (internal org signal)
Cost control mechanisms failing across surveyed companies
Vendor capacity constraints emerging as systemic issue
Go to the source
Pragmatic Engineernewsletter.pragmaticengineer.com
Publisher excerpt: Details from 15 tech companies on the rapid growth of token spend, and their responses to it. Also: AI vendors can’t keep up with demand, plummeting morale at Meta, and more.