Comcast Isn't a Media Company. It's a Broadband Toll Booth Bankrolling One.

Picture the cable that runs into your house. It carries your Netflix, your work calls, your kid's homework — and it does not care which. Comcast charges you for the pipe, not the water. In 2024 that pipe, the segment Comcast calls Connectivity & Platforms, threw off $45.1 billion in connectivity revenue at a 38.2% Adjusted EBITDA margin.¹ That is not retailer economics. That is toll-booth economics. And somewhere upstream of all that cash sits a streaming service bleeding money, a film studio, theme parks, and seven cable channels nobody under forty watches. The pipe pays for all of it.

The official story is that Comcast is a content company that also happens to do cable. The 10-K says the reverse, and it says it in numbers. The broadband business is the company. NBCUniversal is the passenger — an expensive, glamorous passenger riding a fare it could never afford on its own.

The pipe earns the money; the screen spends it

Here is the structure most coverage gets backwards. Comcast's broadband network was built once and now sells access at a margin that looks like software, not telecom — 38.2% on $45.1 billion, even after $8.3 billion of capital spending kept the network current.¹ That margin is the engine. The whole company generated $12.5 billion of free cash flow in 2024,² and the connectivity segment is where the overwhelming majority of it is manufactured. Meanwhile, Peacock — the crown jewel of the media side — grew revenue 46% to $4.9 billion and still lost money.² One business prints cash with no marginal cost; the other consumes it buying sports rights and original shows. The cable that ignores what it carries is, with quiet irony, funding the things it carries.

What Comcast actually spun off — and what it kept

When Comcast announced in November 2024 that it would spin off 'NBCUniversal' networks, the press shorthand made it sound like the media empire was leaving.⁴ It wasn't. What got cut loose — completed as Versant Media in early January 2026¹⁰ — was only the legacy cable channels: USA, CNBC, MSNBC, E!, Syfy, Oxygen, Golf Channel, plus oddments like Fandango and Rotten Tomatoes.³ That bundle generates roughly $7 billion in revenue a year,³ and it is the part of the empire most exposed to cord-cutting — the channels whose audiences age out and whose carriage fees face pressure with every dropped cable package. NBC broadcast, Peacock, Telemundo, the studios, and the theme parks all stayed.⁴ Read correctly, the spin-off is a confession: Comcast quietly amputated the declining cable channels so the broadband cash could concentrate on the one media bet it still believed in.

This is the cross-subsidy in motion. The broadband engine funds the streaming bet long enough for that bet to mature — and Peacock's trajectory says the bet might be working. Its 2024 Adjusted EBITDA loss narrowed by nearly $1 billion year-over-year,² and by the first quarter of 2026 it had reached 46 million paying subscribers, even while still posting a $432 million quarterly loss.⁸ The popular 'Peacock is just burning billions' framing misses the direction of travel. The losses are real. They are also shrinking, on a service the pipe can comfortably afford to wait on. That patience is the whole point of owning a cash cow.

The engine is starting to lose customers

Every cross-subsidy lives or dies on one assumption: that the cash cow keeps giving cash. Comcast's just began to wobble. The company lost 411,000 domestic broadband customers in 2024, ending at 31.8 million.¹ In the fourth quarter alone it shed 139,000 — worse than the roughly 100,000 its own cable CEO had telegraphed to investors in December. President Mike Cavanagh called it 'disappointing,' and the stock fell 11% on the day.⁷ The erosion has a source: fiber overbuilders and fixed-wireless from the mobile carriers are, for the first time, offering homes a real alternative to the cable pipe. By the first quarter of 2026 the bleeding continued — another 65,000 broadband subs gone — and Cavanagh told analysts plainly, 'The competitive environment remains intense.'¹¹

Isn't this just a healthy company moving money around?

The fair objection is that none of this is alarming — Comcast generated $19.2 billion of free cash flow in 2025, more than enough to fund Peacock and buy back stock besides.⁵ True today. But a cross-subsidy is only as durable as the gap between the cow and the cost, and both sides of that gap are moving the wrong way at once. The broadband engine is shrinking subscribers into a competitive market that did not exist five years ago, while media investment — sports rights, streaming content, the scale Peacock needs to reach 46 million subscribers — is climbing. There's an honest counter to the alarm, too: the broadband losses are partly a measurement artifact. Comcast's own Q3 2024 filing showed that, stripping out the expiration of a federal subsidy program, the company would have ADDED 9,000 broadband customers that quarter.⁶ The structural erosion is real but smaller than the headline losses suggest. So the verdict is not collapse. It is compression — a margin engine being squeezed at the exact moment the media side needs it most, with no second engine standing by to take over.

Comcast spent a decade building a toll booth on the internet and a media empire that drinks from it. For years the booth's serene indifference to what it carried was the whole genius: it charged you for the pipe, never the water, and used the proceeds to bankroll the water other people would pay extra for. The structure still works. But the traffic is, for the first time, finding other roads — fiber here, a wireless signal there — and the media side is thirstier than ever. The question was never whether the broadband cow could fund Peacock. It's whether it can fund Peacock long enough for Peacock to learn to feed itself. The cow is still giving. It's just giving a little less, every year, on schedule.