Pattern analysis
Distributions across all 15 people, and the claims those distributions actually support. Every insight below is computed from the dataset at page load and prints the numbers behind it — none of them are written by hand.
What the data says
Each card states its own supporting count so you can check it against the table.
Technology holds 46% of the wealth in this dataset
high7 of 15 people work in Technology, but they account for 46% of the combined net worth — headcount and wealth concentration are not the same thing.
Based on 7 of 15 · 46%
STEM study is near-universal in tech here (100%) and a minority outside it (17%)
high10 of 15 people studied a STEM field. Within tech-adjacent industries that is 9 of 9; outside them it is 1 of 6. A technical education tracks the industry, not wealth itself.
Based on 10 of 15 · 67%
8 of 15 left at least one programme without finishing it
high11 also completed a degree — the two groups overlap heavily, because most of the "dropouts" here abandoned a second or third programme after finishing a first. The pattern is leaving an unfinished credential for a specific opportunity already in hand, not skipping education.
Based on 8 of 15 · 53%
87% are based in United States
high13 of 15. This partly reflects where deep capital markets and scalable consumer platforms are, and partly reflects the sampling: this dataset is drawn from the top of a Forbes list that skews toward US-listed equity.
Based on 13 of 15 · 87%
Founders sit at a higher median than heirs ($236B vs $125B)
mediumAcross 9 founders and 3 heirs, the founder median is 1.9× the heir median. Inheriting reliably places you at this tier; founding rarely does, but the survivors who make it cluster higher. This dataset only contains survivors, so it cannot tell you the odds of either route.
Based on 9 of 15 · 60%
Inherited wealth clusters in Retail
medium4 of 15 people inherited or partly inherited their position, and 3 of those are in Retail — 3 of the 3 Retail entries in this dataset. Established consumer businesses pass down; newly created platforms have not had time to.
Based on 4 of 15 · 27%
"long-term thinking" is the single most widely shared mental model
mediumAttributed to 11 of 15 people across every industry in the dataset — it is the one habit that does not track industry, wealth type, or education. Note that practising it is not free: choosing a worse near-term outcome requires being able to survive one.
Based on 11 of 15 · 73%
Everyone in Retail here shows "long-term thinking"
lowAll 3 people in Retail are credited with "long-term thinking", against 73% across the whole dataset. Industry structure appears to select for particular kinds of reasoning — or, just as likely, our reading of each person is coloured by knowing what industry they are in.
Based on 3 of 3 · 100%
Insights appear only when their threshold is met by the data. Adding or removing people changes which cards render — if a claim stops being true of the dataset, it stops being shown.
Industry distribution
Number of people per industry.
Six industries across 15 people. Compare this against the wealth view to the right — they rank differently.
Wealth by industry
Combined net worth per industry, in USD billions.
Same people, weighted by money instead of headcount. A single person can move an entire industry to the top here, which is exactly what has happened.
STEM vs non-STEM
Share with a documented STEM education.
- STEM10
- Non-STEM5
Read this together with the industry chart, not on its own. STEM study predicts which industry someone entered far better than it predicts wealth — the non-STEM fortunes here are in retail, luxury and investing.
Founder, operator, investor or heir
How the fortune was primarily built.
Hover any label on the table page for a definition. “Mixed” means more than one route contributed comparably — here, inheriting a business and then transforming it.
Wealth source
Whether the wealth was built, inherited, or inherited and grown.
“Self-made” is a Forbes-style label about building rather than inheriting a fortune. It says nothing about starting conditions, which varied widely even within this group.
Fields of study
Distinct fields across all recorded degrees, counted once per person.
People who studied more than one field appear under each, so these bars sum to more than 15.
Country distribution
Country of citizenship or primary residence.
This is mostly a fact about the sample. The list this dataset is drawn from skews heavily toward US-listed equity, so the concentration is partly real and partly selection.
Age distribution
Current age, in decade buckets.
Skewed old, which is what compounding looks like: these are fortunes that have had decades to grow, not fortunes built quickly.
Mental model frequency
How often each mental model is attributed.
The most interpretive chart on this page. It measures what we could find public evidence for — which correlates with how much someone has written and been interviewed, not only with how they think.
Skill categories
How often each skill category appears in a skill stack.
Almost nobody here has a single-skill stack. The combinations are the interesting part — see the Learnable Patterns page.
What 15 people cannot tell you
Every chart here describes a sample of 15 chosen for being at the very top of one wealth list. That makes it useful for spotting what these particular fortunes have in common, and useless for causal claims. There is no control group: the people who studied the same subjects, built the same skills, and did not become billionaires are not in any dataset. Read these distributions as a description of who is here, never as a formula for getting here.