WeWork Didn't Die of Ego. It Died of a Mismatch No CEO Could Fix.

WeWork signed a lease for fifteen years and rented you the desk for one month. That is the whole story, compressed into a single sentence. The landlord got a tenant who could not leave; the member got a tenant who could leave whenever they liked. WeWork stood in the middle, holding the long obligation in one hand and the short one in the other, betting forever that the two would never drift apart. In November 2023 they had drifted far enough that the company filed for bankruptcy in New Jersey, listing $18.65 billion in debts against $15.06 billion in assets.⁴¹ The gap that killed it was sitting in plain sight the entire time.

The official autopsy is about a founder. A charismatic CEO, a private jet, a $47 billion vanity number, a governance scandal that scared off Wall Street. All of that happened, and almost none of it is the cause of death. WeWork's problem was never the man at the top. It was the shape of the balance sheet underneath him.

It sold a short product on the back of a long promise

Strip the beer taps and the 'elevate the world's consciousness' language away, and WeWork was an arbitrage on duration. It committed to landlords for years, often a decade or more, then sold that same square footage to members in slices they could cancel almost at will. In good times this is a beautiful machine: you lock in your cost, you ride rising demand, you keep the spread. In a downturn it is the worst possible structure in commercial real estate, because you keep paying the long lease while the short member simply walks. WeWork held all of the downside of an empty building and none of the upside of a committed tenant. The revenue was recession-sensitive; the obligations were not. That asymmetry is not a flaw a better CEO fixes. It is the business.

Now run the numbers the S-1 finally put on the record. Revenue grew from $436 million in 2016 to $1.822 billion in 2018 — a 318% climb that looked like a software company finding its rocket.³ But losses grew almost as fast, from $430 million to $1.611 billion over the same stretch.³ This is the tell. A scaling business should lose less per dollar of revenue as it grows; WeWork lost roughly as much. Every new location added more long-dated lease liability than it added durable profit, so growth did not relieve the trap — it deepened it. By the first half of 2019 the company booked $1.535 billion in revenue and still lost $690 million.³ The faster it ran, the bigger the hole.

The $47 billion was a number SoftBank told itself

Here is where the founder myth and the real mechanism part company. WeWork is endlessly described as 'the company once valued at $47 billion.' It was — but only in a private room. That figure was set in January 2019, when SoftBank put in another $2 billion.⁵ No public market ever ratified it. According to CNBC, SoftBank's own successive rounds are what carried the valuation from roughly $17 billion up to $47 billion — a price one motivated investor printed by reinvesting into its own prior mark.⁸ When the S-1 went out to the actual market on August 14, 2019, the same market that had been quoted dreamy numbers looked at the losses and the lease liabilities and balked.² The IPO was withdrawn on September 30, 2019.⁵ The valuation did not collapse because sentiment turned. It collapsed because, for the first time, people who were not SoftBank had to underwrite it — and the unit economics underneath could not carry the weight.

The S-1 is also where a famous embellishment needs correcting. The phrase everyone remembers — 'community-adjusted EBITDA,' the metric that backed out rent and marketing and salaries to make losses look like profit — did not actually appear in that August 2019 filing. It came from earlier private investor materials, including a 2018 bond document; the S-1 itself used drier labels like 'Adjusted EBITDA before Growth Investments.'⁶ The detail matters because it shows the pattern: the most flattering math lived in the private rooms, where SoftBank set the price, and got quietly sanded down the moment the document had to face a regulator and a public order book.

Wasn't this just a runaway founder, though?

The honest objection is that the governance really was a mess, and it really did blow up the IPO. The dual-class voting, the self-dealing, the exit package reported at up to $1.7 billion — these were not invented by critics. But notice two things. First, even that package is widely overstated: Fortune later corrected the record to say only about half of the up-to-$1.7 billion was actually executed, and WeWork's own Executive Chairman called the billion-dollar-walkaway story 'totally false.'⁷ Second, and more important: imagine a saintly founder and a pristine board. They would still have signed fifteen-year leases against month-to-month revenue. They would still have lost $1.611 billion on $1.822 billion in 2018.³ They would still have walked into a 2020 office-demand collapse holding all the long obligations. Good governance might have produced a smaller, slower failure. It could not have changed the arithmetic. The founder was the headline. The duration mismatch was the cause.

WeWork went public in October 2021 the unglamorous way — a SPAC, at roughly $9 billion in enterprise value, about a fifth of the number SoftBank had once printed for itself.⁵ Two years later it was in bankruptcy court.¹ The tidy lesson is 'beware charismatic founders,' and it is the wrong one, because it lets every disciplined operator off the hook. The real lesson is colder and more useful. A company can be perfectly run and still be structurally doomed if it owes long and earns short. WeWork rented out the future one month at a time, while paying for it fifteen years in advance — and waited, the way that arrangement always must, for the one bad year that would never let it catch up.