- US connected TV (CTV) hits about $38 billion in 2026, up 14.5%, and passed primetime linear on upfronts. [1]
- It's 43.8% of TV viewing but only 7.7% of ad spend. That gap is the whole story. [2]
- 84% of CTV is programmatic across a fragmented field. Frequency and measurement are the real work. [2]
- Netflix, Prime Video, and other ad tiers behave like walled gardens. Plan them as such. [3]
- Stop splitting "TV" from "digital video." Plan reach across screens as one system.
Connected TV is where the audience already is. It is not, yet, where the money is. And the distance between those two facts is the most straightforward reach opportunity left in video advertising.
US connected-TV ad spend reaches about $38 billion in 2026, up 14.5%, and it has already edged past primetime linear on upfront commitments. [1]
US CTV ad spend: eMarketer forecast via StackAdapt [1]. The 2028 figure is corroborated by MNTN Research [5]. 2028-29 are projections.
The headline growth isn't the interesting number, though. This one is.
43.8% of US TV usage. That's 20.2% of all time spent with media. The eyeballs are already here.
7.7% of total US ad spend. The budgets are still parked on the screen people left.
That mismatch (43.8% of viewing against 7.7% of spend) is underpriced reach. [2] It won't last. Spend is climbing double digits and CTV upfront commitments ($17.73 billion) have already passed primetime linear ($16.98 billion). [2] But for now there is genuine arbitrage for brands willing to do the hard part.
The hard part is fragmentation, and it's not a detail
The reason the money lags the attention isn't inertia. It's that buying CTV well is operationally difficult.
About 84% of CTV spend (roughly $32 billion of the $38 billion total) is transacted programmatically, spread across Amazon, Netflix, Roku, YouTube, and Disney, and only three companies are expected to hold more than 10% of CTV ad sales in 2026. [2][5] Buy each of those platforms in isolation and no single one's tools can give you deduplicated reach or a real frequency cap. You'll pay to hit the same household eight times while missing the one next door. Cross-publisher identity, log-level data, and independent measurement aren't sophistication here. They're the minimum not to waste the budget.
Together, CTV and digital video now command about 23% of US ad spend (the largest share of any channel), and 55% of marketers expect attention to become the standard way CTV is measured and bought. [4][6]
The walled gardens are turning CTV into retail media.
Netflix ad revenue passed $1.5 billion in 2025 and is set to roughly double in 2026. [3] Prime Video, Netflix, and their peers offer strong first-party data and weak interoperability, exactly the retail-media pattern. Plan them as distinct surfaces with their own frequency and attribution, not as interchangeable programmatic inventory you can dedupe away.
How to actually play it
So play it deliberately. Make CTV the default TV line and treat linear as a declining reach complement, not the base. Claim the arbitrage where it's strongest (reach extension over linear and social), but solve frequency first: insist on cross-publisher identity and a household cap before you scale spend. Plan the walled gardens, Netflix and Prime Video, as their own surfaces with bespoke measurement, the way you'd treat retail media. And budget TV, CTV, and online video as one reach-and-frequency system, pushing vendors toward attention and cross-screen outcomes rather than platform-siloed impressions.
The CTV story is usually told as growth: bigger every year, past linear soon. The more useful version is the gap: an audience that has already moved, and a budget that hasn't. That gap is a discount on reach, and discounts on reach don't stay open. The brands that win the next two years are the ones that fix frequency and measurement now, while the inventory is still cheaper than the attention it carries.
Sources
- StackAdapt · Connected TV statistics (eMarketer 2026 forecast)
- Digital Applied · Connected TV advertising 2026: performance marketer's guideNielsen usage, programmatic share, upfront figures
- Adwave · CTV advertising 2026 update
- IAB via MediaPost · IAB 2026 ad spend forecast: CTV, social, search, linear shares
- MNTN Research · CTV ad spend forecasts to 2028
- Teads · Connected TV advertising 2026 trends
Frequently asked questions
How big is CTV advertising in 2026?
eMarketer forecasts about $37.95 billion in US connected-TV ad spend for 2026, up 14.5% from $33.14 billion in 2025. On upfront commitments, CTV ($17.73 billion) has already edged past primetime linear ($16.98 billion), and CTV is broadly expected to surpass linear TV ad spend overall around 2028.
What is the 'arbitrage' in CTV?
Attention has moved faster than money. CTV accounts for 43.8% of total US TV usage and 20.2% of time spent with all media, but only 7.7% of total ad spend. That gap between share of viewing and share of budget is underpriced reach. The opportunity is claiming it before the spend catches up and prices normalize.
Why is fragmentation the main operational problem?
About 84% of CTV spend is transacted programmatically across Amazon, Netflix, Roku, YouTube, and Disney, and only three companies are expected to hold more than 10% of CTV ad sales in 2026. No single platform's tools give you deduplicated reach or frequency across that set, so cross-publisher identity, log-level data, and independent measurement stop being nice-to-haves and become table stakes.
Should I treat Netflix and Prime Video like retail media?
Increasingly, yes. Their ad tiers are scaling fast (Netflix ad revenue passed $1.5 billion in 2025 and is expected to roughly double in 2026), and they behave like walled gardens: strong first-party data, limited interoperability, bespoke measurement. Plan them as their own surfaces with their own frequency and attribution, not as interchangeable programmatic inventory.
How should I structure TV and video budgets now?
As one system, not 'TV versus digital video.' Digital video plus CTV now hold the largest share of US ad spend at about 23%, ahead of social (18.4%) and search (16.2%). Plan reach and frequency across linear, CTV, and online video together, and push measurement vendors toward attention and cross-screen outcomes rather than platform-siloed impressions.
When will CTV overtake linear TV ad spend?
On upfront commitments it already has, with CTV at $17.73 billion ahead of primetime linear at $16.98 billion. Across TV ad spend overall, CTV is broadly expected to surpass linear around 2028. So the crossover is not a single moment but a staggered one that has already begun on the forward commitments.
How will CTV be measured going forward?
The industry is moving toward attention and cross-screen outcomes rather than platform-siloed impressions, with 55% of marketers expecting attention to become the standard way CTV is measured and bought. Because no single platform gives you deduplicated reach across the field, cross-publisher identity, log-level data, and independent measurement become the baseline. Push your measurement vendors toward outcomes across linear, CTV, and online video together.
How long will the CTV arbitrage last?
Not long. Spend is already climbing double digits, up 14.5% in 2026, and CTV upfront commitments have passed primetime linear, so the gap between 43.8% of viewing and 7.7% of spend is closing. The window favors brands that fix frequency and measurement now, while the inventory is still cheaper than the attention it carries.
Mikołaj Salecki
Writes about media, tech, and AI business for people who actually run digital. Former agency lead. Skeptic of frameworks that read better than they perform.
Get the next issue by email.
One letter, once a week. Sharp coverage of media, tech, and AI business. No filler.
