Netflix Q3 2026 Earnings: The Ad Forecast Cut Is About Who Owns the Pipe

Netflix's third quarter was fine on the main lines. Revenue grew 12% to $12.9 billion, operating margin was 34.6%, and full-year guidance was held at $51.1 billion. The change that mattered for this piece was one sentence in the letter: Netflix now expects about $2.7 billion of ad revenue in 2026, down from about $3 billion, "due to slightly slower than expected rollout of some ad product capabilities." On the call co-CEO Greg Peters said Netflix was "a bit behind our aggressive target for the year" and had "seen some delays in the rollout and partner integrations for some of those new ads products," while "general demand remains strong."

Ads are about 5% of Netflix's revenue, so $300 million is small on the year. I think the reason given for it is the more useful part. Netflix didn't say advertisers stopped buying. It said the tools and connections that let them buy weren't ready on time, and Netflix sells some of its inventory through buying tools that belong to other companies. That's worth reading against what happened to the rest of the ad market this year.

The sellers that grew are the two that can show an advertiser what the ad led to, a purchase on Amazon or a search on Google. Spotify and The Trade Desk can't, and both grew about 3%. Spotify is the useful case here, because it owns its buying tool, an ad exchange and Ads Manager, and that didn't save it; its own update blames softer pricing. So I think the data is the part that counts. My Trade Desk stock page records the fee side: Amazon's DSP charges close to no fee and Google capped its fees, and advertisers have been moving spend toward the tools that come with the data. Netflix is growing far faster than either of the slow pair, from a small base. Netflix sells programmatically through Amazon's DSP, Google's DV360, The Trade Desk and Yahoo as well as its own suite. It didn't say which partner integrations ran late, so this is my inference, but I think some of the slip sits in exactly those connections.

Netflix and Amazon are both rated strong on proprietary data, but the stock pages earn those ratings on different data. Netflix's rating rests on viewing patterns that tell it what to commission, which is a real edge in programming. Amazon's rests on purchase intent, which is the thing an advertiser pays to reach and to measure. Amazon is also rated strong on transaction embedding, and Netflix intact, with its note crediting the ad tier as "a second transaction layer." The Trade Desk is weakened on proprietary data for exactly the reason in the table: it has no purchase data and competes against two sellers that do. My reading, which the framework doesn't score separately, is that Netflix's ad business today borrows the transaction layer from whoever's buying tool the advertiser uses.

A TV ad used to be sold on reach. A streaming ad is increasingly sold on outcomes, and the outcome has to be measured somewhere. Amazon's DSP pitch for Netflix inventory pairs it with Amazon's first-party insights and clean-room measurement. That puts Amazon's data next to the impression when a buyer uses its tool, and Netflix doesn't have purchase data of its own to set against it. Over time that shifts pricing power toward whoever holds the proof.

• Peters said Netflix wants to "increase the availability of our first-party measurement" and to "advance our full-funnel optimization solutions." Both are about proving outcomes, and that's where Amazon and Google start with purchase and search data that Netflix doesn't have.
• Netflix said this year's delay was in rollout and "partner integrations." The Trade Desk's Q2 release lists Netflix joining its Sellers and Publishers 500+ program this year. Each new buying route adds demand, and each one runs on the partner's schedule as well as Netflix's.
• When the sellers with purchase data cut their fees, a seller without it loses share. The Trade Desk's growth fell to 3% in Q2 and it guided Q3 below last year's $739 million. Netflix isn't a DSP, but it sells into the same buyers, who increasingly pick the tool that comes with the data.

The strongest objection is that I'm treating a timing slip as a structural problem. Netflix's ad business is still growing over 80% this year, and Peters said demand is strong and advertisers keep broadening. Peters called Netflix's own ad tech build "the fastest build-out of a comparable ad tech stack in our industry," and it chose to take "a modest short-term hit for a better long-term strategic position" by holding out for the integration model it wants and to "retain our first-party data advantage." On that reading, the $300 million is the cost of not handing its data to partners on their terms, which is the opposite of the dependence I'm describing.

The second part of the objection is about what Netflix sells. The most valuable TV ads have always been sold on attention and reach, not on a click to a checkout, and Netflix has the largest premium audience in streaming. Live sports and events are growing, and Ted Sarandos said live events "add fuel to our ad revenues." Netflix is also taking the ad tier to 15 new countries in 2027, starting with twelve in March, nine of them in EMEA, which adds inventory on its own terms. A brand advertiser buying a Christmas NFL game is paying for the audience, and Netflix doesn't need a purchase graph to sell that. On this view, Amazon and Google win the performance budget and Netflix wins the TV budget, and both can grow.

I think that's right about the TV budget, and it's why I don't expect Netflix's ad line to stall. Where I disagree is on what it does to the moat. A brand budget sold on reach is priced against every other premium screen, including YouTube on the living-room TV and Prime Video, both of which can also show measured outcomes. Netflix's own priorities on the call, first-party measurement and full-funnel optimization, are an admission that the TV budget is moving toward proof too. Building that proof without purchase data is possible, but it's slower, and this quarter showed it running behind Netflix's own plan. Until Netflix can sell outcomes on its own data, I'll treat its ads as a growth driver and won't credit them as a separate moat.

Netflix will give 2027 guidance in January, and its CFO said ads remain one of three revenue levers. The falsifiable claim above is the plain test, checked against the Q4 2027 letter in January 2028: 2027 ad revenue of about $5 billion or more from a $2.7 billion base, near this year's pace, would mean Netflix can grow ads at walled-garden speed while part of its demand comes through other companies' tools. Growth of 50% to 80% wouldn't break it. A January 2027 guide at that level would be an early warning, not the test itself. Before then, Amazon, Alphabet and The Trade Desk report Q3 in the next three weeks. If Amazon and YouTube ads keep growing double digits while The Trade Desk shrinks, the split in the table is holding. If The Trade Desk recovers, the independent buying route is healthier than I think, and Netflix's partner integrations matter less.

This also bears on my September piece on Netflix and the content-scale floor, which named advertising above $6 billion for two years as the print that would force a review of its transaction embedding. This quarter pushes that line further out than the July forecast implied, and I've added a revision to that piece.

I'd read the other subscription businesses adding ads the same way. A streaming or audio service with a large audience can grow ad revenue quickly at first, because the inventory is new. How far it goes depends on whether it also owns where the advertiser buys and how the result gets measured, and right now Amazon and Google own most of both.