A Short Taxonomy of Bundle Collapse

The bundle was supposed to save the business.

Sometime around the spring of 2024 — and the pivot was marked, for me, by the May 2024 announcement that Disney+, Hulu, and Max would sell as a combined package — the streaming industry collectively conceded that the Great Unbundling was over. The services that had spent a decade explaining why the cable bundle was a relic, a regressive cross-subsidy, a tax on consumers who only wanted the good parts, now stood up and said, with straight faces, that actually what the consumer really wanted was a bundle. Just a better one. A smarter one. A streaming-native one.

I watched this happen with some amount of grim amusement. I had written the previous fall about what the death of the universal remote really meant — that the subscription bundle was a child of the cable era and that it could not simply be reassembled in software. I did not think that view would be validated as quickly as it was.

In the eighteen-ish months since the Disney-Hulu-Max announcement, the industry has tried a lot of bundle shapes. Co-branded stacks. Telco-plus-streamer sleeves. Sports-plus-entertainment packages. Plus-Plus tiers. Family tiers. Household tiers. Wallet-based promotional stacks routed through credit card co-brands. A joint sports product — Venu — announced with fanfare on February 6, 2024, blocked by a federal judge in August 2024, and abandoned by its three partners in January 2025 before it shipped a single minute.

Most of these bundles have collapsed. The ones that have not collapsed yet are trending, by any reading, toward collapse. And the failures are not random. They fall into a small number of recurring modes, and the modes rhyme, and once you see the modes you cannot stop seeing them.

This essay is a short taxonomy. Four modes. Each of them is a different way a streaming bundle dies.

Mode One: The Dual-Sub Stack

The first and most common mode is what I will call the dual-sub stack. You take two standalone subscription services, staple them together, discount the combined price by some amount, and launch the result as “a bundle.”

On a whiteboard, the math looks compelling. Service A costs fifteen dollars. Service B costs twelve. Together they would be twenty-seven. You offer them together for twenty-two. Consumers save five dollars a month. Both services get credit for a new incremental subscriber. Churn goes down because users are now anchored to two content libraries instead of one. Everybody wins.

In practice, almost nobody buys the dual-sub stack. The consumer who would have bought both services separately buys the bundle and saves five dollars, which is not a new acquisition — it is margin compression dressed up as growth. The consumer who would have bought only one service is not actually going to add a whole second service just to save money on the one they wanted. They look at the bundle, compare it to their single-service subscription, and keep the single-service subscription.

The dual-sub stack, in other words, does not expand the market. It just re-prices the intersection of the two existing markets. And it re-prices that intersection downward.

This is what has happened with most of the headline “bundle” announcements of the last eighteen months. Disney has not, to date, publicly disclosed a specific take rate for the Disney+/Hulu/Max bundle — an absence that tells you something in itself, given how prominently the bundle was positioned at launch. What has been disclosed is that the bundle’s monthly price moved up in October 2025, from $16.99 / $29.99 to $19.99 / $32.99, which is the move a service makes when discount-driven acquisition is not paying for itself. The stack exists. You can buy it. The services have a line item in their investor decks that proudly reports the bundle as a retention lever. But the actual behavior — the movement of subscribers from one-service to two-service — is much smaller than the whiteboard said it would be, and the discount-driven margin compression is much larger.

The dual-sub stack is the bundle shape that looks most like a bundle on paper. It is, for exactly that reason, the one that fails first.

Mode Two: The Zombie Middle Tier

The second mode is what happens when a single service tries to bundle itself, internally, by creating a tier structure that is supposed to satisfy everyone and instead satisfies no one.

Almost every streamer now has some version of this. There is an ad-supported tier at the bottom, at a promotional price. There is an ad-free premium tier at the top, at a price that keeps climbing. And there is, in between, a middle tier that nobody at the company can really explain.

The middle tier was supposed to be the bundle inside the service. A little cheaper than premium, a little richer than the ad tier. It was supposed to capture the subscriber who was price-sensitive but did not want ads, or who wanted ads but wanted some kind of upgrade in quality. On a whiteboard, it looks like a reasonable compromise.

In practice, the middle tier is a zombie. Nobody chooses it deliberately. The subscribers who land on it are mostly legacy — they were on the premium tier when it was cheaper, got price-migrated to the current middle tier during some reorganization, and have not yet bothered to check whether the new structure makes sense for them. They will check, eventually. And when they check, some of them will move up to premium and some of them will move down to the ad tier and some of them will cancel entirely.

The middle tier is a zombie. Nobody chooses it deliberately. The subscribers who land on it are legacy, and when they check what they are paying for, they leave.— the thesis

The zombie middle tier is a bundle in the sense that it is trying to be many things to many people at once. It is failing for the same reason the dual-sub stack fails. The bundle is not expanding the market. It is cannibalizing the two good tiers that bracket it.

Netflix did the honest thing. It eliminated the Basic plan for new customers in 2023 and migrated existing Basic subscribers off the tier in June 2024, collapsing its pricing architecture into a cleaner three-point menu: ad-supported, standard, premium. Most of the other major streamers are still defending their middles, and their middles are still emptying. The ones who keep the middle tier are usually doing it because eliminating it would require admitting, publicly, that the previous pricing architecture was wrong. That is a harder conversation than continuing to report the tier.

Mode Three: The Co-Branded Tourniquet

The third mode is the most embarrassing one. The co-branded tourniquet is a bundle announced, usually with press-release fanfare, in response to a churn crisis. Two services — typically ones that are not natural partners — stand up together on a stage, announce a joint offering, describe it as strategic, and then quietly abandon it within a quarter or two when the underlying numbers do not move.

The Venu Sports arc is the canonical example, and it did not even make it to market. Disney, Fox, and Warner Bros. Discovery announced on February 6, 2024 a joint sports streaming service. It was positioned as the answer to the cord-cutting-and-sports problem. FuboTV sued on antitrust grounds, and on August 16, 2024, federal judge Margaret Garnett granted a preliminary injunction blocking the service from launching. The partners began appealing. And then, on January 10, 2025, Disney, Fox, and Warner Bros. Discovery announced they would not pursue Venu at all — the same week Disney announced a merger of Hulu + Live TV into Fubo that resolved the underlying lawsuit. Eleven months from announcement to abandonment. The service never launched.

The thing to notice about Venu is that it was the correct idea, in the abstract. A single destination for the sports that are otherwise scattered across three disconnected services is, structurally, what the user actually wants. I have written before about why a consolidated sports experience would outperform a fragmented one by a very large multiple — the audience will find the game if the game is findable.

The reason Venu collapsed was not that the product idea was wrong. It was that the three partners could not agree on whose P&L the bundle would protect, what the sub-economics of a joint venture between three rivals would look like, and — most importantly — whose churn problem the bundle was solving for. All three companies were trying to use the same bundle as a tourniquet on three different wounds. When the courts intervened, the partners were ready to walk away. The bundle had not been strategically necessary to any of them. It had been politically expedient for all of them.

Every co-branded tourniquet launched since has some version of this problem. Two partners, each solving their own quarterly churn, stapling together a temporary offering and calling it strategy. The offerings get announced with phrases like “long-term partnership” and “reimagining the consumer experience.” They get deprioritized, quietly, within two to three quarters when the partners figure out that the bundle was not saving either of their balance sheets, and then they get unwound in an even quieter press release some weeks after that.

The co-branded tourniquet fails because it was never designed to succeed. It was designed to show motion on a quarterly earnings call. The motion was the product. Once the quarter had closed, the bundle’s job was done.

Mode Four: The Sports-Shaped Hole

The fourth mode is the one I find most structurally interesting, and it is the one most relevant to the work I do every day. Call it the sports-shaped hole.

Every bundle that has been attempted in the last eighteen months has had, at its center, an unresolved question about sports. Sports is the content category with the most retention value per minute, the highest ad load tolerance, the lowest substitutability, and the worst fit with the rest of the streaming library. Sports lives on a different calendar. It has a different rights structure. It is licensed in ways that make it hard to include in a general-entertainment bundle and hard to exclude from one without creating a visible gap.

So every bundle ends up building around a sports-shaped hole.

Some bundles put sports in a separate tier and charge extra for it, which makes the bundle feel like a cable bundle with an extra sports charge, which is the exact thing the streaming pitch was supposed to eliminate. Some bundles include only a partial sports lineup, which generates constant support tickets asking where the other games are. Some bundles exclude sports entirely and watch their churn spike in October and February and then again during playoff season, when the users who were willing to tolerate a sports-free service for a few months finally give up and go to wherever the games are.

The underlying issue is that sports rights are currently structured across too many services for any single bundle to solve the sports problem. The NBA’s eleven-year deal announced on July 24, 2024, for roughly $76 billion, split rights across Disney (ESPN / ABC), NBC (NBCUniversal / Peacock), and Amazon (Prime Video), with the first season of the new contract beginning 2025–26. The NFL has games on CBS, Fox, NBC, ESPN, Amazon, and a Sunday Ticket package that moved exclusively to YouTube TV beginning with the 2023 season in a reported $2 billion annual deal. The regional sports network business, meanwhile, has collapsed into itself — Diamond Sports Group filed for Chapter 11 on March 14, 2023 and eventually emerged as Main Street Sports Group, with the surviving networks now pushing direct-to-consumer wrappers because the cable-distributed RSN model is no longer solvent.

No bundle currently in market solves all of that. None can, structurally. The rights are not available to be consolidated at bundle-construction time. And so every bundle announcement in the last eighteen months has included some version of the line “we are working to bring more sports to the platform over time,” which is a sentence that means, in practice, that the sports-shaped hole will remain a sports-shaped hole for the foreseeable future.

A bundle with a sports-shaped hole is not a bundle. It is a fragment marketed as a unified experience. Users figure this out quickly. Most of them do not object loudly. They just keep a second subscription to whatever service carries the games they care about, and the bundle becomes the thing they use for everything other than the thing they most wanted the bundle to solve.

The Bundle That Works Looks Least Like a Bundle

If the four failure modes all have in common that they try to look like bundles — announced as bundles, marketed as bundles, invoiced as bundles — then the implication of the pattern is that the bundle which actually works will look, on its surface, almost nothing like a bundle at all.

I think that is right, and I think the shape of the successful bundle is already visible in outline.

It is a single service with a single interface and a single remote action. You open one app. You browse one library. You press play on one title and the thing plays. Behind the scenes, the content may be coming from six different rights holders and ten different billing arrangements. The user does not know and does not care. The experience is unified.

That is what a real bundle is. It is not an invoice decision. It is an experience decision. The cable bundle, in its time, worked not because it discounted the per-channel price but because it restored a single remote — one set of buttons, one channel guide, one expectation about how the television worked. You did not think, when you clicked up and down through channels, about which content company owned which channel. You just watched television.

The industry has not yet built the thing I am describing. Most of what is being marketed as a bundle today is an invoice staple — two subscriptions sold together, two billing events, two apps, two remotes. That is not a bundle. That is two services with a shared line item.

The successful bundle will be something different. It will be one service with one remote that happens, under the hood, to have licensing agreements with many rights holders. It will not be launched with a press event. It will not be co-branded. It will not have a dedicated landing page explaining the partnership. It will just be the service that works, and it will grow because it is the thing users actually want, which is not a bundle discount but a single act of turning on the television.

What The Next Eighteen Months Will Show

I suspect we are at the tail end of the bundle-announcement era. The four failure modes have repeated enough times that even the executives most invested in the bundle narrative are starting to notice that the announcements are not producing the subscriber numbers the announcements predicted. The dual-sub stack is not converting. The zombie middle tiers are being collapsed one by one. The co-branded tourniquets are unwinding. The sports-shaped holes remain unresolved.

The next eighteen months, I think, will look different. Fewer bundle announcements. More consolidation. More services buying each other rather than partnering with each other. More direct-to-consumer sports services rolled into general-entertainment services rather than offered as separate tiers. More library content flowing between companies as licensing deals replace failed partnerships. The single-remote experience will start to emerge — not from any one company’s strategic masterstroke, but from the accumulation of consolidation decisions driven by the fact that none of the standalone strategies are working.

The companies that will be around in 2030 are the ones that are already designing for that single-remote outcome. The ones that are still announcing bundles in 2027 will be the ones that did not notice the pattern had changed.

The bundle, as currently marketed, is dead. The bundle, as experienced by the user, has not even been built yet.

That is where the next decade of this business is going to be won.