Roku, Inc. ROKU

Key data

The business

Roku makes two things that work as one. It sells cheap streaming players and Roku-branded TVs, close to cost, to get its operating system into as many homes as possible. Then it makes its money on the Platform side: the ads on the home screen, the branded buttons on the remote, the cut it takes when you subscribe to a streaming service through Roku, and its own free channel. The hardware is the cost of entry. The home screen is the toll booth every stream passes through. Advertising is the engine, and the last 10-Q flags one advertising customer, called Customer J, large enough to name as a concentration. The Frndly TV service it bought in 2025 adds subscription dollars. International is still under 10% of revenue, so this is, for now, a bet on the American living room. The moat is the installed base and the screen real estate no rival controls.

The numbers

The five quarters below show the turn.

The story is the last two columns. A year ago Roku made $10.5 million in a quarter. This June it made $164 million, and diluted earnings of $1.08 came in at nearly double the $0.61 analysts expected, the fourth straight quarter well ahead of them. Revenue grew 22% year over year, steady, not spectacular. What changed is what fell to the bottom.

Revenue compounded about 14% a year from 2021 to 2025, a straight line through the ad recession that dragged net income to a $710 million loss in 2023. The number that matters is operating income: minus $5.6 million for all of 2025, then positive $198 million in the first half of 2026 alone. That is operating leverage, revenue climbing while cost holds, and it is why the profit inflected so hard.

Gross margin on the whole business went from 44.8% to 49.7% in a year while operating income swung from a $23 million loss to a $146 million profit. With ARPU no longer disclosed, gross margin remains the cleanest gauge of how well Roku monetizes each home. What the market may not fully credit is that the leverage looks structural, not a one-off ad bump: capex is almost nothing, about 0.2% of revenue, and the cost base barely moved. The print that settles it is operating margin staying double digits through the back half.

Management

Anthony Wood founded Roku and still controls it through Class B super-voting stock, 16.4 million shares against 132 million public Class A. Insiders have been sellers, not buyers: zero purchases and 33 sales worth $14.4 million over the past year, the largest from officers Fuchsberg and Collier, all after the stock ran from $79 to $160. Plan status is not disclosed in the filings I read, so read those as officers taking money off a tripled stock, not a signal either way. More telling is the capital allocation. Roku bought back $150 million of its own stock in 2025 and over $100 million more in the first half of 2026, its first buybacks ever, at prices below today's. Stock-based compensation runs about 4.6% of revenue, moderate for the industry.

How it fails or surprises you

One advertiser leaves, or the ad market cools. The 10-Q names a single advertising customer large enough to be a concentration risk. Platform advertising drove the margin gains, so if that customer pulls back or CTV ad budgets soften, the gross margin and the earnings retrace as fast as they rose. Watch gross margin slipping back below 46% or revenue growth falling into the low teens.

The leverage proves cyclical, not structural. Operating income went from a $5.6 million loss in 2025 to $198 million in six months. The honest doubt: some of that is a strong ad cycle, not permanent discipline. If sales, content, and stock-comp costs re-inflate as growth normalizes, the margin gains fade. The print that would prove the read wrong is operating margin falling back to low single digits in the second half.

Ad load, subscriptions, and international all turn on at once (right tail). The market prices about 14% revenue growth out to 2028. If Roku lifts ad load on its household lead, compounds the Frndly subscription base, and finally turns on international beyond today's sub-10%, growth reaccelerates toward 20% with margins still climbing. Two straight quarters of 20%-plus Platform growth would reveal it first.

Closing thoughts

This is a question a print will answer, not a permanent unknown. Whether the 2026 profit turn is durable operating leverage or a well-timed ad cycle gets settled by the next two operating-margin reports, nothing sooner. The fatter tail is the upside: capex near zero and a barely-moving cost base say the leverage is largely real, which the 66-times trailing multiple half-hides because last year's earnings were still depressed. What is at risk if Customer J or the ad market breaks is a hard multiple compression from a rich starting point. What it is worth if the leverage holds is a re-rating onto a fast-compounding free-cash-flow base that is only now becoming visible.

The bet is still that Roku keeps putting its operating system inside more American TVs and streaming boxes, and keeps turning those living rooms into advertising dollars on its home screen, with the money it earns from ads and subscriptions growing faster than what it spends. What breaks it is ad demand softening or the cost base re-inflating. The one pair that tells you first is gross margin and revenue growth, read together each quarter. If gross margin holds near 50% while revenue keeps growing 20%, the turn was real.