6/2/26

Humans in the Loop Ep 6: Jensen's Tokenmaxxing Paradigm vs. Enterprise Reality

Nvidia's Jensen Huang says a $500K engineer who doesn't burn $250K in AI tokens is underperforming. Morgan Stanley and Goldman are backing the framing. "Tokens consumed" is becoming the new AI productivity KPI.

But in cost sensitive regulated industries, burning tokens doesn't mean generating more value. In fact, if the cost of the employee plus tokens used doesn't lead to direct revenue growth to cover the cost of this "AI enabled" employee, then you're upside down on your ROI and your CFO is going to shut it down fast.

Next week on Humans in the Loop, we're asking: is the token economy real, and how should boards measure it?

Previous

Humans in the Loop Ep 7: Are FDEs the right model for AI adoption?

Next

Humans in the Loop Ep 5: You Vetted the Vendor. Did You Vet the Model?