Cathie Wood just called 2.8% productivity growth the warm-up act.
The Signal
US nonfarm productivity is clocking 2.8% year-over-year, which sounds modest until you remember the twenty-year average before COVID was 1.4%. Wood's bet is this isn't the ceiling, it's the floor. Ark's thesis: we're in the early innings of AI deployment actually showing up in output-per-hour numbers, not just demos and pilot programs.
Here's why that matters. Productivity growth is the only real way economies get richer without just working more hours or burning more resources. The last productivity boom, the internet wave of the late 90s, gave us about five years of 2.5-3% annual gains. That translated into higher wages, lower inflation, and the room to invest in new categories of work. Wood's betting AI does the same thing, except bigger and faster, because the implementation curve is steeper. Companies aren't waiting five years to roll out email. They're spinning up AI agents in months.
The counterargument is that we're seeing measurement artifacts, not real gains. Maybe companies are just cutting headcount and calling it productivity. Maybe the gains are concentrating in three sectors while everyone else treads water. But if Wood's right and this accelerates past 3%, past 4%, we're looking at a fundamentally different economy by 2028. One where the returns to building infrastructure for AI agents, not just the agents themselves, become obvious to even the skeptics.
The Implication
Watch which companies are reporting margin expansion with flat or falling headcount. That's where the productivity story is real, not aspirational. And if this trend holds, expect a rush into tools that help smaller companies access the same leverage the big players already have.
Source: Bloomberg Tech