Comcast Stacked the Whole Media Pyramid. Then It Started Taking It Apart.

On May 22, 2025, Comcast opened Epic Universe in Orlando — the first theme park the company ever built from scratch, rather than inherited.⁹ Six months earlier, it had announced it was giving away a chunk of the empire that made parks possible: USA Network, CNBC, MSNBC, E!, Syfy, Golf Channel — a $7 billion-a-year slice of cable television, spun off into a separate company it would no longer own.⁷ One hand was pouring concrete in Florida. The other was quietly dismantling the very layer of the business that, fourteen years earlier, had been the whole point of the deal.

The official story is that Comcast assembled a beautiful vertical stack — it owns the pipe into your house, the channels that travel down it, a European twin in Sky, the studios that feed everything, and parks to merchandise the characters. The real story is that the stack was a thesis about reinforcement, and the reinforcement stopped working. Cord-cutting hollowed out the distribution moat, and Comcast is now narrowing on purpose the empire it spent more than $60 billion widening.

The two-step deal almost everyone misremembers

Start with the founding myth, because the myth is wrong in a way that matters. People say Comcast bought NBCUniversal in 2011. It didn't. On January 28, 2011, Comcast closed on a 51% controlling interest, handing GE about $6.2 billion in cash for the right to consolidate the asset and name the strategy.¹ Full ownership came only on March 19, 2013, when Comcast paid $16.7 billion for GE's remaining 49% — and threw in another $1.4 billion to buy the buildings themselves, 30 Rockefeller Plaza and CNBC's New Jersey home.² Two steps, two prices, two years apart. The detail matters because it shows Comcast's confidence escalating: control first, then conviction. By 2013 the company wasn't just managing NBCUniversal — it was betting the house on the idea that a cable distributor should also be a content factory.

The whole bet, in one sentence

Here is the thesis Comcast sold to regulators and to itself: own the pipe and the programming so each one strengthens the other. The cable network feeds subscribers to the channels; the channels make the pipe worth paying for; the studios fill both; the parks turn the studios' characters into rides and toys. Every layer is supposed to make the layer above and below more valuable. That is the entire logic of an adjacency stack — you don't buy adjacent businesses because they're good businesses, you buy them because they make your core business harder to leave. For a decade, it held. Then the core itself began to leak.

Sky was the stack's most expensive expression. In early 2018, Comcast made a public offer for the European pay-TV operator at £12.50 per share — a 16% premium meant to top a sitting bid from 21st Century Fox.⁴ It was supposed to be the opening move. It became, instead, the floor of a bidding war. By the time the dust settled, Comcast had won Sky at £17.28 per share, a roughly 38% leap over its own first number, for total cash consideration of £30.2 billion — about $39.4 billion.³ The same vertical logic, transplanted to Europe: a pipe plus the content to fill it. Comcast paid auction-war prices for the privilege of owning the thesis twice.

Where the money actually comes from now

Strip the romance away and look at the 2024 books, and the stack reveals what it really is: a connectivity business wearing a media costume. Of Comcast's $123.73 billion in full-year 2024 revenue, the Residential Connectivity & Platforms segment — the pipe — accounted for $71.57 billion, very nearly 58% of the whole company. Media came to $28.15 billion, Studios $11.09 billion, and the celebrated Theme Parks, the part everyone wants to talk about, just $8.62 billion — under 7%.⁵ The adjacencies are real, but they orbit a center that is overwhelmingly broadband and cable. And that center is exactly what cord-cutting attacks.

The mechanism of the unwind is straightforward once you see it. The stack was supposed to be mutually reinforcing — but reinforcement runs in both directions, which means it can also be mutually draining. When cable subscribers leave, the cable networks lose the captive audience that made carriage fees flow. The pipe stops needing the channels to justify itself, because customers increasingly want broadband, not bundles. The two assets that were designed to prop each other up start to weigh on each other instead. So in November 2024, Comcast did the unsentimental thing: it announced a tax-free spin-off of USA, CNBC, MSNBC, Oxygen, E!, Syfy, and Golf Channel — plus digital assets like Fandango and Rotten Tomatoes — into a new public company, roughly $7 billion in annual revenue, cut loose as linear economics deteriorated.⁷ NBC, Bravo, Telemundo, the parks, the studios, and Peacock stayed. The cable-network layer — the one the 2011 deal was built around — was the layer sent away.

But Peacock was supposed to be the bridge

The strongest defense of the stack is that Comcast saw cord-cutting coming and built a bridge across it — Peacock, the streaming service meant to carry the content layer into the post-cable world. If the bundle dies, the argument goes, the same shows and studios just travel down a different pipe. It's a fair objection, and Peacock is a real product. But the numbers say the bridge is not yet carrying its own weight. Peacock ended 2024 with 36 million subscribers — flat against the prior quarter and short of the roughly 37.6 million analysts expected — and posted an adjusted EBITDA loss of $1.79 billion for the year on $4.9 billion of revenue.⁶ A bridge that loses nearly $1.8 billion a year is a bridge still under construction, not one you've finished crossing. The streaming layer may eventually replace the cable-network layer Comcast is shedding. As of the most recent reporting, it is replacing it at a steep operating loss — which is precisely why the company is keeping Peacock while spinning the legacy channels off. It is betting on the new bridge by abandoning the old road.

Comcast's adjacency stack was never wrong, exactly. It was a coherent bet that distribution and content would keep needing each other, made by a company confident enough to pay auction prices in two countries to own both ends of the wire. The bet's flaw wasn't the assembly — it was the assumption that the bottom of the pyramid would hold. When it didn't, the same architecture that made the pieces stronger together made them heavier together, and the only move left was subtraction. Epic Universe rises in Orlando while the cable networks float free as Versant.¹⁰ Same company, opposite math. The genius was building the stack. The harder, later wisdom was knowing which floor to take out before the whole thing leaned on it.