NEA’s Tiffany Luck says enterprises are still figuring out their AI ROI
TL;DR
TechCrunch interviews NEA partner Tiffany Luck about enterprise AI spending as the earlier tokenmaxxing push runs into budget reality. The setup: CEOs urged employees to use AI as much as possible, then costs spiked. TechCrunch says Uber reportedly burned through its annual AI budget in months. Some companies are cutting Claude licenses for parts of their orgs, while Meta reportedly shut down an internal AI usage leaderboard.
Nauti's Take
This is where the AI hype has to grow up. Tokenmaxxing sounds like innovation culture, but without measurement it is mostly a neat label for runaway spend.
Companies should stop asking how many prompts employees generate and start asking which decisions, processes, and outcomes actually improve. That is where sober AI infrastructure beats demo magic.
Briefingshow
Enterprise AI is moving from enthusiasm to cost control. If usage is pushed up without measuring productivity gains, AI turns into an expensive software subscription layer. The next advantage will come less from model access and more from disciplined workflows, cost governance, and measurable output.
Sources