Can You Still Build a 100-Year Company?

The case for permanence in an economy addicted to exits

In 578, a Korean carpenter named Shigemitsu Kongō crossed the sea to Japan to build Shitennō-ji, one of the first Buddhist temples in the country. The firm he started to do the job - Kongō Gumi - kept building and repairing temples for the next fourteen hundred years, through forty generations of the same family. It was already centuries old when Marco Polo was born. It outlasted the samurai. It survived the Meiji government’s deliberate campaign to dismantle Buddhism, which tore down tens of thousands of temples, and it was still taking commissions after the war levelled most of Osaka in 1945.

Now hold that up against the rest of us.

It’s not a flattering comparison.

Most companies don’t outlive their seed round. The average tenure of a company on the S&P 500 has fallen for decades, from around sixty years in the 1950s to under twenty today, and that’s the survivorship-biased top of the market, with everything below it doing much, much worse.

By any honest reading of the numbers, your odds of building a true hundred-year company are abysmal.

Your odds of building anything resembling Kongō Gumi are probably worse than zero.

But I want to argue the case for trying anyway; in fact, I want to make a case for aspirational permanence.

What wanting it gets you, which is nothing

First: a dose of realism.

In 2008, in the midst of a collapsing economy, a note-taking app launched against the grain of everything venture capital wanted to fund. Investors were chasing social networks and ad-supported web apps; Evernote offered neither. Its CEO, Phil Libin, called it the anti-social network, built for the ninety per cent of your life when you aren’t bragging about your dog or your holiday. Libin promised, repeatedly, that Evernote was going to be a hundred-year company, and that a company thinking in generations behaved differently from one built to be flipped. He invested early and globally, shipped native apps for every major platform, and treated the whole sprawling problem of human memory as his to own.

For a while it worked. I was an early user, obsessed with that green elephant. Evernote crossed a hundred million dollars in recurring revenue, won millions of paying customers, passed 250 million users, and was briefly a billion-dollar company; for a few years nothing else in the category came close. Libin had the intention and the rhetoric, and (critically) the believers.

Evernote is now owned by a holding company with a reputation for buying and gutting software businesses. Bending Spoons closed its acquisition in early 2023, and within months it cut 129 staff, telling the press the app had been unprofitable for years and was unsustainable. A few months after that it laid off nearly all the remaining US and Chile employees and moved the operation to Europe. Libin’s successors had leant on a consumer freemium model while Notion shipped the collaboration features the market actually wanted, and they bled executives the whole way down.

Kongō Gumi’s owners never said a word about lasting, but they lasted fourteen centuries. Evernote built its entire identity around lasting a hundred years, and couldn’t hold it for fifteen.

Wanting it, clearly, doesn’t get you there.

What does?

What actually correlates with lasting

In fairness, the Kongō Gumi story is a survivorship anecdote.

For every Kongō Gumi there are ten thousand temple-builders who started in the same century and are now a smear in the archaeological record.

But it’s still worth asking: survivorship aside, what correlates with longevity?

Look at the rare firms that clear a hundred years and they share a handful of unglamorous traits. They stayed inside their circle of competence; they avoided fatal debt; they preserved a recognisable culture after the founder died. They also carried the trait the business books tend to bury: they weren’t standing in the wrong place when something tried to kill them.

Kongō Gumi ended up proving the rule by breaking it. During Japan’s bubble economy in the 1980s, the firm borrowed to speculate in real estate, and when the bubble burst, the debt did what fourteen centuries of war and fire couldn’t. In January 2006, carrying ¥4 billion in liabilities, the family handed the business over to the Takamatsu Construction Group, where it survives today as a subsidiary. The company that outlasted the atom bomb couldn’t outlast a property loan.

We credit the survivors with skill, because we like a story with a hero in it. But the founders who reach a hundred years seldom won a hundred battles. More often, they never lost the single battle that would have ended them, and never took the blow when they were too weak to absorb it. They won because they avoided dying, and they got lucky.

Skill, survival, luck.

You can do something about skill. You can do something about survival. You can’t manufacture luck, but you can read the weather.

And I still think the weather right now, against every intuition, is looking damned good for founders who actually want to last.

Why the thing killing companies is good news

The world is fast and getting faster; cruel and getting crueller. How could that possibly help anyone build something that lasts?

We reinvent whole technologies, and make them obsolete, in a few short years now rather than a generation. A rival can copy or leapfrog your hard-won technical edge before you’ve finished congratulating yourself. At that pace, surely, all incumbents die sooner or later.

But that same pace is why you should crave permanence. If your average competitor is built to win a five-year cycle and then evaporate, the rarest thing on the field is a company that intends to stay, and the churn that buries everyone else is what makes the founders who last worth betting on.

The fair objection: why should anyone want a company to last a hundred years? Most shouldn’t last ten. Joseph Schumpeter named the churn creative destruction in 1942, and he was right; an economy where firms never die is an economy where capital and talent stay trapped in yesterday’s ideas. But a company that dies because something better replaced it and a company that dies because its owners built it to be flipped are different deaths. One recycles value. The other strip-mines it. Immortality for its own sake is a vanity project; the prize is the option to stay whenever staying is the better trade, and almost nobody is building for that option.

Since 2010 or so, tech founders and their backers played one game: blitzscaling. Raise an enormous pile of money, grow at any cost, capture the market, and worry about durability some other decade. They won often enough to build some of the largest companies in history; but they also built a vast field of organisations designed to be sold rather than to last, and they treated their customers and their own people as ore to strip-mine on the way to an exit.

And people can tell.

Customers know the difference between a company built to serve them and one built to be flipped to a private equity firm in eighteen months. Employees want to work somewhere that will still exist, and still recognise itself, in ten years.

Those customers and those employees are your opening. In a market drowning in the disposable, the people you most want to hire notice the founders who plan to stay. Founders who build for a century are making a competitive bet as much as a values statement: they tell their best people that the work will compound instead of evaporate, and they prove to customers that they’ll still honour the deal after the next funding round. “We plan to be here” is, in a disposable economy, a unique selling point.

The conditions are generous, in other words, to anyone who gives enough of a shit to stick around. The remaining question is what sticking around looks like. As I see it, there are four pieces, and they have to hold together.

What a tiny team can do now

The first piece is staying adaptable, and the conditions for that have never been better.

Founders today can do what needed a department a decade ago and an army a generation ago. They pay a fraction of what software, analysis, design, legal drafts, customer support, and research used to cost. A handful of people with good judgment and modern tools now operate at a scale that used to demand enormous headcount, and the bureaucracy that headcount drags behind it.

Old companies die when they lose the ability to adapt, piling on process after process until the org chart becomes a monument to problems they solved a decade earlier. They grow so attached to the edge that once made them successful that it becomes the liability they can no longer bring themselves to change.

Every hire adds drag. The more people you bring on, the more they have to coordinate; the more they coordinate, the more process they write; the more process they write, the slower they all turn. Craigslist showed what refusing that drag looks like: Craig Newmark’s classifieds site served tens of millions of people for three decades on a headcount of around fifty, and it did so before the current generation of tools existed. Founders today can hold the line at four people doing the work of forty, and stay small, and therefore nimble, far deeper into their growth than Newmark ever could.

Lean teams stay adaptable, and founders who stay adaptable across a hundred-year horizon outlast the ones who bet everything on a single product or market. Nobody can know what the world looks like in 2075, but the goal was never to predict the future. It was to build a team that can become something it can’t yet imagine. Small, automated teams turn faster, and they get more chances to become the next thing before the current thing kills them.

Change everything but the core

The second piece pulls the other way. Founders who can only adapt, and never commit, court a different death: they build a company that can become anything and means nothing. You have to be willing to change almost everything except the few things you know, in your core, you must never change.

The people who built every enduring institution held one reason for being fixed while they rotated products and markets around it. The folks at Nintendo made playing cards for most of a century before they touched a video game. Fusajiro Yamauchi started the firm in 1889 selling handmade hanafuda cards in Kyoto; the first console didn’t arrive until the late 1970s. Across that whole stretch they treated the product as negotiable and the purpose, play, as fixed.

Founders keep falling in love with the scaffolding, or the favoured product, or the clever business model, or the current tech stack, and treating their actual purpose as optional, which is how they get killed in a platform shift. When the product goes obsolete, founders who organised everything around that product have nothing left to be, while founders who organised everything around a purpose just build the next form.

If you want a century, the most important early work is to name the small handful of commitments you’d rather shut the company down than abandon, then treat everything else as temporary.

Money on a hundred-year clock

Even a lean team with a fixed core dies if the people writing the cheques are on a five-year clock.

The people who run traditional venture capital work against century-building almost by design. They operate on fixed fund lifecycles and forced exits, and their portfolio math needs most winners to either go exponential or die fast, because they have to return capital to their own investors on a clock. A founder who wants to grow steadily and profitably for forty unspectacular years before becoming formidable is a bug those investors can’t tolerate.

But the people writing cheques have diversified. More investors now prize cash flow over growth-at-all-costs. More founders bootstrap past the point where they’d once have raised; Valve has stayed private since Gabe Newell and Mike Harrington founded it in 1996, and no fund lifecycle could force its hand while it spent years turning Steam into the default store for PC games. Some investors have built holding companies and evergreen funds so they can hold things rather than flip them. And because founders now pay so much less to operate, they need far less outside money to reach durability.

When founders need less capital, they keep more control. Reach sustainability without handing equity to investors on a strict timer and you keep the one thing a hundred-year company can’t do without: the freedom to make decisions on a long horizon. Founders seldom die because they ran out of options. They die because they sold the right to choose before the long game could pay off.

The only thing that compounds

Stay lean, stay fluid, keep control, fix your irreducible core. Do all of it and you still have to protect the asset no one can buy from you.

Products decay, patents expire, and rivals erode a market position faster now than ever before. The one thing that compounds across generations is culture: the shared, mostly unspoken agreement among your people about why the work gets done and how it gets done differently here. People can destroy a culture faster than you can blink, and to keep one alive past your own exit you have to engineer it on purpose. Codify it, then live it, through who you hire, who you promote, what you celebrate, and what you refuse to tolerate.

The people at every century-old company passed this test in some form; they solved succession. They worked out how to hand their essence to people who never met the founder, who joined for reasons the founder never anticipated, working in conditions the founder never saw. The Kongō family handled it with a flexibility most modern boards would envy: when a generation lacked a capable heir, they brought in a son-in-law and gave him the Kongō name, choosing the steward over the bloodline. The logistics were never the hard part. The hard part was carrying purpose and identity intact across generations of strangers.

Your people are the whole game now. Once the tools handle the doing, what they still own is judgment, taste, and character, the intangibles nobody can copy overnight.

The verdict

Can you build a hundred-year company today? Yes. Not easily, and not without an unfair amount of luck you can’t manufacture. Kongō Gumi survived the long centuries partly by being good and partly by not standing where the worst blows fell, and then a single bad decade of borrowing took it down anyway. Skill plus survival plus luck, and you only control the first two.

But the conditions favour you now in a way they didn’t a generation ago. With modern tools, a tiny team stays adaptable far longer. Your customers and your best hires, worn out by disposable companies, reward you for planning to last. The money has diversified enough that some of it will fund your patience. And when any rival can clone your technical edge in months, you compete on the assets they can’t copy once everything copyable has been copied: the trust and judgment your people carry.

You can’t get to a hundred years as the accidental byproduct of a great five-year exit, and as the folks at Evernote learned, you won’t get there just by announcing you want to either. You can choose to aim for it at the start, but you have to renew that choice every day, for forty years, against every incentive pulling the other way.

My version of this is small on purpose. I run a one-person studio, and I hold no illusions that an agency with my name on it outlives me; agencies are scaffolding, and scaffolding comes down. The bet I’m making is on the writing. Essays compound the way temples do: they get repaired, requoted, argued with, and handed on to people who never met the author and who joined the argument for reasons the author never anticipated. The studio pays for the work. The work is what I expect to still be standing in a hundred years.

Shigemitsu Kongō, on the day he stepped off the boat in 578, wasn’t writing a hundred-year plan; he was a carpenter taking a temple commission. Forty generations of his family did the rest.

What are we going to do, in our turn?