Peloton Built a Factory for a Pandemic That Ended

On December 21, 2020, Peloton agreed to pay $420 million for Precor, a 40-year-old commercial fitness equipment maker with over 625,000 square feet of factory space. Five months earlier its stock had traded at $33.

By the time the Precor deal closed, Peloton was worth more than Ford. The company was not buying a competitor. It was buying capacity, because it believed the demand in front of it was permanent.

I wrote about a similar instinct in Nike’s direct-to-consumer reversal: a company reading a temporary edge as a structural one, then spending real capital to defend a position that was already dissolving. Peloton’s version is starker, because the mistake was not strategic drift. It was a single, specific error: treating a demand spike as a moat.

This is piece four of The Moat That Broke, a series on advantages that failed and the mechanism that killed each one.

Weather Is Not a Moat

A moat is a structural reason demand keeps returning to you rather than a competitor, something a rival cannot copy quickly even if they have the money and the will. A demand spike is different.

It is a period where conditions outside your control push more people toward your category than usual. Both can look identical from inside the building, especially if the spike lasts long enough for a management team to build its planning assumptions around it.

The test that separates them is simple and almost nobody applies it in the moment: if the external condition reversed tomorrow, would the demand still be there. Gyms were closed. People were at home. Peloton’s bikes and treads were one of the few ways to exercise indoors with any social component at all.

None of that was Peloton’s doing, and none of it was permanent. It is a different failure mode from Intel losing its process lead to TSMC, where the moat was real and the company simply stopped reinvesting in it. Peloton’s problem was upstream of execution. It never had the moat it thought it was defending.

Retention looked like proof

Peloton did have something real underneath the spike. Once a household bought a bike, a meaningful share of them kept paying the monthly subscription, and Peloton’s own churn numbers were genuinely strong for a hardware-adjacent subscription business. That retention was the actual asset.

But retention among people who already own the hardware is a different quantity from new households wanting to buy the hardware in the first place, and Peloton’s leadership spent 2020 and 2021 building fixed costs against the second number using evidence from the first.

Fixed costs do not know the weather changed

A factory, a warehouse full of finished bikes, and a headcount built to serve a subscriber base do not shrink automatically when demand normalizes. They are commitments made in one climate that have to be paid for in whatever climate arrives next.

That is the trap: the decision to build capacity is made at the exact moment the data looks best, which is also the exact moment it is least trustworthy.

What Peloton Actually Built

Peloton’s revenue went from $1.83 billion in fiscal 2020 to $4.02 billion in fiscal year 2021, the twelve months ending June 2021, a 120 percent jump, according to the company’s own investor relations reporting.

Connected fitness subscriptions, the recurring number that mattered most to the market, grew 114 percent year over year to roughly 2.33 million by the fourth quarter of calendar 2021. The stock, which had listed at $29 in September 2019, closed at an all-time high near $171 on January 13, 2021.

Management read this as proof of a durable position and moved to lock it in. In May 2021 Peloton broke ground on Peloton Output Park, a planned $400 million, roughly 1.9 million square foot manufacturing plant in Troy Township, Ohio, expected to employ around 2,000 people and start production in 2023.

Seven months earlier it had agreed to buy Precor specifically to add owned manufacturing capacity and shorten delivery windows, a deal that closed in early 2021. Peloton was building the picks-and-shovels business to service a gold rush it assumed would keep going.

It did not. By early 2022 gyms had reopened, competitors including Tonal and Lululemon’s Mirror had crowded into the same category, and hardware orders slowed sharply.

Peloton disclosed roughly $1.1 billion of inventory sitting unsold in fiscal 2022, and by its fiscal third quarter that year it had booked around $182 million in inventory impairment charges, with cumulative reserves reaching about $255 million by September 2022 across excess accessories, returned equipment and unusable components, per the company’s own SEC filings.

On February 8, 2022, founder and CEO John Foley stepped down, replaced by former Spotify and Netflix finance chief Barry McCarthy, and Peloton cut about 2,800 jobs the same day. The Ohio factory, never having produced a single bike, was cancelled that day too. The remaining parcel of land sold for $4.2 million in September 2024, a rounding error against the $400 million plan it replaced.

Peloton’s headcount had peaked above 6,700 at the end of June 2021. By October 2022, after a second layoff round, it was down to roughly 3,825.

In July 2022 the company announced it would exit owned manufacturing entirely, handing production to Taiwan’s Rexon Industrial and shutting its own Tonic Fitness Technology facility, an admission that the entire capital program behind Precor and Output Park had been a bet the company was now unwinding.

Why it broke

The mechanism is not that Peloton grew too fast, or that its product was bad, or that its brand curdled.

The mechanism is that Peloton’s team confused the size of a temporary market with the durability of a competitive position, and financed permanent fixed costs, a factory, a $420 million acquisition, a headcount nearly triple pre-pandemic levels, against a number that was mostly weather.

The subscription retention underneath it was real, but it was a fraction of the size of the hardware boom that had funded the capital spending. When the pandemic-shaped demand normalized, the fixed costs did not normalize with it. Peloton was left paying pandemic-era rent on a post-pandemic business.

What Peloton Never Broke

It is worth being precise about what did not fail, because the lazy version of this story implies Peloton was bad at everything. It was not.

Even during the worst of the inventory crisis, subscriber churn stayed low by industry standards, monthly workout volume per subscriber held up, and the content and instructor programming, the actual product experience, kept its Net Promoter Score well ahead of most consumer hardware brands.

Peloton’s connected fitness subscriptions did not collapse when hardware sales did. They kept climbing through 2022 and peaked above 3.1 million in the March 2023 quarter, nearly a year after the CEO change and the factory cancellation. The company had built one genuinely durable thing. It had simply built far more fixed cost on top of it than the durable thing could carry alone.

How to Tell a Spike From a Moat

1. Separate the number that grew from the number that stayed

Before you commit capital, isolate the metric that reflects an external shock, units sold, foot traffic, signups in a single window, from the metric that reflects behavior after the shock ends, repeat purchase, renewal, usage twelve months later. Peloton had both numbers in front of it in 2020 and built its factory against the wrong one.

2. Ask what happens if the condition reverses tomorrow

Run the test explicitly, in a meeting, with a written answer. If gyms reopened tomorrow, if the shortage ended tomorrow, if the competitor exited tomorrow, does the demand curve hold. If the honest answer is “no, but it’s been holding for eighteen months,” that is not evidence of a moat. It is evidence the condition has not reversed yet.

3. Match the duration of your cost to the duration of your evidence

A short-cycle cost, inventory, temp staff, a leased line, can be built against a spike because it can be unwound at the same speed the spike fades. A long-cycle cost, a factory, an acquisition, a permanent headcount tier, needs multi-year evidence, not eighteen months of one unusual year.

The unit economics work in scaling a business applies here directly: cheaper growth only stays cheap if the fixed costs behind it are matched to demand you can actually count on, not demand you are hoping repeats.

4. Price and build for the subscription you can prove, not the one you are projecting

Peloton’s real asset, subscriber retention, was smaller and slower than the hardware wave sitting on top of it. Build your fixed cost base against the smaller, provable number. If the bigger number turns out to be real too, you can add capacity later, on a shorter timeline than an Ohio factory ever offered you.

5. Put a name on the decision-maker who owns the reversal case

Every capital commitment made during a demand spike should have one person whose job is explicitly to argue the demand does not last, and that argument should be written down before the money moves, not reconstructed afterward to explain what went wrong.

Peloton’s board approved the Precor deal and the Ohio plant within months of each other, at the top of the curve, without a documented answer to what happens if the curve is temporary.

A Graph Shaped by a Closed Gym Is Not a Business Model

Peloton did not lose because a competitor out-executed it. It lost because it looked at a graph shaped by a closed gym and called it a business model. The subscription retention it pointed to as its moat was real, but it was never large enough to carry the fixed costs the company built on top of a pandemic-shaped assumption.

The lesson generalizes past exercise bikes.

Any founder or operator staring at a number that has only existed for eighteen months, sitting on top of an external shock nobody controls, should ask whether they are building a moat or building a factory for the weather. Peloton built the factory. It sold the land four years later for one percent of what the plan cost.

The contrast worth holding onto is AG1, a company that also sells a recurring health habit through a single hardware-adjacent product line. AG1 built its subscription business by owning its own demand generation from the start, so the subscription was the whole business, not a costume worn by a hardware boom. Peloton had the costume. AG1 never needed one.