Another data source showing the same trends consistently over the last couple of years:
https://www.stlouisfed.org/on-the-economy/2025/nov/state-gen...
Interactive data tracker here: https://www.genaiadoptiontracker.com/
I'll note that despite this, most AI providers are very compute constrained (see: the skyrocketing backlogs of each cloud provier; Claude's 9's until a few weeks ago until Anthropic, along with Google -- of all companies -- renting capacity from Elon.)
That could be a material reason for the overall low adoption. I think the big AI players are carefully balancing multiple conflicting priorities with their limited compute: acquiring and keeping users with free or heavily subsidized tokens, while ensuring they can satisfy their paying customers' SLAs. If there was more compute they would probably invest more in marketing and pushing adoption further (and indeed, a Google exec was quoted saying exactly that after a big Gemini launch.)
Something else to consider: If only 10% of users use GenAI daily, that leaves a potential 45x growth minimum on the horizon (the remaining 90% of users x 5 weekdays, assuming everyone chills on weekends, and ignoring that agents use way more tokens.) Which is why every hyperscaler has been scrambling to build out more capacity.
The typical question at this point is, where's the payoff? Two points:
1. Despite this relatively low level adoption, AI company revenues are already spiking at unprecedented rates. (Not just the frontier labs, even the smaller startups, as reported by investors, enterprise customers as well as Stripe, their payment processor: https://news.ycombinator.com/item?id=46730182)
2. The St. Louis Fed study above notes that there has been a ~1.89 percentage points increase in national labor productivity since 4th quarter 2022, which could be explained by their data on GenAI adoption.
They are careful to call out that causation is impossible to prove, but the possibility that GenAI may already be having a noticeable impact on national-level economic indicators only 3+ years since launch is wild.