What You’ll Find This Week

Quick question before this week's article: who does your company actually work for?

Not who you serve. Not who's named in the mission statement. Who you're structurally bound to put first. Most founders have never answered that question, because the standard incorporation stack answered it for them before they thought to ask.

This week, I'm looking at what happened when the most famous attempt to escape that default got unwound in public, and at the companies that built escape hatches that actually held: Vanguard, Bosch, Rolex, and a grocery chain whose CEO needed $1.5 billion to buy back his own mission.

Here’s what you’ll find:

Don’t Miss Our Latest Podcast

This Week’s Article

Do you know who your company is beholden to? Shareholders.

Not customers. Not employees. Not the mission statement you agonized over before committing it to your pitch deck. Nobody sat down and chose this. It's a default so totally ingrained that it stops feeling like a choice. Like gravity.

Most founders never think about this because the answer is virtually decided for you before you ever think to question it. The decision is baked into your incorporation documents, your board's reflexes, and your investors' expectations.

Except gravity can be beaten. A handful of companies have done it. Eric Ries calls these companies incorruptible: engineered so that when the pressure comes to trade the mission for money, it’s impossible. These companies beat gravity by building protections into the corporate structure itself.

OpenAI was supposed to be incorruptible. A nonprofit sat on top of the for-profit, holding the leash, so no investor and no market could ever bend the mission toward money. That structure held for about a decade. Then it all fell apart last October. The for-profit became a public benefit corporation. The nonprofit that was supposed to be in control was left with only a minority stake. It still appoints the board, but control that once rested on owning everything now rests on terms that can be renegotiated, and this restructuring is the proof. By this spring, the company was reportedly prepping an IPO at more than a Trillion dollar (yes, with a “T”) valuation.

OpenAI began as a charity. Donations funded it, and its assets were legally pledged to the benefit of humanity. The restructuring moved 74% of the economics to Microsoft, investors, and employees, and left the foundation holding 26% of a company it used to own outright. The people who engineered the shift are the same people positioned to profit from it.

And to make matters worse, somewhere along the way the word "safely" was quietly dropped from the mission. Nobody had to break a single rule to pull it off. Two state attorneys general reviewed the deal and waved it through. The shift happened in the open, in press releases, with lawyers, blessed by the regulators who were supposed to be the last line of defense.

The rules were never as strong as the pull. The structure had a leak. Gravity found it.

What happened to OpenAI happens to almost every company. But it usually happens quietly, at incorporation, buried deep in paperwork nobody reads. There's no drama, because there's no resistance.

Every company is a fiduciary to someone: the party whose interest they’re bound to put first. A lawyer has a client. A trustee has a beneficiary. And the standard incorporation stack names yours before you've ever thought to ask: shareholders. Everything else, customers, employees, the mission, becomes a means to their end.

Most founders assume this is the law. But in reality, there is no statute that demands you squeeze customers for the last dollar. The pressure comes from consensus: your lawyer, your board, and your investors all treat shareholder primacy as the only “adult” way to run a business. And that's what makes it so durable. A law is written down: you can challenge it, amend it, repeal it. This rule is distributed: copied into every charter, every term sheet, every board's reflexes, deal after deal, before anyone thinks to question it. There's nothing to strike down, because the rule is written everywhere and legislated nowhere.

Once you become aware of this pattern, you can't help but see it everywhere. I see it most in the place I spend all my time: innovation.

Ask why big companies are so bad at it and you'll get the usual list. Bureaucracy. Risk aversion. No talent. No budget. I've written whole pieces about some of these, and they're real. They're also symptoms. Underneath all of them sits the same gravity.

Companies innovate constantly, as long as the innovation is additive. New products, new markets, new revenue streams: all of it gets funded, because none of it requires an existing, profitable product to die. The gravity shows up when the new business competes with the existing one. Kodak invented the digital camera in 1975 and put it in a drawer, because every digital sale would have come out of film's margins. Blockbuster had the chance to buy Netflix for $50 million and passed, because streaming would have eaten the late fees that made their revenue model work. Both had the idea in the building. What they couldn't do was put the future customer ahead of this quarter's shareholder.

The default of shareholder primacy forbids it.

So the company swears it wants innovation, funds a lab, hires the team, and then strangles every initiative that threatens the core. There is no innovation problem. The fiduciary is the problem.

So if the pull is that strong, can a real business resist it? One of the most successful financial companies in history was built to.

Vanguard runs more than eight trillion dollars. Its structure is almost stupidly simple, and almost nobody copies it. Vanguard is owned by its own funds, and those funds are owned by the people invested in them. The customers own the company. There is no outside shareholder with a hand out.

John Bogle, Vanguard's founder, built it that way on purpose. The structure became the weapon. With no outside owners to pay, Vanguard runs its funds at cost. The fees still exist, but they only have to cover operations, with no profit margin stacked on top for somebody else. That gap compounds: lower fees pulled in more money, more money meant more scale, and more scale let fees fall further. Vanguard didn't outsell the industry. It out-structured it. In 2000, the average stock fund charged 0.99%. Today it charges 0.40%, and analysts have a name for the force behind that collapse: the Vanguard effect. Fidelity got dragged all the way to the floor, launching index funds that charge nothing just to compete.

This is what people miss about choosing the right fiduciary: it's a moat. Fidelity can zero out fees on a handful of funds and subsidize them from everything else it sells. Copying Vanguard outright means running the entire firm at cost, and that requires handing the company to your customers. No owner votes for that.

Hold Vanguard next to OpenAI. Both built structures meant to keep the mission in charge. The difference is who profits from breaking them. At OpenAI, dismantling the structure made the people in the room richer, so they dismantled it. At Vanguard, dismantling the structure would mean customers voting to raise their own fees. A structure survives when the people who could break it are the same people it protects.

Market Basket ran the right hierarchy for decades: customers and employees first, shareholders last. In 2014, the board fired CEO Arthur T. Demoulas for his dedication to that fiduciary choice, and ultimately what saved him wasn't governance. It was 25,000 employees walking off the job, customers boycotting 71 stores, revenue collapsing 90%, and nearly $600 million in losses before the board sold him the company back for $1.5 billion. I've written about it before.

Loyalty won. Then last year the same fight broke out again: Arthur T. placed on leave, longtime executives fired, the same gravity pulling at the same company. The right fiduciary, defended by culture alone, has to win every round. Structure only has to be written once.

That's the strange part. None of this is a startup experiment.

Rolex has no shareholders. A foundation set up by its founder owns the whole thing, which is why it pours money into the watches instead of dividends and answers to no one on Wall Street. Robert Bosch wrote the same arrangement into his will almost a century ago, so that no buyer, no heir, and no shareholder could take control of the company after he was gone. The dividends flow to a charitable foundation that holds no votes. The votes sit with trustees who can't pocket the profits. Money and power, deliberately separated. Tata, one of the biggest enterprises on earth, is about two-thirds owned by charitable trusts.

These aren't curiosities. They're among the most durable companies alive, and most founders have never heard of the structures holding them up. Nobody is hiding them, either. The people who guide founders just have no reason to bring them up. Investors, lawyers, and accelerators all get paid through the standard model. There's no carry in a company built for longevity over shareholders.

I won't pretend this is free. It isn't, and the research is blunt about the bill.

Companies built to be incorruptible tend to grow more slowly. Study after study of foundation-owned firms lands in the same place: less top-line growth, but far longer survival (one line of research puts it near 600% higher odds of reaching forty years), leaner debt, steadier headcount, less management churn. And the earnings hold up. Returns on equity match family- and investor-owned peers, with far less volatility along the way. The price is the growth rate, not the profits. You're trading a steeper curve for a longer life.

But most founders today are never offered that trade. The grow-at-all-costs mandate comes from the capital, not the founder: a venture fund has to return its money inside roughly a decade, and that math requires fast growth, an exit, and a shareholder-first charter to protect both. Plenty of founders would happily choose the longer life. But the capital on offer arrives with the wrong fiduciary already written into the term sheet.

The structure isn't a force field either. Mozilla is the cautionary tale. It's built exactly the way the theory says, a nonprofit sitting on top of the company that makes Firefox, and the structure did protect the mission. Nobody bought Mozilla, nobody stripped it for parts, nobody bent it toward an exit. What the structure couldn't do was write a business model. Mozilla lives on a single check: Google pays to be Firefox's default search engine, and that one deal has supplied most of Mozilla's revenue for two decades. When the check wobbles, layoffs follow, and they have, repeatedly. A structure can keep you from being captured. It can't make anyone buy what you sell.

The gravity is real. OpenAI is proof that even the most deliberate attempt to escape shareholder primacy gets toppled once the numbers grow big enough. But it isn't destiny. Vanguard turned it into a weapon. Rolex and Bosch and Tata have held the line for generations. What separated them was whether the caring was load-bearing before the pressure arrived.

Eric asks founders one question I can't shake: Who would you rather die than betray?

Your customers? Your employees? The mission on the wall? If you have an answer, you already have a fiduciary. The only question left is whether you wrote it down anywhere that protects it. In the documents. In the ownership. In the structure that's still standing when you're not in the room.

Because if you didn't, the default already answered for you. You work for the shareholders. And the day you decide to change that is the only chance you get, because it is always too early, right until the morning it's too late.