Two real datasets, asked about the same word -- "entrepreneur" -- and they point in opposite directions on age. That's not a contradiction to resolve in favor of one over the other. They're measuring two different stages of the same process, and the gap between them is the actual finding.

The participation curve: who's attempting it, self-reported, skews young. The Global Entrepreneurship Monitor surveys thousands of US adults every year and asks who's currently starting or running a business under 3.5 years old -- a measure called Total Entrepreneurial Activity, or TEA. For every year GEM tracked this from 1999 through 2021, the highest TEA rate belonged to the 25-34 or 35-44 age bands. That flipped in the two most recent survey years: 18-24 year-olds now lead, tied with 25-34, both at 25% TEA -- the highest rate GEM has recorded for that age group in the survey's 25-year history. Worth being honest about a real gap here: GEM's public releases highlight the youngest cohorts because that's the headline they're reporting, but don't state the exact current-year percentages for the 35-44, 45-54, and 55-64 bands -- so this piece can confirm the youth-led reversal is real without claiming a full five-band picture it doesn't have.

The success curve: who actually builds something that grows, tracked through Census and IRS records, skews toward the mid-40s and 50s. A 2018 NBER study (Azoulay, Jones, Kim, and Miranda) didn't survey anyone -- it linked administrative business-registry and tax data to track real firm outcomes. Among the top 0.1% of startups ranked by growth in their first five years, the average founder started the company at age 45. A 50-year-old founder is 1.8 times more likely to produce that kind of upper-tail growth than a 30-year-old founder is. And among firms that reached a successful exit -- an acquisition or an IPO -- the average founder age was 46.7, older still.

Why they diverge -- a measurement difference, not a mystery. GEM counts every real attempt: anyone who says they're taking active steps to start something, regardless of what happens next. It's a rate of activity. The NBER study only exists downstream of an attempt already made -- it can't see anyone who never started a company, and it's specifically measuring which of the people who did start one ended up building something that scaled. Younger people are starting more things right now; the founders most likely to be among the small number who build something that grows large have historically skewed a full generation older. Both statements can be true about the same population at the same time, because they're answering "how many people tried" and "which of the people who tried actually broke through" -- two different questions that happen to share the word "entrepreneur."

The credentialed-founder cohort sits on the older curve, not the younger one -- and it's not new. Kauffman Foundation data shows Americans aged 55-64 had a higher rate of entrepreneurial activity than 20-34 year-olds in every single year from 1996 through 2007, roughly a third higher on average, and that age group's share of new entrepreneurs rose from about 15% in 1996 to over 25% by 2017-2019 -- a real, sustained rise that predates the current youth-participation surge by two decades. A credentialed founder in their late 50s starting a company after a full career isn't an exception to either curve; it's the demographic the success-outcome data has consistently favored, tracked separately and for longer than the recent youth story GEM is currently reporting.

What's still genuinely unknown, worth naming rather than guessing at. Whether the recent shift toward younger participants in GEM's survey is also producing more young founders in the NBER study's upper-tail growth tier isn't answerable from what's public yet -- the NBER study's data runs through an earlier period, and GEM's youth surge is only two years old as of its most recent report. Whether the age-of-success curve is itself shifting younger, holding steady, or diverging further from the participation curve is a real open question this piece can raise but not close.