What does proof look like?
Gabe Brown did not set out to prove anything. He lost his crop to hail four years running in the 1990s on a 5,000-acre North Dakota farm. Banks stopped lending. Necessity — not philosophy — forced him to stop buying fertilizer, stop tilling, and start paying attention to what the soil actually did when left alone.
Two decades later, the data: soil organic matter rose from 1.7% to 6.1%. Water infiltration went from half an inch per hour to over eight inches. Input costs dropped to near zero. The ranch runs profitably without a single government subsidy, in a county where most farms cannot survive without them.
I keep returning to Brown's story because it is so ordinary. A man went broke and tried something different. The something different worked. The USDA now studies his methods. Thousands of farms replicate them. But the timeline — a decade before the soil biology fully established — matters as much as the results. This is patient evidence. Quarterly reports cannot see it.
Here are fourteen more.
Who did it at corporate scale?
Will Harris, White Oak Pastures (Georgia). Inherited an industrial cattle operation. Spent twenty years converting it to a vertically integrated regenerative farm raising ten species. Revenue: over $20 million annually. A 2019 Quantis lifecycle assessment found the farm is a net carbon sink — its soil sequesters more carbon than its livestock emit. The industrial equivalent is a net emitter. The farm employs 160 people in a county where the next largest employer has fewer than 20.
Patagonia. Revenue at the time Yvon Chouinard transferred ownership to a climate trust: $1.7 billion. B Corp score: 166.0 out of 200 (median business: 50.9). The company spent 1% of sales on environmental causes, offered lifetime repair guarantees, and ran a Black Friday ad telling people not to buy its jacket. Sales increased 30%.
That last detail still stops me. A company told its customers to buy less, and they bought more. The usual explanation is "great marketing." The better explanation is that people are starving for something that means what it says.
Interface. In 1994, Ray Anderson — CEO of the world's largest commercial carpet tile manufacturer — read Paul Hawken's The Ecology of Commerce and publicly called his own company a destroyer. He launched Mission Zero: zero environmental impact by 2020. By 2019, greenhouse emissions down 96%. Waste-to-landfill down 91%. Water per unit down 89%. Stock price up approximately 7x.
The mechanism was simple and brutal. Every pound not wasted was a pound not purchased. Environmental accountability and cost reduction pointed in the same direction. Interface did not sacrifice margin for mission. They discovered margin inside mission.
Mondragon Corporation (Basque Country). A federation of worker-owned cooperatives, 80,000 worker-owners, $14 billion revenue. During 2008, when Spanish unemployment hit 30%, Mondragon's layoff rate stayed below 1%. Workers voted to cut their own pay rather than eliminate jobs. When the crisis passed, the full workforce was intact. Competitors who had gutted their teams recovered slower.
Pay ratio: 6:1, highest to lowest. The average large US corporation runs at roughly 350:1. Despite that compression, Mondragon attracts talent, maintains technological competitiveness, and has survived for seven decades. Something other than maximum extraction retains people.
Virta Health. Reversing Type 2 diabetes — a condition conventional medicine treats as chronic and progressive — through continuous remote monitoring and individualized nutrition. Published results: 60% remission at one year. No surgery. No new drugs. Annual US cost per diabetes patient: approximately $16,750. Virta's intervention: a fraction. Scale this and you remove hundreds of billions from healthcare expenditure while patients get healthier. Not a tradeoff. A correction.
Does it work at the scale of a nation?
Costa Rica. By the 1980s, forest cover had dropped from over 50% to roughly 25%. The government began paying landowners directly for maintaining forest, biodiversity, and watershed protection. By 2020, forest cover recovered to 52% — exceeding the 1940s baseline. Eco-tourism overtook banana exports as the largest source of foreign revenue. Over 98% of electricity now comes from renewable sources. Three decades. Multiple administrations. Measurably wealthier, healthier, more biodiverse than the extractive model produced.
Freiburg Vauban (Germany). A car-free, energy-positive district. 65% modal shift to cycling and walking. Buildings produce more energy than they consume. 5,500 residents. Completed 2006. Twenty years later, property values consistently exceed surrounding areas. The air smells different there — I noticed it within minutes of arriving. Less exhaust, more bakery.
Kerala, India. Per capita GDP roughly 1/50th of the United States. Literacy rate: 99%. Life expectancy: matches US figures. Achieved through sustained investment in public education and healthcare. This is the data point that refutes the claim that positive-sum thinking is a rich-country luxury. Kerala did it poor.
What about networks with no center?
Wikipedia. Over 60 million articles. 300+ languages. Zero paid content writers. It demolished the encyclopedia industry — Encyclopaedia Britannica, valued at roughly $1 billion at its peak, ceased print publication in 2012. Replaced by something free.
Linux. Runs 96.3% of the top one million web servers. Estimated development cost if built by a single corporation: $16.8 billion. Contributed by thousands of developers, most not paid by Linux itself. The largest cooperative engineering project in human history, and it has no CEO.
Bitcoin and Ethereum. Combined market capitalization exceeding $2.5 trillion. Millions of participants coordinating economic activity without a central authority. Trustless is the wrong word — it is trust distributed rather than trust concentrated.
Grameen Bank. Nine million borrowers, 97% women, 98% repayment rate. Higher repayment than most commercial banks. Muhammad Yunus won the Nobel for proving what the banking industry said was impossible: the poorest people on Earth are creditworthy when the structure is cooperative rather than extractive.
What do the patterns tell us?
Fifteen cases. Agriculture, manufacturing, healthcare, finance, governance, technology, national policy. A 5,000-acre ranch and a country of five million. Here is what repeats:
Every one started from crisis. Brown lost his crops. Anderson confronted his own company's destruction. Costa Rica deforested to ecological collapse. Transformation almost never begins from comfort. It begins when the old way stops working so completely that experimentation becomes cheaper than continuation.
Every one measured in decades. Brown's soil: ten years. Interface's Mission Zero: twenty-five years. Costa Rica's reforestation: thirty years. Any evaluation framework shorter than a decade will systematically miss these outcomes. Quarterly reporting is structurally blind to them.
Every one created more value than it captured. Linux created $16.8 billion in value and captured approximately zero. Wikipedia destroyed a billion-dollar industry and replaced it with something free. Patagonia grew while giving ownership away. The pattern is consistent: positive-sum systems generate surplus, not scarcity.
Every one built new supply chains. White Oak Pastures built its own processing facility. Interface redesigned manufacturing from raw materials forward. Costa Rica created entirely new economic sectors. This is not a surface change. It is infrastructure.
Every one became an attractor. Mondragon retains workers through crises. Patagonia receives far more applications than positions. Linux attracts the world's best kernel engineers for free. When a system creates more than it extracts, people want in.
What question remains?
We did not write this to inspire anyone. We wrote it to establish a factual baseline. The question of whether positive-sum systems work at scale is no longer open. It has been answered — by a carpet company, a ranch, a bank, a nation, an operating system, and an encyclopedia.
The remaining question is harder: what prevents this from working more often? What are the real failure modes, the structural limits, the honest objections?
That is next week. We intend to try to tear down everything we just built.