In a striking shift, enterprise leaders are beginning to question the value of their AI investments as mounting costs and uncertain returns spark a wave of retrenchment. Microsoft has reportedly canceled most of its Claude Code licenses, citing cost concerns, while Uber’s COO acknowledged that AI spending is becoming increasingly difficult to justify. This signals a broader trend away from “tokenmaxxing”—the practice of burning through AI tokens without clear business outcomes—toward more disciplined, ROI-focused AI deployment strategies. Ali Ansari, CEO of model‑training firm Micro1, describes this as a “healthy swing” in enterprise AI usage. The key question now is whether companies will recalibrate AI use for efficiency or overcorrect and stifle innovation.