NEA’s Tiffany Luck says enterprises are still figuring out their AI ROI
TL;DR
TechCrunchs Equity podcast talks with NEA partner Tiffany Luck about why many enterprises still cannot clearly tie AI spending back to ROI. The backdrop is the tokenmaxxing wave: companies pushed broad AI usage, then faced large bills, reduced Claude access in parts of their orgs, and dropped internal usage leaderboards. Luck points to room for startups that help companies measure AI cost, usage, and business impact instead of simply adding more tools to teams.
Nauti's Take
This is the sober phase after the first AI rush. Tokenmaxxing sounded like progress, but in many cases it was just the most expensive version of experimentation without a clear scorecard.
The interesting companies will not be the ones selling another chatbot; they will be the ones translating AI work into operating metrics. The piece is VC- and podcast-heavy, but the signal is clear: enterprise AI matures when CFOs and operations teams get a real vote.
Briefingshow
The AI hype cycle is moving from access to impact. When budgets can disappear in months, usage alone no longer proves productivity. The hard part is connecting cost per workflow, time saved, quality gains, and revenue impact.
That creates a clearer market for measurement, governance, and FinOps-style tooling around AI.
Sources