Peloton Spent $47M Buying a Bike Factory in 2019 — and Walked Away From Owning Factories in 2022

In October 2019, weeks after its IPO, Peloton bought a factory. Not a stake, not a supply contract — the whole thing: Tonic Fitness Technology in Taiwan, one of the plants that stamped out the frame under its connected bikes, for about $47 million.¹⁸ It looked like a company at the peak of its confidence reaching down the value chain to own the metal, not just the software. Two years and nine months later, Peloton announced it was getting out of owned manufacturing entirely.⁴ The factory it had bought to control its own destiny went quiet, and Peloton handed the work to a supplier it had been using all along.

The story that stuck is neat and slightly wrong: Peloton bought its sole bike supplier and then shut it down. Tonic was never the sole supplier, the 2022 wind-down was framed as a suspension rather than a shutdown, and the Tonic factory was the smallest piece of a much larger, much costlier retreat. The real story isn't a supplier being killed. It's a vertical-integration bet placed at exactly the wrong moment in the demand cycle.

Tonic was one supplier among several — not the only one: Peloton's own filings name the partners it kept, even as it bought one outright

Read Peloton's own words and the 'sole supplier' framing falls apart. The November 2019 shareholder letter calls Tonic 'one of our long-time' bike manufacturing partners — plural, deliberately.² A year later, announcing its $420 million deal for Precor, Peloton described the new U.S. plants joining 'its existing manufacturing network with its third party manufacturers and the Tonic facilities based in Taiwan.'³ Tonic sat alongside the network, not on top of it. The other partner that mattered was Rexon Industrial, which Peloton's own supply-chain chief later confirmed had been building its stationary bikes and Tread all along.⁵ So the 2022 move wasn't Peloton losing its only factory. It was Peloton deciding that owning one factory made no sense when a contract manufacturer could do the same work without tying up capital — and it already had that manufacturer on speed dial.

The $47 million wasn't the bet — the $420 million was: Tonic was a toe in the water; Precor and a planned U.S. plant were the plunge

Focus on Tonic and you miss the actual gamble. Tonic was small — roughly $47 million, a bolt-on to control a slice of production during a demand surge.¹⁸ The real vertical-integration bet came later and cost an order of magnitude more. In December 2020, with pandemic demand at full boil, Peloton agreed to buy Precor for $420 million to get U.S. manufacturing capacity.³ It then began building its first-ever U.S.-owned factory in Wood County, Ohio. This was the strategy in full: own the plants, control the output, never be caught short of bikes again. It made sense only if the line up and to the right kept going. It didn't.

When demand cracked, all three legs buckled in the same window. In March 2022 Peloton canceled the Ohio factory and cut around 2,800 jobs.⁷ In July it announced it was exiting owned manufacturing outright, moving Bike and Tread production to Rexon and suspending Tonic through the rest of the year.⁴ And the biggest asset proved the hardest to shed: the attempt to sell Precor collapsed after the buyer's offer 'dramatically dropped,' so Peloton kept running it as a standalone subsidiary instead.⁶ The factories you buy at the top are the ones nobody wants to buy at the bottom.

Why owning the factory turned from asset to anchor: vertical integration converts a variable cost into a fixed one — brilliant on the way up, brutal on the way down

Here is the mechanism, worked down. Contract manufacturing is a variable cost: you pay per bike, and if orders halve, your bill roughly halves. Owning the plant converts that into a fixed cost: the building, the tooling, the workforce, and the obligation exist whether or not a single bike ships. During the pandemic surge, that trade looked like genius — owning capacity meant you controlled it, and control was the scarce thing when everyone wanted a bike. But the same fixed cost that guarantees supply on the way up becomes dead weight on the way down. When Peloton's demand reversed, the owned plants didn't flex. They just sat there, absorbing cash. Rexon, by contrast, could scale down with the orders because it wasn't Peloton's problem to keep busy.⁵ That is the whole difference between renting capacity and owning it — and it only reveals itself when the direction of demand changes.

Wasn't owning capacity the right call during the shortage?: in 2020 the logic was real; the error was timing, not intent

The fair objection is that in 2020, integrating was defensible. Peloton genuinely could not make bikes fast enough; waitlists stretched for weeks, and a competitor's supply was a customer Peloton lost. In that world, owning capacity was a rational hedge against a supply constraint that was actively costing sales. That's true — and it's exactly why the mistake is instructive rather than stupid. The error wasn't the intent; it was treating a demand spike as a demand level. Peloton committed hundreds of millions to fixed capacity as if the surge were the new baseline, then discovered the baseline was much lower. And notice what it kept: not the owned plants, but the partner relationship with Rexon, the flexible arrangement it had never actually abandoned.⁴⁵ The company ended up 'going back to nothing but partnered manufacturing' — right where it started, minus the money spent proving it should never have left.⁵

The tidy legend says Peloton killed the supplier it depended on. What actually happened is stranger and more useful: it spent real money to own its factories at the exact moment owning them was worth the most, then discovered the price of that ownership at the exact moment it was worth the least. Tonic was the cheap lesson. Precor was the expensive one — the plant nobody would buy back. Vertical integration doesn't reward control; it rewards being right about your own future. Peloton was right about 2020 and wrong about 2022, and it paid the difference in buildings it couldn't sell.