Betting the company: when one move is big enough to sink you

The plane that cost three Boeings

In 1965 Boeing was worth about $375 million. It agreed to build a plane that would cost somewhere between $1.2 and $2 billion to develop: more than three times the value of the whole company.

There was no cash to pay for it. There was only debt, which climbed past $2 billion, with $1.2 billion of it owed to banks, a borrowing record for any company at the time. What made the bet thinkable was one customer: on April 13, 1966, Pan Am ordered 25 of the new 747s for $525 million.

Then things went wrong at once. Engine failures left up to 20 finished aircraft sitting at Everett with no working engines. Between early 1970 and October 1971, Boeing's workforce fell from about 80,400 to roughly 37,200. Its president later put it plainly.

“It was really too large a project for us.”William Allen, Boeing's president, as recorded in the Boeing 747 case

Boeing survived. The 747 paid down its debt and flew for fifty years. Pan Am, which had staked its own balance sheet on the same plane, lost $26.42 million in 1969, before its first paying 747 passenger had even boarded. It sold its Pacific division in 1985 and ceased operations in 1991.

Same plane, same years, two companies that each bet everything. That is the decision in one story: a move so large the company may not survive it failing, where surviving still doesn't guarantee winning. Most of the bets in this guide were never undone, and many of those still went badly. The rest of this guide is about what separated the bets that paid from the ones that merely stood.

What bet-the-company decision is

Every large investment carries risk. What makes this one different is where the downside lands: not on a product line or a quarter, but on the firm. That changes who should decide, how the bet should be sized, and what has to be written down before the money goes out.

Why it is so hard

On paper the rule is simple: bet big only when the odds are good. In a real company, three things get in the way.

The ways companies did it

The bets on record take four shapes. Each puts the company at risk in a different way.

What 22 companies did

Here is every decision on record, placed by the year it was made and marked by how it turned out. Hover a mark, or open the table, for the company and the outcome.

Of the 21 decisions that can be judged, 18 were never undone, 1 was abandoned and 2 were kept only in part.

The bet abandoned outright was Ford's. Its electric truck was not a bad product: in the fourth quarter of 2023 the Lightning held 72% of the electric-pickup segment. The segment was simply far smaller than 150,000 trucks a year. Ford was right about the direction and wrong about the slope, and it had built a factory for the slope. It ended production of the all-electric truck in December 2025, with $19.5 billion in special charges.

Microsoft's OpenAI bet, recorded as two decisions in 2019 and 2023, was not abandoned; it was reopened. A bet on a partner you don't control is a bet whose terms the partner can bring back to the table. In October 2025 Microsoft gave up its right of first refusal on OpenAI's new capacity, and its stake fell to roughly 27%, in exchange for OpenAI's commitment to $250 billion more of Azure.

Never undone is not the same as won. Pan Am's jumbo order stood, and the airline collapsed. Boeing's 2010 choice to re-engine the 737 rather than build a new plane stood too, and the MAX was grounded worldwide from March 2019 to November 2020. Airbus kept building the A380 until its orders were delivered; 251 were, against a forecast of nearly 1,200 very large aircraft. Intel kept IDM 2.0 while its 18A process was reportedly stuck below 10% yields in late 2024. Beyond Meat stayed committed to scale while its revenue and stock collapsed.

The bets that held and paid were more often sized to something already paid for. Boeing's 747 had Pan Am's order for 25 aircraft. Tesla built the expensive Model 3 variants first and sold the $35,000 car only from February 28, 2019, after the ramp. Eli Lilly's obesity bet was a filing for a molecule already approved for diabetes, not a new plant.

What the share price did

For many of the decisions there is a clean market record: the company's total return over the three years after the decision, against the S&P 500.

Three years is a short window for bets that play out over a decade. Boeing's 2010 decision to re-engine the 737 beat the index by 63.1 points over the next three years. The two fatal crashes that grounded the MAX came in October 2018 and March 2019, long after the window had closed. Meta's metaverse bet trailed the index by 53.6 points after one year and beat it by 56.6 after three, while its losses kept growing. The share price measures the company, not the bet.

How to run this decision

If you are weighing a bet of this size, here is what the record suggests doing, in order.

1. Write the downside in the company's own size

   Not a percentage of revenue: what the company would be if the bet failed. Boeing's 747 cost more than three times what Boeing was worth. If the company cannot survive the downside, say that out loud before you approve it, not after.

   Seen in: Boeing, 1965 · Pan Am, 1966
2. Separate interest from orders in the model

   Reservations, waitlists, refundable deposits and survey intent are interest. Ford's reservations needed a $100 refundable deposit. Apply a conversion rate you can defend from something you have actually sold, and if you have never sold anything like it, say so.

   Seen in: Ford, 2022
3. Build capacity behind revenue where you can

   Tesla sold the expensive versions of the Model 3 first and used them to fund the road to the cheap one. That ordering is available more often than it is used, and it turns a bet on demand into a bet on execution.

   Seen in: Tesla, 2017
4. Size the part you can't undo to what has been shown

   The whole bet can be large. The part that cannot be unwound in a year should match demand you have seen, not the plan. Eli Lilly's obesity bet was a filing; Ford's was a factory. Ask which kind of commitment you are making.

   Seen in: Eli Lilly, 2022 · Ford, 2022
5. Name who is allowed to say stop, and give them a number

   Meta's losses grew every year and the bet stood. Beyond Meat stayed committed as revenue collapsed. Name the person, the measure and the date at the outset, while nobody's reputation is yet attached to the answer.

   Seen in: Meta, 2021 · Beyond Meat, 2021
6. Ask what it costs to be right but slow

   Nvidia's executives have said it took roughly ten years before Wall Street believed CUDA was worth anything. Ford was right about electric trucks and early by more than its factory could absorb. Being early is a different risk from being wrong, and it needs its own number.

   Seen in: Nvidia, 2006 · Ford, 2022
7. Commit to what can't be changed later, and ship the rest unfinished

   Apple launched the iPhone mispriced, without an App Store and tied to one carrier, then cut the price from $599 to $399 and apologized in public. It fixed those within about a year. The architecture was the part that could not be retrofitted.

   Seen in: Apple, iPhone, 2007

Warning signs

What it has looked like when this decision was going wrong.

• Capacity is sized to reservations, waitlists or a forecast of a whole new category.Ford, 2022 · Airbus, 2000
• The losses grow every year, and the plan says that is expected.Meta, 2021
• The bet is funded with more debt than the company is worth.Boeing, 1965
• The cheaper option is chosen precisely to avoid making a bet the moment calls for.Boeing, 737 MAX, 2010

All 22 cases

Each case is one company at one decision: what was on the table, what it chose, why, and what followed, with every fact sourced.

Questions

What is a bet-the-company decision?

A decision that commits so much money, debt or reputation to one move that if it fails, the company may not survive. Boeing's 747, which cost more than three times what Boeing was worth when it began, is the classic example.

What are examples of companies that bet the company?

Boeing with the 747, Pan Am with its order for 25 of them, Airbus with the A380, Apple with the iPhone, Nvidia with CUDA, Amazon with AWS, Meta with the metaverse and Ford with the electric F-150.

Why do bet-the-company decisions fail?

On record, the bet that was abandoned outright, Ford's electric F-150, was sized to refundable reservations rather than paid orders. Others stood and still went badly: Airbus sized the A380 to a forecast and ended production after 251 aircraft.

How do you reduce the risk of a big strategic bet?

Separate interest from paid orders, build capacity behind revenue where possible, keep the irreversible part of the commitment sized to what has been shown, and name in advance who can stop the bet and on what number.