The DialHow broadcast television actually works — and what the networks and their local stations are worth now.

A television station is a six-megahertz hole in the air

Everything else about the business is downstream of one physical fact: the FCC licenses a broadcaster to radiate on a 6 MHz-wide channel over a defined geographic area, and nobody else may use that frequency there. That license is the asset. The studio, the anchors, the news van — those are costs.

The channel numbers you grew up with are frequency allocations. Channel 2 is 54–60 MHz; channel 7 is 174–180 MHz; channel 14 is 470–476 MHz. Two bands were carved out in the 1940s and 50s — VHF, which propagates far and bends around terrain but needs an enormous antenna, and UHF, which is closer to line-of-sight but supports far more channels. A full-power UHF station radiating up to 1,000 kW from a tall tower covers roughly a 60-mile radius.

Since the 2009 digital transition there's a wrinkle: the number on your screen is a virtual channel, carried in the signal's metadata. WNBC still says "4" on your TV, but it hasn't transmitted on RF channel 4 in years. The branding survived; the physics moved.

What fits in the pipe

A 6 MHz channel under ATSC 1.0 — the digital standard the US has used since 1998 — carries exactly 19.39 Mbps. Not "about." That is the fixed payload of 8VSB modulation in a 6 MHz channel, and it is the single most consequential number in over-the-air picture quality.

That budget is why your local station's picture sometimes looks softer than the same network feed on cable. Subchannels — the .2 and .3 feeds running MeTV, Antenna TV, a weather radar loop, a shopping channel — are free money for the station. Each one is also bandwidth stolen from the main HD feed, because they all share the same 19.39.

ATSC 3.0, and why the transition is a fight

The replacement standard, marketed as NextGen TV, throws out 8VSB for OFDM, wraps everything in IP, and swaps MPEG-2 for HEVC. Practically: roughly 25 Mbps of far more efficient video in the same 6 MHz, 4K and HDR, per-household targeted advertising, and the ability to datacast to things that aren't televisions at all.

"Not backward compatible" is the whole problem. Until now the FCC has required stations moving to 3.0 to keep simulcasting a 1.0 signal, which means finding a partner station's spectrum to host it — an expensive, market-by-market game of musical chairs. In October 2025 the Commission proposed ending that requirement. The broadcasters' trade association wants the top 55 markets fully converted by February 2028 and everyone else by February 2030; public broadcasters have asked the FCC not to set a hard date at all, because their viewers skew toward exactly the households least likely to buy a new television. As of this writing there is no final order, and the tuner-mandate and DRM questions are still open.

The network is not the station

This is the part most people have backwards. ABC, CBS, NBC and Fox do not transmit anything. A network is a programming supplier and a national advertising sales house. It owns a handful of stations outright — the O&Os, concentrated in New York, Los Angeles, Chicago and a dozen other big markets — and in every other market it rents reach from a station somebody else owns.

There are about 1,390 commercial full-power television stations in the US, plus roughly 390 noncommercial ones, distributed across 210 markets. Nielsen draws those markets — DMAs, designated market areas — county by county, with no gaps and no overlaps, so every American household sits in exactly one. Market 1 is New York. Market 210 is Glendive, Montana, with about 3,500 television households.

Each network feeds its schedule by satellite to roughly 200 affiliates, who retransmit it locally. The day is carved up by contract:

A network affiliation is not permanent. Affiliates switch — sometimes spectacularly, as when Fox bought its way into the NFL in 1994 and pulled twelve CBS stations over to the new network in a single stroke, a move that is still the reference case for what network programming is worth to a local station.

Follow the money — and watch one arrow flip

Three streams of money run through broadcast television, and the direction of the third one is the single biggest change in the industry's economics in the last twenty years.

1 & 2 — the advertising, which everybody understands

National advertisers buy the network; local advertisers — car dealers, personal injury lawyers, furniture stores, and every other election year, political campaigns — buy the station. Each side keeps what it sells. That split has been stable since the 1950s.

3 — retransmission consent, which nobody outside the business understands

The 1992 Cable Act gave every commercial station a choice, renewed every three years: demand must-carry, where the cable system has to carry you but pays nothing, or elect retransmission consent, where the cable system must negotiate for your signal and may be told no. For over a decade almost nobody could actually extract cash. Cable operators paid in kind instead — that is literally how Fox News got its initial carriage.

Then, starting around 2006, stations began winning real money. And because the network supplies the programming that makes the station worth carrying, the network came back for its cut. That cut is reverse compensation: the affiliate pays the network a per-subscriber fee for the right to keep the affiliation.

This is why a station group's headline retransmission revenue is nearly meaningless on its own. The number that matters is net retrans — gross receipts from cable, satellite and streaming bundles, minus what the network takes back. Gray Media's net retransmission revenue was $150 million in the second quarter of 2026, up 10% from $136 million a year earlier, even as the gross number had been falling. Both the subscriber count and the reverse-comp rate move, and they don't move together.

Twenty years, in four numbers

In 2005 the networks were an advertising business with a ratings problem. Today they are a sports-rights business with a toll booth attached, and the toll booth has started to leak.

The audience, meanwhile

Broadcast's advertising currency is the rating among adults 18–49 — one rating point is one percent of that population. Here is the entire 2025–26 primetime season:

Two honest caveats. Nielsen's methodology has changed, and today's numbers exclude a great deal of delayed and streamed viewing that eventually gets counted elsewhere. And the Big Four still sold a record $9.3 billion of upfront advertising in 2006 — the absolute dollars did not fall as fast as the ratings, because scarcity of mass reach props up the price per viewer. But the direction is not in dispute, and the networks' answer to it has been the same for a decade: buy more sports.

So what is a broadcast network actually worth?

Hard to answer directly, because three of the four are buried inside conglomerates that never report them separately. Fox is the exception, and Fox's own segment disclosure is the most revealing document in American television.

Fox Corporation splits into a Television segment — the Fox network, its owned stations, and Tubi — and a Cable Network Programming segment that is overwhelmingly Fox News. Here are both, over the first nine months of fiscal 2026:

The broadcast business converts about ten cents of every revenue dollar into operating profit. The cable business converts forty-two. Broadcast carries the transmitters, the local news payrolls, the affiliate relationships, the FCC licence obligations, and — above all — the sports rights.

The sports problem, which is also the sports answer

The NFL's current media agreements run 2023 through 2033 and total roughly $110 billion — about $10 billion a year, with CBS, Fox and NBC each paying somewhere in the $2.0–2.2 billion range annually and ESPN/ABC around $2.7 billion. There is an opt-out roughly seven years in, which means this all gets repriced around 2029, with Amazon, Netflix and YouTube in the room.

Sports is why broadcast still matters — it is the last reliable way to put fifteen to twenty million Americans in front of the same thing at the same time, and the rights are contractually tied to free over-the-air reach. It is also what makes the margins thin. The networks are, in a real sense, pass-through vehicles for sports leagues now.

The revealed preferences

• Disney floated selling ABC in 2023. Analysts put it around $8 billion — for a network that reaches every US household. Disney didn't sell. It sold its 50% of A+E to Hearst instead, for about $1.2 billion, closing August 2026.
• Comcast spun off its cable networks and kept NBC. Versant — CNBC, USA, Syfy, Golf, E!, MS NOW — began trading on 5 January 2026 at a deliberately modest valuation. The broadcast network stayed. That is the clearest statement anyone has made about relative value: cable nets are the melting asset, broadcast is still worth holding for sports and reach.
• Paramount, under Skydance since August 2025, went and bought Warner Bros. Discovery — $31.00 a share, roughly $111 billion, DOJ-approved in June 2026 and now tied up in litigation with an antitrust trial calendared for March 2027. Note what the prize was: the studio, HBO and the streaming subscribers. CBS was already in the building and nobody talked about it much.

The fair summary: the parents are worth a great deal. The networks, valued as standalone assets stripped of their parent's streaming service and their sports contracts, are worth surprisingly little — a single-digit-billions business attached to a multi-billion-dollar annual obligation. What they still own is the last mass audience in America, and an obligation to pay dearly for it.

The local affiliates are where this gets grim

Local stations were rolled up into a handful of public companies during the 2013–2019 merger wave, financed with debt, underwritten on an assumption that retransmission revenue would keep compounding. It didn't. The equity is what absorbed the difference.

S&P has Sinclair, Gray, Scripps and Cox Media Group all on negative outlook, with debt maturities concentrating from 2026 onward and refinancing risk if free cash flow doesn't go to paydown. Scripps announced a cost-cutting programme and layoffs in 2026; its net leverage sits at 3.9×.

The revenue mix explains the fragility. Across the sector roughly half of station revenue is advertising, about 41% is retransmission, and the remaining 9% is digital. Depending on the group, somewhere between a third and half of all revenue rides on retrans — a line that is now flat to declining, and on which the network takes its cut first.

Fighting on two fronts at once

Affiliates are now in conflict with their distributors and with their own networks, simultaneously.

Against distributors: blackouts. E.W. Scripps pulled 54 stations off DirecTV in 36 markets on 31 May 2026 — Cleveland, Denver, Detroit, Phoenix, Tampa, Baltimore — and kept them off for five weeks before a deal on 10 July. This is now a routine negotiating instrument rather than a crisis.

Against their networks: in 2026 the affiliate boards of all four networks filed joint comments at the FCC. Two complaints. First, they have no ability to negotiate their own rates with virtual providers like YouTube TV — those terms are network-dictated, take it or leave it. Second, and more fundamentally, the networks are now delivering the same premium programming directly to consumers through Peacock, Paramount+, Fox One and Hulu, bypassing the local station that pays for the affiliation. The affiliates have also asked the FCC to protect local sports rights from being pulled up to the national level. Chairman Carr's FCC has opened a proceeding into the network-affiliate relationship.

Consolidation is the plan, and it is stuck in court

The industry's answer is scale: bigger groups, more leverage against both distributors and networks. The test case is Nexstar's acquisition of Tegna, and it is a mess.

The one thing going right

Political advertising. 2026 is a record midterm: BIA projects $9.7 billion of local political spending, with AdImpact putting roughly $5.6 billion of it on broadcast television — about 30% above 2022. Stations are required to sell candidates time at the lowest unit rate, but the volume more than compensates, and it lands disproportionately in swing-state markets.

It is also a drug. Even-year political revenue masks the underlying trend, odd years are brutal by comparison, and no station group has ever been able to borrow against it.

Four things to watch

• The ATSC 3.0 endgame. No final FCC order yet. If the simulcast requirement dies and a hard sunset lands, broadcasters get a genuinely modern pipe — addressable advertising, datacasting, automotive and emergency services they can sell outside television entirely. If a tuner mandate doesn't come with it, they get a standard most households can't receive.
• Whether the ownership-cap repeal survives. The FCC replaced a statutory limit by rule. That is exactly the argument that will be made against it. If it holds, expect the remaining groups to consolidate down to three or four; if it falls, Nexstar–Tegna may be the high-water mark.
• The 2029 sports repricing. The NFL's opt-out arrives with streamers holding more cash than the networks. A package moving off broadcast wouldn't just cost a network ratings — it would strip the affiliates' reason to pay reverse compensation.
• Whether the affiliate model survives the networks' own apps. The FCC proceeding into network-affiliate relations is the first regulatory look at this in a generation. There is no obvious remedy that doesn't take money from one side and hand it to the other.

None of this touches the transmitters. A station will still be a 6 MHz channel radiating over a county map drawn by Nielsen, and somebody will still own that licence. The open question has never been whether broadcast television works. It is who gets paid for it.