We built beautiful engines for content delivery. But forgot to fuel them.
Streaming promised everything: personalization, control, infinite choice. A direct line to what you love, when you want it, on the device you happen to be holding. It delivers on cable’s old promise of the “4 Any’s”: Any Content, Anywhere, Anytime, Any Device — in a way the cable industry never saw coming.
But here in 2025, as the streaming giants get older and their balance sheets more demanding, a fundamental tension is surfacing: the platforms are fully grown — but the business model is still figuring itself out. The infrastructure is sophisticated. The UX is slick. The content recommendation is decent, if a little same-y. Streaming isn’t struggling with how to deliver content and it’s struggling with how to pay for it. And that, in turn, is driving the churn.
Note: These AI ideas aren’t gospel — just possibilities. They are sketches, not blueprints. If you’re already building in this space (or know someone who is), drop me a line. Always happy to swap notes, remix ideas, or connect dots we haven’t seen yet.
A Maturing Platform, A Stalled Model
Most of the major players — Netflix, Disney+, Hulu, Max — have passed the early-stage chaos. They’ve gone global. Super Bowl. They’ve launched ad tiers. They’ve consolidated libraries and restructured programming. What they haven’t done is build a monetization system that feels as smart as their delivery pipeline.
The numbers tell the story:
U.S. viewers now spend $61/month on streaming, a 27% increase since 2023 — and suddenly a lot closer to the old cable bundle [AdImpact, 2025].
Over half of Gen Z and millennial users canceled at least one streaming service in the past six months [NewscastStudio, March 2025].
· Platforms are pushing ad-supported tiers harder than ever, yet CPMs remain fragmented, and brands remain skeptical. FAST has surged to fill the weekday lean-back gap — but whether the revenues are truly following is still an open question. [Deadline, 2025]
We’ve hit a ceiling. Not because people don’t want content — but because the economics of how we surface and support that content haven’t caught up.
Five Business Gaps Holding Back Streaming Monetization
Let’s be blunt; the ad-supported side of streaming still feels like an awkward bolt-on. There are five big friction points slowing things down:
Audience Understanding Is Shallow
We’ve built personalization for viewers — but not for advertisers. Identity is fractured across devices. Most identity companies don’t resolve identity outside the advertising bid stream. Churn makes it worse. And the privacy backlash means traditional tracking is vanishing.
Creative Is Expensive and Generic
Ads don’t flex to genre, mood, or moment. Most are static, recycled, and forgettable — filler content that adds no value and often breaks the emotional rhythm of a viewer’s session. In a world where everything else is increasingly personalized, ads still feel like they’re built for broadcast. If creative is going to earn its place, it needs to become content — not interruption.
Contextual Matching Is Weak
You could watch a quiet art-house drama and get served an over-the-top gaming ad. Thematic and emotional mismatches aren’t just jarring — they’re inefficient. And then you might see the ad twice, or every ad break.
Measurement is Broken
There’s no standard cross-platform attribution. Advertisers can’t reliably connect ad exposure to outcomes. And optimization, if it happens at all, is slow and reactive, usually after an ad campaign has already ended.
Monetization Forecasting is Blurry
Without strong data and prediction, platforms can’t accurately value their own inventory. They’re flying blind on pricing, pacing, and performance.
Keep reading for paid subscribers
The next section explores how AI isn’t just a shiny new tool. It’s a chance to rewire the economics of streaming from the inside out. From smarter audience modeling to real-time creative swaps, we’ll walk through five ways AI might actually close the monetization gap and what happens if it doesn’t.
How AI Can Close the Gap
This is where AI enters not as a gimmick or a gimmick-chaser but as a toolset that can help streaming’s business model finally grow into its platform.
1. Guesswork to Granularity: AI for Audience Modeling
AI can fill in the gaps where traditional identity fails. Through behavioral clustering, federated learning, and on-device modeling, platforms can build richer (and privacy-safe) profiles.
But granular, real-time data about user behavior and playback experience is essential to train, validate, and continuously improve these models. Platforms like Datazoom are leading the way here by providing detailed, event-level telemetry that connects moment-by-moment viewer behavior with broader personalization and performance efforts.
YouTube’s predictive audience tools already do this at scale.
Amazon’s ad stack leverages purchase behavior to create deeply contextual segments without relying on cookies or logins.
2. One-Size Ads to Generative Creative
Generative AI can create ad variants that match tone, pacing, and genre. It can localize on the fly. And it can adapt based on real-time data — tweaking colors, language, or even the emotional arc.
Netflix’s rumored in-house ad engine is aimed at matching ads to viewing mood.
Synthesia and Runway already offer scalable tools for hyper-adaptable creative.
Generative AI isn’t limited to producing standalone spots. It’s also being used to dynamically insert branded elements directly into shows themselves. Modifying existing content to serve contextual product placement. Platforms like Mirriad are already enabling virtual brand integrations that shift based on geography, user profile, or even viewing device and many companies are exploring this as a feature to other services they provide.
Imagine an ad that feels like it belongs after the scene you just watched — or better yet, within it. The best ad creative can match — or even elevate — the tone of the story it’s placed beside. But it takes effort. AI can’t replace taste, but it can offer more shots on goal.
3. “Ad Spots” to “Ad Scenes”: AI-Powered Metadata
Streaming platforms sit on mountains of data — but most don’t index content beyond title and genre. AI can go deeper: parsing scenes for emotion, tone, character archetypes, and more.
Vionlabs and IRIS.TV offer contextual tagging systems that power smarter ad placement.
Think of it as “programmatic, but for feeling.”
4. Lagging Reports to Real-Time Feedback
AI-enhanced measurement enables mid-flight course corrections. Underperforming creatives can be swapped. Uplift can be tracked through search trends, purchases, or social signals.
Innovid and TVScientific are building attribution models designed for FAST and AVOD worlds.
The holy grail? A feedback loop where the ad engine learns from every frame.
5. From Gut Feel to Precision Planning
Forecasting tools powered by AI can help platforms plan yield more effectively — deciding how much ad inventory to open up, what CPM floors to set, and how to optimize between AVOD and subscription.
Google’s Publisher AI and The Trade Desk’s Koa platform are leading indicators of this shift.
Sidebar from the Near Future: When the Interface Becomes Intelligent
One last thought — what if the next real disruption comes not from a content platform, but from the operating system itself?
Imagine a TVOS platform — not just optimized for streaming, but built from the ground up for advertising. A generative AI layer at the system level could go far beyond recommendations:
An AI-generated, mood-aware lineup built just for you each night.
Ads created in real time to match the tone of the show you’re watching.
A voice prompt that lets you say, “Something slow and thoughtful — maybe sci-fi?” and gets exactly that.
The Trade Desk is already exploring this future, developing an open TV operating system designed specifically for advertisers — potentially rewriting not just the interface, but the economics underneath it.
It’s speculative now. But not far off. Apple, Amazon, and Roku all have the pieces. Whoever gets there first might reshape not just content discovery — but the very economics of television.
The Churn Loop and How AI Might Break It
This isn’t about improving ad performance; it’s about fixing the whole experience. Because churn isn’t a pricing problem, it’s an attention problem. If AI can help make the ad experience more relevant, respectful, and rewarding, viewers might finally feel like they’re getting value again.
We’re already seeing new ad formats test this idea: squeezebacks that keep content visible while ads run in a sidebar, interactive overlays that let users explore without leaving the stream, and dynamic creative that adapts mid-flight. Platforms like Innovid and BrightLine are pushing this forward — hinting at a future where the ad doesn’t interrupt the experience, it participates in it.
Fix the ad experience, and you might fix the churn.
What Happens Next?
We’ve passed the infrastructure race. The pipes are built, and the platforms are mature. But we’re still using broadcast-era thinking to monetize post-linear behavior.
AI offers a chance but not a guarantee to close the loop:
To connect viewers to brands in a way that feels intentional.
To turn ad breaks into editorial choices, not interruptions.
And to finally align the creative, emotional, and economic layers of streaming.
We’re no longer building the system. We’re figuring out what should flow through it — and how to make sure it pays.
🍸 Cocktail Endnote
The Long Game
A low-ABV, highball-style drink built for the season and the season pass.
The highball is less a specific recipe and more a template — a modular framework of spirit + soda (or tonic) + flavor. It’s a drink meant to go the distance: sessionable by design, light enough to pace yourself, and easy to build variations on depending on mood, season, or setting.
In other words it’s built for repeat engagement. Which, coincidentally, is exactly what streaming platforms need right now.
In a moment where ad-supported viewing is rising and audience attention is stretched thin, this isn’t a time for heavy pours or showy tricks. It’s a time for drinks — and strategies — that can sustain. Like the first warm evenings of spring, what we need right now is clarity, lightness, and something that feels fresh without being fragile.
The thyme brings in an herbal lift — a reminder that the season is shifting, and so is the streaming landscape. Floral, aromatic, quietly complex. Just enough personality to stand out, but built to keep going. This one’s for the planners, the schedulers, the ad ops teams looking to balance endurance with flavor.
Ingredients
1.5 oz dry vermouth
0.5 oz fino sherry
1 oz grapefruit cordial (see note below on how to prep your own or buy El Guapo’s Paloma Mixer)
Top with soda
2 sprigs fresh thyme — one muddled, one for garnish
Orange twist
Method
Lightly muddle one sprig of thyme in a highball glass. Add vermouth, sherry, and grapefruit cordial over ice. Top with soda water and stir gently. Garnish with an orange twist and the remaining thyme sprig.
🍊 About the Grapefruit Cordial
If you don’t have grapefruit cordial on hand, it’s easy to make your own. Combine equal parts fresh grapefruit juice and sugar (1 cup each is a good start), then add a touch of zest and gently heat until the sugar dissolves — but don’t boil it. Let it cool, strain, and bottle. Keeps in the fridge for about two weeks.
Think of it as grapefruit syrup with a citrus backbone — sweet, tart, and just complex enough to stand up in a highball.
Pairing note
Built for sustainability, not spectacle. The herbal top note adds a breath of spring and a reminder that not everything needs to be loud to be lasting. Like streaming’s next phase, it’s about quiet precision, modular frameworks, and pacing for endurance. Distribution is a marathon. Sip accordingly.


