Roku Loses Money on Every Box It Sells. That's the Plan, and It's Working.
In 2024 Roku's hardware lost $80M — devices cost more to make than they sold for. That's not a stumble; it's the point. The box is bait. The $1.9B in platform profit at a 53.5% margin is the catch.
Comes with a free Profit-Engine Map template — plus a worked example for Roku.
Walk into a store this holiday season and you can buy a Roku streaming stick for the price of a couple of movie tickets. Take it home, plug it into your TV, and a company you'll forget about by Tuesday quietly takes up residence on your home screen — and starts working. Here is the part that should stop you: Roku made that stick at a loss. In fiscal 2024, Roku's hardware brought in $590.1 million and cost $670.4 million to build and ship — an $80.3 million gross loss before a single salary was paid.1 The company is not bad at making boxes. It is selling them below cost on purpose. The box was never the product. You are.
The official story is that Roku is a hardware company that got into advertising. The truth is the reverse, and it was never hidden. Roku built an ad business and used cheap hardware to deliver an audience to it. There was no pivot. The subsidy on the stick is the down payment on a much larger transaction you don't see — the one that happens on your screen every day after.
“Our business model is focused on growing active accounts, and then monetizing those active accounts through our platform business. The way we grow active accounts is we sell streaming players, we license to TV companies and we license to operators.”6
The box is the cost of acquiring you, not the thing being sold
Almost every consumer-electronics company in the world treats the device as the prize: design it, mark it up, book the margin. Roku does the opposite. It treats the device as a marketing expense that happens to look like a product. The cheaper the stick, the more homes it lands in, and every home is a permanent storefront Roku now controls — the home screen, the ad inventory on it, and a cut of every subscription that runs through it. The $80 million it 'lost' on hardware in 2024 is better understood as customer-acquisition cost, paid once, against a customer who then pays rent indefinitely. Most companies pay to acquire customers and pray they monetize. Roku built a business where the acquisition channel ships in a retail box with a barcode.
And the platform it acquires you for is genuinely large. In 2024, platform revenue hit $3.5 billion, up 18% year over year, at a 53.5% gross margin — roughly $1.9 billion in gross profit.13 That figure dwarfs the hardware loss by more than twenty to one. Notice, too, that the platform is not just ads. It has two roughly co-equal legs: video advertising on one side, and distribution on the other — the revenue-share Roku takes from streaming services, premium subscriptions sold through its store, and the cash content owners pay for those branded buttons on the remote. In full-year 2024, distribution actually grew faster than advertising.3 The remote in your hand is, quietly, a billboard.
| Devices (hardware) | Platform (ads + distribution) | |
|---|---|---|
| Revenue | $590.1M | $3,522.8M |
| Cost of revenue | $670.4M | $1,636.8M |
| Gross result | −$80.3M loss | ~$1,885.9M profit |
| Gross margin | Negative | 53.5% |
| Role in the model | Acquire the customer | Monetize the customer |
The right-hand subtraction is the $80.3M hardware loss in 2024 — the toll Roku pays to plant its OS in a home.1 The left-hand term compounds: once the stick is plugged in, every ad impression and every subscription routed through the home screen accrues at a 53.5% margin, with almost no incremental cost. Sell the box at a loss once; collect on the screen forever. That asymmetry is the whole machine.3
The model finally stopped being a leap of faith
For years the bet was theoretical: keep losing money on boxes, keep scaling the platform, and eventually the math turns. In 2024 and 2025 it did. Q3 2024 was Roku's first-ever quarter above $1 billion in total revenue.8 Q4 2024 was its first quarter above $1 billion in platform revenue alone, up 25%.3 By full-year 2025, platform revenue reached $4.145 billion and the company posted its first positive net income, with record free cash flow and Adjusted EBITDA margin up 255 basis points.4 Tellingly, Roku also stopped reporting household counts and ARPU after 2024 — a signal that it now wants investors watching the cash engine, not the audience-growth story it used to lean on.8 The loss leader has done its job. The flywheel spins on its own.
Isn't it dangerous to bleed money on hardware forever?
The fair objection is blunt: a business that structurally loses money on its physical product is one supply-chain shock away from trouble, and Roku is still doing it — devices posted a $22.9 million gross loss as recently as Q3 2025.7 But that misreads the loss. The subsidy isn't a leak; it's a deliberate, sized investment in installed base, and it has stayed small relative to the platform it feeds. The deeper risk is more interesting, and Roku can't fully control it. The entire model rests on Roku owning the home-screen operating system — the layer that decides what ads show, which apps surface, and who gets a cut. That moat holds as long as TV makers are happy to license Roku's OS. It erodes the day they don't. Samsung runs its own Tizen OS; LG runs webOS. A world where every major TV manufacturer internalizes its own software is a world where Roku's cheap hardware buys a shrinking share of screens, and a loss leader that no longer leads to the platform is just a loss. The model is proven. It is not permanent.
A loss leader only works if it leads somewhere you own. Roku eats $80M on hardware because that hardware deposits its OS — the home screen, the ad slots, the distribution toll — into a home it then monetizes at 53.5% for years. The discipline is twofold. First, make sure the cheap thing acquires a customer you can actually charge again, not a one-time buyer who walks away. Second, make sure you control the layer where the money is made: the moment a supplier or partner can internalize that layer themselves, your subsidy stops buying a moat and starts buying nothing. Cheap hardware is only strategic when it's the gateway to a fortress you hold the keys to.
Roku figured out something the rest of the electronics aisle missed: in streaming, the screen is worth more than the device that lights it up. So it gave the device away at a loss, kept the screen, and built a billion-dollar toll on everything that crosses it. The $80 million it spends losing on hardware each year isn't a weakness in the model — it's the model, paying for the only thing that matters: a permanent seat in the living room. The genius was deciding the box was never the business. The danger is that one day, someone else owns the screen — and a loss leader with nowhere to lead is just a company losing money.
Profit-Engine Map
A one-page map that pulls a business apart into the hook that gets the customer in the door and the engine that quietly earns the margin. Use it to see where the real profit lives, how the two halves are wired together, and what breaks if the link is cut. Blank to dissect your own P&L; filled as the worked example of a business whose advertised product is not where it makes its money.
The worked example unlocks with a subscription. See plans →
Sources
Where this comes from — the filings, records, and reporting behind it.
- 1FY2024: Devices revenue $590.1M, devices cost of revenue $670.4M, devices gross loss -$80.3M; Platform revenue $3,522.8M, platform cost $1,636.8M, platform gross profit ~$1,885.9M (~53.5% margin); total net revenue $4,112.9M.
- 2Q1 2024: Devices revenue $126.5M vs. cost $132.6M (gross loss -$6.1M); Platform revenue $755M vs. cost $360.6M (gross profit $394.4M, ~52% margin).
- 3FY2024: Platform revenue $3.5B up 18% YoY; Q4 2024 first quarter exceeding $1B in platform revenue (up 25% YoY); platform gross margin 54.1% in Q4 and 53.5% for full year 2024; devices revenue guidance for 2025 of $660M (up 12% YoY).
- 4FY2025: Platform revenue $4.145B up 18% YoY with gross margin 52.0%; Roku achieved positive net income in 2025, expanded Adjusted EBITDA margin by 255 basis points, and reported record free cash flow.
- 5Roku's 2017 S-1 prospectus established the foundational model: grow active accounts through hardware and licensing, then monetize through the platform. Anthony Wood stated: 'The only reason we sell hardware is to acquire customers' (attributed, per contemporaneous Nasdaq/Variety reporting of the IPO).
- 6Anthony Wood stated at IPO: 'Our business model is focused on growing active accounts, and then monetizing those active accounts through our platform business. The way we grow active accounts is we sell streaming players, we license to TV companies and we license to operators.'
- 7Q3 2025: Platform gross profit $547.8M; devices gross loss -$22.9M, reflecting ongoing structural negative margins in hardware even as platform scales.
- 8In Q3 2024, Roku posted its first-ever quarter of more than $1B in total net revenue; the company announced it would stop reporting streaming household counts and ARPU starting Q1 2025, shifting investor focus to platform revenue and profitability.