19% of people shaping AI who founded a dated organization have founded two or more, waiting a median of 7 years before a second
Data as of
- Repeat founders
- 221 (19%)
- Median 1st-to-2nd wait
- 7 yr
- Median 2nd-to-3rd wait
- 4 yr
- With prior work role
- 63%
Across 1,136 people who founded or co-founded a dated company, standalone lab or nonprofit, we estimate that 221 (19%) are repeat founders who started two or more organizations, and 915 (81%) founded one. Those 221 repeat founders account for 506 of all 1,421 founding events (36%) across 491 of the 1,266 distinct founded organizations (39%). Their median interval from a first founding to a second is 7 years (middle half 3 to 11 years), and the median interval from a second founding to a third is 4 years (middle half 2 to 7 years).
Key findings
- 221 of 1,136 founders (19%) started at least two organizations: 170 people (77% of repeat founders, 15% of all founders) founded two organizations, and 51 (23% of repeat founders, 4% of all founders) founded three or more, led by J. Craig Venter and Kyle Vogt with 5 each.
- Second-or-later foundings reached 27% of all founding events in 2020–2026: across calendar decades with at least 50 founding events, repeat follow-on foundings accounted for 10% of founding events in the 1980s (6 of 61), 17% of founding events in the 1990s (21 of 122), 16% of founding events in the 2000s (32 of 203), 17% of founding events in the 2010s (91 of 526) and 27% of founding events in 2020–2026 (131 of 487).
- Founders who go straight into a second venture do so in a median of 4 years: the 138 repeat founders (62%) who did not start an outside work role between their first and second foundings waited a median of 4 years (middle half 2 to 8), and the 83 repeat founders (38%) who took an outside work role between ventures waited a median of 11 years (middle half 8 to 17).
- Nonprofits and standalone labs rose from 14% on a first founding to 19% on a third or later founding: companies fell from 86% of repeat founders' first foundings (190 of 221) to 82% of second foundings (182 of 221) and 81% of third-or-later foundings (52 of 64).
- 63% of repeat founders held a dated prior work role before their first founding, led by universities at 54.3% of those with dated prior work: among the 140 repeat founders with dated prior work across 229 parent organizations, 76 (54.3%) worked at universities, 74 (52.9%) worked at companies and 38 (27.1%) worked at institutes and labs, with Stanford University leading individual prior workplaces at 15 repeat founders.
How common repeat founding is
Of the 221 repeat founders, 170 (77%) founded two organizations and 51 (23%) founded three or more. Specifically, 40 people founded three organizations, 9 founded four, and 2 founded five: J. Craig Venter (The Institute for Genomic Research (1992), Celera Genomics (1998), J. Craig Venter Institute (2006), Human Longevity (2014) and Diploid Genomics, Inc. (2026)) and Kyle Vogt (Justin.tv (2007), Socialcam (2011), Twitch (2011), Cruise Automation (2013) and The Bot Company (2024)). Among the 623 founders whose first founding occurred in or before 2016 (giving at least 10 years of follow-up), 168 (27%) have founded a second organization, including 107 (17%) within 10 years of the first.
Repeat-founder rates vary across role categories. Builders have the highest repeat-founder rate among dated founders at 32% (75 of 234), followed by public voices at 25% (33 of 132), lab leaders at 20% (21 of 103) and researchers at 14% (92 of 667). Counted by headcount of repeat founders, researchers make up the largest group with 92 of the 221 repeat founders (42%). Across calendar decades with at least 50 founding events, second-or-later foundings represented 10% of founding events in the 1980s (6 of 61), 17% of founding events in the 1990s (21 of 122), 16% of founding events in the 2000s (32 of 203), 17% of founding events in the 2010s (91 of 526) and 27% of founding events in 2020–2026 (131 of 487), with 2020–2026 setting the series high at 27%. Of all 285 second-or-later founding events, 131 (46%) took place in 2020–2026 (99 second foundings and 32 third-or-later foundings).
How long repeat founders wait between organizations
Across all 221 repeat founders, the median time from the first founding to the second is 7 years, with the middle half spanning 3 to 11 years. Within that group, 13 people (6%) founded two organizations in the same calendar year (0 years), 61 (28%) founded their second within 3 years, 91 (41%) within 5 years, and 160 (72%) within 10 years. The remaining 61 repeat founders (28%) waited more than 10 years before starting their second organization. For the 51 people who went on to a third founding, the median wait from the second founding to the third is 4 years (middle half 2 to 7 years), 3 years shorter at the median than the wait from the first to the second. Across all 285 consecutive founding pairs, the median interval is 6 years (middle half 3 to 10), and the total career span from a repeat founder's first founding to their latest founding has a median of 8 years (middle half 4 to 14, and 12 years among those with three or more foundings).
Range dots showing the median and interquartile range of years between foundings across 7 interval measures for 221 repeat founders.
Source: People of Superintelligence career records from biographies and OpenAlex, as of 22 October 2026.
Data table
| Interval measure | Founders or pairs | Median (yr) | 25th pct (yr) | 75th pct (yr) | Same year (0 yr) | Within 3 yr | Within 5 yr | Within 10 yr | Over 10 yr |
|---|---|---|---|---|---|---|---|---|---|
| 1st to 2nd founding (all repeat founders) | 221 | 7 | 3 | 11 | 13 (6%) | 61 (28%) | 91 (41%) | 160 (72%) | 61 (28%) |
| 2nd to 3rd founding | 51 | 4 | 2 | 7 | 5 (10%) | 19 (37%) | 33 (65%) | 48 (94%) | 3 (6%) |
| 1st to 2nd: no outside role between | 138 | 4 | 2 | 8 | 13 (9%) | 60 (43%) | 83 (60%) | 121 (88%) | 17 (12%) |
| 1st to 2nd: outside role between | 83 | 11 | 8 | 17 | 0 (0%) | 1 (1%) | 8 (10%) | 39 (47%) | 44 (53%) |
| 1st to 2nd: same organization type | 180 | 6 | 3 | 10 | 12 (7%) | 52 (29%) | 77 (43%) | 137 (76%) | 43 (24%) |
| 1st to 2nd: switched organization type | 41 | 9 | 5 | 18 | 1 (2%) | 9 (22%) | 14 (34%) | 23 (56%) | 18 (44%) |
| 1st to latest founding (total span) | 221 | 8 | 4 | 14 | 13 (6%) | 52 (24%) | 72 (33%) | 138 (62%) | 83 (38%) |
| All consecutive founding pairs | 285 | 6 | 3 | 10 | 18 (6%) | 83 (29%) | 128 (45%) | 216 (76%) | 69 (24%) |
| 2nd to later consecutive pairs | 64 | 5 | 2.8 | 8 | 5 (8%) | 22 (34%) | 37 (58%) | 56 (88%) | 8 (13%) |
| 1st to latest founding (3+ foundings) | 51 | 12 | 8 | 18.5 | 1 (2%) | 4 (8%) | 4 (8%) | 20 (39%) | 31 (61%) |
The interval between the first and second founding separates by whether a founder took an outside work role in between. The 138 repeat founders (62%) who recorded no new outside work stint between their first and second foundings waited a median of 4 years (middle half 2 to 8 years), with 83 (60%) launching their second organization within 5 years. The 83 repeat founders (38%) who started a dated work role at an outside parent organization after their first founding and before their second waited a median of 11 years (middle half 8 to 17 years), 7 years longer at the median than those without an outside role. The most common intermediate outside workplace between a first and second founding is Google, with 13 people (Anthony Levandowski, Boris Sofman, David Silver, Eric Schmidt, Fei-Fei Li, Navneet Dalal, Nick Frosst, Nitish Srivastava, Pete Warden, Thomas Hofmann, Timnit Gebru, Tristan Harris and Yang Zhilin). Repeat founders who kept the same organization type on their first and second foundings (180 people, 81%) had a median interval of 6 years (middle half 3 to 10), and the 41 repeat founders (19%) who switched organization types had a median interval of 9 years (middle half 5 to 18), 3 years longer at the median than those who kept the same type. Same-year foundings (0 years) include Anders Heyden (23°N and Ludesi in 2023), François Pachet (Imagine All The People and Ynosound in 2023), Matthew E. Peters (Envive and Spiffy in 2023), Ziad Obermeyer (Dandelion Health and Nightingale Open Science in 2023) and Sabina Jeschke (Arctic Brains AB and Quantagonia GmbH in 2021); the longest intervals between a first and second founding reach 37 years, including Tony Sale (Alpha Systems in 1968, then The National Museum of Computing in 2005, 37 years), John Henry Clippinger (Brattle Research Corporation in 1980, then Institute for Data Driven Design in 2014, 34 years), Yoshua Bengio (Mila in 1993, then LawZero in 2025, 32 years), Roy Rada (Hypermedia Solutions Limited in 1993, then Rada Research Trust in 2023, 30 years) and Lawrence J. Fogel (Decision Science, Inc. in 1965, then Natural Selection in 1993, 28 years).
What repeat founders build: companies, nonprofits and standalone labs
Across all 506 founding events by repeat founders, 424 (84%) are companies, 70 (14%) are nonprofits and 12 (2%) are standalone research labs. Across the 915 one-time founders, 753 (82%) founded a company, 139 (15%) a nonprofit and 23 (3%) a standalone lab. Across founding stages, companies fell from 86% of repeat founders' first foundings (190 of 221) to 82% of second foundings (182 of 221) and 81% of third-or-later foundings (52 of 64, and 80%, or 41 of 51, on third foundings specifically). Combined nonprofits and standalone labs rose from 14% on first foundings (24 nonprofits and 7 labs) to 18% on second foundings (37 nonprofits and 2 labs) and 19% on third-or-later foundings (9 nonprofits and 3 labs), with nonprofits alone moving from 11% to 17% and 14%.
Stacked horizontal bars showing the share of companies, nonprofits and standalone labs across 5 founding stages.
Source: People of Superintelligence career records from biographies and OpenAlex, as of 22 October 2026.
Data table
| Founding stage | Foundings | Companies | Company share | Nonprofits | Nonprofit share | Standalone labs | Lab share | Nonprofits + labs |
|---|---|---|---|---|---|---|---|---|
| 1st founding of repeat founders | 221 | 190 | 86% | 24 | 11% | 7 | 3% | 31 (14%) |
| 2nd founding of repeat founders | 221 | 182 | 82% | 37 | 17% | 2 | 1% | 39 (18%) |
| 3rd founding of repeat founders | 51 | 41 | 80% | 7 | 14% | 3 | 6% | 10 (20%) |
| 3rd or later founding of repeat founders | 64 | 52 | 81% | 9 | 14% | 3 | 5% | 12 (19%) |
| All repeat-founder foundings | 506 | 424 | 84% | 70 | 14% | 12 | 2% | 82 (16%) |
| One-time founders | 915 | 753 | 82% | 139 | 15% | 23 | 3% | 162 (18%) |
Across their full founding histories, 163 of the 221 repeat founders (74%) founded only companies, 13 (6%) founded only nonprofits, and 45 (20%) founded organizations of more than one type. Among those 45 cross-type repeat founders, 27 started a company on their first founding and later founded a nonprofit or standalone lab, including Christian Guttmann (HealthiHabits in 2016, then Nordic Artificial Intelligence Institute in 2026), Clara Shih (Hearsay Systems in 2009, then New Work Foundation in 2026), Andrew Yang (Stargiving in 2000, then Venture for America in 2011), Andy Konwinski (Databricks in 2013, then Laude Institute in 2025) and Joseph Gordon-Levitt (HitRecord.org in 2005, then Creators Coalition on AI in 2025). Another 15 started a nonprofit or standalone lab first and later founded a company, including J. Craig Venter (The Institute for Genomic Research in 1992, then Celera Genomics in 1998), Neil Zeghidour (Kyutai in 2023, then Gradium in 2025), Sam Rodriques (FutureHouse in 2023, then Edison Scientific in 2025), Tamay Besiroglu (Epoch AI in 2022, then Mechanize in 2025) and Conor Grennan (Next Generation Nepal in 2006, then AI Mindset in 2024), and 3 (Bernard Vauquois, Margaret Masterman and Yoshua Bengio) founded both a nonprofit and a standalone lab. Exactly 15 organizations were co-founded by two or more repeat founders, each with 2 repeat founders: Teknowledge (Douglas B. Lenat and Jerry Kaplan), iRobot (Helen Greiner and Rodney A. Brooks), Justin.tv (Emmett Shear and Kyle Vogt), SeaMicro (Andrew Feldman and Gary Lauterbach), Databricks (Andy Konwinski and Ion Stoica), Cerebras Systems (Andrew Feldman and Gary Lauterbach), Black in AI (Rediet Abebe and Timnit Gebru), CoreWeave (Brian Venturo and Michael Intrator), Weights & Biases (Chris Van Pelt and Lukas Biewald), Wadhwani AI (Romesh Wadhwani and Sunil Wadhwani), Cohere (Ivan Zhang and Nick Frosst), TheSequence (Jesus Rodriguez and Ksenia Se), Exafunction (Douglas Chen and Varun Mohan), Codeium (Douglas Chen and Varun Mohan) and Epoch AI (Jaime Sevilla and Tamay Besiroglu).
Where repeat founders worked before their first founding
Before starting their first organization, 140 of the 221 repeat founders (63%) held a dated non-degree work role at 229 parent organizations, and 81 (37%) have no dated prior work role before their first founding. Among the 915 one-time founders, 683 (75%) held a dated prior work role across 779 parent organizations and 232 (25%) have no dated prior work role before their founding. Repeat founders with dated prior work spent a median of 9 years in prior roles before their first founding (middle half 4 to 15 years), shorter than the median of 12 years (middle half 7 to 20) for one-time founders with prior work.
The sector mix of prior work differs between repeat founders and one-time founders. Among the 140 repeat founders with dated prior work, universities lead at 54.3% (76 people, including 41, or 29%, who worked only at universities before founding), followed by companies at 52.9% (74 people, 49 only at companies) and institutes and labs at 27.1% (38 people, 11 only at institutes and labs), with 39 repeat founders (28%) having worked in two or more sectors before their first founding. Among the 683 one-time founders with dated prior work, universities lead at 72% (494 people, 244 only at universities), followed by companies at 47% (321 people) and institutes and labs at 27% (183 people). Repeat founders with prior company work waited a median of 5 years between their first and second foundings, those with prior institute or lab work waited a median of 6.5 years, and those with prior university work waited a median of 7 years.
Horizontal bar chart of the 8 parent organizations where the most repeat founders worked before their first founding. The data table lists all 15 organizations with at least 3 repeat founders.
Source: People of Superintelligence career records from biographies and OpenAlex, as of 22 October 2026.
Data table
Across individual prior employers, Stanford University leads with 15 repeat founders (25% of the 59 founders who worked there before their first founding: Andre Esteva, Ashutosh Saxena, Astro Teller, Barbara Hayes-Roth, Douglas B. Lenat, Fei-Fei Li, Gordon Wetzstein, Ian Bremmer, Jay M. Tenenbaum, Jerry Kaplan, Julien Chaumond, Kunle Olukotun, Louis Rosenberg, Richard Socher and Timnit Gebru), followed by Massachusetts Institute of Technology (12 of 75, 16%) with 12 repeat founders. Among the 15 parent organizations where at least 3 repeat founders worked before their first founding, McGill University has the highest repeat-founder share of alumni founders at 43% (3 of 7).
Career data comes from the biographies on this site and from OpenAlex author records, as of 22 October 2026. People whose OpenAlex record may be incomplete or mixed are left out of citation figures.