The hard cap
Thirty-two rosters, one negotiated number. How the cap is set, what a contract actually pays, what ‘guaranteed’ really guarantees, and every door — draft, tags, waivers, trades, compensatory picks — through which players arrive and leave.
How NFL teams make money — and how it is policed
National TV money shared 32 ways, the players’ 48 per cent, the Packers’ public accounts as the league’s only open window, spending floors, debt and ownership rules, private equity — and what happens to cap cheats.
Everything else in this guide happens beneath a number; this section is about the machine that makes the number. The NFL cleared roughly $23B of revenue in its 2024 fiscal year, about 60% of it arriving through nationally shared streams — television above all — and the collective bargaining agreement pipes roughly 48 cents of every one of those dollars to the players. The cap is not a policy choice made fresh each spring: it is the players’ share, less benefits, divided by 32.
The league’s books are otherwise closed — 31 clubs publish nothing. What follows is the plumbing anyway: the CBA’s revenue split drawn as one dollar, the television rows that feed it, the single public window Green Bay opens every July, the cash rules that bind owners more tightly than the cap does, and the audits that punish the rare club caught keeping two sets of books.
One dollar of NFL money
Band heights are proportional, so the areas are honest. The ≈60/40 national–local split is reported, and the media-vs-Ventures widths are indicative — the league publishes no exact split — but the 55/45/40 shares and the 48–48.8% band are the CBA’s own numbers (Art. 12 §6), and the figure’s arithmetic lands on ≈48¢ by itself: run 55/45/40 over a 60/40 dollar and the blend the lawyers banded is simply what falls out.
This is the landing section’s little formula diagram with the pipes drawn in — the cap section works the five steps one by one, and its history chart shows what sixty years of these pipes widening has done to the number.
The dollar, in words
Call the league’s ≈$23B one dollar. About 60¢ of it arrives through national streams — league media much the biggest piece, NFL Ventures and the postseason the rest — and is pooled into 32 identical cheques; the other ≈40¢ is local money (gate, sponsorship, premium seating) each club keeps and earns for itself. The players’ slice is taken bucket by bucket: 55% of league media, 45% of Ventures and postseason, 40% of local — a blend that lands almost exactly on 48¢, and the CBA bands it there: a 48% floor for 2021–2030, a 48.5% ceiling, and a media kicker that lifts 17-game seasons as far as 48.8% when new television deals grow fast enough. From the players’ ≈48¢, projected benefits come off first — $77.6M per club in 2026 — and what remains, divided by 32, is the cap: $301.2M. The other ≈52¢ runs stadiums, front offices, debt service and profit, under rules of its own, below.
The bargain underneath
The 2020 CBA was executed March 15, 2020 — days after that spring’s cap had been set, days before the world shut — and runs through the 2030 season. Its core trade: a 17th regular-season game for a banded share of All Revenue. The 55/45/40 blend is floored at 48% (47% in 2020) and capped at 48.5%, with the stadium credit never allowed to push below the floor. Then the kicker: a 17-game season is a “Covered Season”, and the share climbs through 48.5, 48.6 and 48.7 to 48.8% as new media deals grow 60, 100, 110 and 120% past the old $7.357B annual average. The 2021 contracts roughly doubled that average — which is why “players get 48.8% in media-kicker years” became the shorthand.
The number is set on projections and settled on actuals. Each spring’s cap is computed from projected revenue; when the season’s real figures land, the machinery trues up — shortfalls and overages flow back through subsequent player-cost calculations, and each club’s own line carries incentive true-ups from deals that paid out (or didn’t) against expectation. 2021 stretched the mechanism to its limit: after 2020’s ~92% attendance collapse, a straight formula application produced roughly a $160M cap, and the sides negotiated a $182.5M floor instead — borrowing the difference from the players’ future share, repaid as revenue returned through the record jumps of 2022 (+$25.7M) and 2024 (+$30.6M, the media deals’ full weight plus the remaining COVID repayment landing at once). The whole borrow-and-repay shape is visible in the history chart of the opening section.
“Benefits” is not a rounding term. The Player Cost Amount funds benefits before it funds salaries — pensions and post-career benefits, insurance, performance-based pay, the veteran salary benefit — and in 2026 that stack runs $77.6M per club on top of the $301.2M cap: $378.8M of total player spending per team, of which the cap is only the visible four-fifths.
One benefit deserves its own line, because it is the league running redistribution on itself. Performance-Based Pay (Art. 28) is a league-funded pool distributed each spring by an index of playtime over compensation: the more you played relative to what you were paid, the bigger the cheque. Funded at $8.5M per club in 2020 and $10M in 2021, growing with projected revenue by up to 5% a year, the 2024-season distribution topped $452M league-wide — about $14.1M per club — with Jets linebacker Jamien Sherwood’s $1,092,206 the largest individual payout.
The television rows
Nine rows explain most of the cap’s modern shape. The five packages signed March 18, 2021 — eleven-year agreements totalling ~$110B — sum to roughly $10B a season, about double the average they replaced: enough deal-over-deal growth to reach the media kicker’s top slot. Around them, the moves since: Sunday Ticket’s $500M-a-year raise on its way to YouTube, Netflix buying Christmas and then a five-game slate, the league swapping NFL Network for 10% of ESPN — and a 2029 opt-out everyone is already negotiating around.
| Partner | Package | Span | ≈ $/yr | Note |
|---|---|---|---|---|
| ESPN/ABC | Monday Night Football + two Super Bowls, flex scheduling | 2023–2033 | $2.7bn | The most expensive of the five 2021 packages at a reported ~$2.7B a year — eleven-year agreements signed March 18, 2021, totalling ~$110B. |
| Fox | NFC Sunday afternoon | 2023–2033 | $2.2bn | Reported ~$2.2B a year for the NFC Sunday afternoon package — the conference-package model unchanged since the 1990s, at forty times the price. |
| CBS | AFC Sunday afternoon | 2023–2033 | $2.1bn | Reported ~$2.1B a year for the AFC Sunday afternoon package — per-network fees are the widely reported approximations; the league has never published exact figures. |
| NBC | Sunday Night Football | 2023–2033 | $2bn | Reported ~$2.0B a year for Sunday Night Football — primetime’s flagship window, and the slot the 2026 Wednesday opener was built around. |
| Amazon Prime | Thursday Night Football, streaming-exclusive | 2022–2033 | $1bn | The first fully streaming package at a reported ~$1.0B a year — Amazon’s TNF began a season early, in 2022. |
| YouTube | Sunday Ticket (out-of-market), exclusive | 2023–2030 | $2bn | Seven years at ~$2.0B a season, up from DirecTV’s $1.5B. 2026 consumer pricing: a $240 new-user promo, $378 returning with YouTube TV, $480 standalone (tiers vary by bundle; treat precise pricing as medium confidence). |
| Netflix | Christmas games → five-game 2026 slate + NFL Honors | 2024–, extended May 2026 | $150M | Bought Christmas in May 2024 at ~$75M a game (~$150M for 2024’s doubleheader), then extended: the 2026 slate runs five games — Week 1 in Melbourne (49ers–Rams, the first regular-season game in Australia), Thanksgiving Eve, a Christmas doubleheader and a Week 18 Saturday window. |
| ESPN ⇄ NFL | NFL Network + linear RedZone + NFL Fantasy for 10% of ESPN | Closed Feb 1–2, 2026 | — | Not a rights fee but an equity swap: the NFL took a 10% stake in ESPN (≈$3B against ESPN’s ≈$30B valuation), leaving Disney ~72% and Hearst 18%. NFL Network folds into ESPN’s direct-to-consumer service for 2026; Disney filings note a post-July-2034 option to reacquire the league’s stake. |
| Everyone | The 2029 opt-out | After the 2029 season | — | The league may opt out of the ~$110B deals beginning after 2029, and has tested the market early: talks with Paramount/CBS (the league reportedly sought ≈+$1B a year) are “on pause”, and Fox’s Lachlan Murdoch said in August 2026 that Fox will not renegotiate before the opt-out. A holding pattern — with the cap’s next great jump waiting inside it. |
Per-network fees are the widely reported approximations — the league never publishes exact figures. The five 2021 packages (ESPN/ABC, Fox, CBS, NBC, Amazon) sum to $10.0B a season.
The Packers window
One bar per July. The Packers are the only club that publishes accounts, so their annual report’s national-distribution line — the equal share of the shared money every club receives — is the league’s only open window. Behind each bar, a green rule marks that calendar year’s salary cap.
national distribution per club (Packers report)that calendar year’s salary cap
* FY2017 is attested only as “roughly $255M”; the oft-cited $255.9M is unverified. “FY” is the season each report covers — the fiscal year ends the following March 31, and the report lands in July; FY2025’s arrived July 24, 2026.
The dare: find the fiscal year in which the equal national cheque failed to cover that season’s entire salary cap. There isn’t one — every blue column clears its green rule — and the gap only widens: by the FY2025 report the distribution is $453.2M per club against a $301.2M cap for 2026. Before a single ticket, beer or local sponsorship is sold, every club’s share of the national money more than pays its whole payroll ceiling, with roughly $150M left over. Note FY2020: gates collapsed ~92% in the COVID season and the national money still grew — the fact that explains why the cap fell only 8%.
Why does only Green Bay have to tell you any of this? Because the Packers are the NFL’s lone public corporation — no billionaire owner, ~537,000 shareholders’ worth of civic ownership, and therefore an annual report. FY2025’s showed total revenue of $753.0M, $299.8M of it local, and net income of $132.5M — but underneath, a $1.1M operating loss on rising player costs, rescued by $133.6M of non-operating income swollen by the club’s share of the NFL Network sale to ESPN, with a $701M reserve fund banked. CEO Ed Policy, on billionaire owners and private equity: “It’s like other teams have an ATM that we don’t have.”
The window, year by year
| FY | National per club | That year’s cap | Note |
|---|---|---|---|
| FY2015 | $222.6M | $143.28M | — |
| FY2016 | $244M | $155.27M | League total $7.808B — a record at the time. |
| FY2017 | ≈$255M | $167M | Attested only as “roughly $255M”; the oft-cited $255.9M precise figure is unverified — interpolate or footnote when charting. |
| FY2018 | $274.3M | $177.2M | Up 7.2% year over year. |
| FY2019 | $296M | $188.2M | About $9.5B league-wide. |
| FY2020 | $309.2M | $198.2M | The COVID season: gates collapsed ~92%, and the national money still GREW — the fact that explains why the cap fell only 8%. |
| FY2021 | $347.3M | $182.5M | — |
| FY2022 | $374.4M | $208.2M | About $12.0B league-wide. |
| FY2023 | $402.3M | $224.8M | — |
| FY2024 | $432.6M | $255.4M | $13.84B league-wide. |
| FY2025 | $453.2M | $279.2M | Reported July 24, 2026 (FY ended March 31): +4.8% to $453.2M per club, ≈$14.5B league-wide. Packers total revenue $753.0M, local $299.8M, net income $132.5M — but a $1.1M OPERATING loss on rising player costs, rescued by $133.6M of non-operating income boosted by the club’s share of the NFL Network sale to ESPN. Reserve fund: $701M. CEO Ed Policy, on billionaire owners and private equity: “It’s like other teams have an ATM that we don’t have.” |
Cash over cap
The cap is an accounting system; the budget is cash. Proration spreads a signing bonus across up to five league years on paper, but the wire transfer happens at signing — cash out the door years before its cap charges arrive — and the funding rule sharpens the point: the present value of what a club has guaranteed must be deposited into a segregated account when the promise is made, so a fully guaranteed contract is not a future problem but a present-day liquidity event. That is why the real constraint on a front office is rarely this year’s cap number, which restructures can manufacture room under, and almost always the owner’s cash budget: how large a cheque he will actually write this March, with the escrow behind it. What “guaranteed” means, vests as and forfeits back is the guarantees section’s subject.
Cash is also where the CBA sets its floors. League-wide, clubs must spend 95% of the caps in cash over each multi-year period (2021–23, 2024–26, 2027–30); each club individually must spend 90% over the same periods, any shortfall paid straight to that club’s players — the current period closes with the 2026 league year, shortfalls due by the following September 15, and the NFL pays if the club doesn’t. There is no per-year floor: a team may run one cheap season inside a period so long as the period’s cash total lands at 90%. The ceiling gets the headlines; the floors are why even the thriftiest club spends a quarter of a billion dollars a year on players across the current period.
The funding rule, as governance
The escrow requirement above is Article 26 §9, and it is short enough to read as four facts. Its mechanics — what it does to guarantee structures, and the collusion fight it fed — live in the guarantees section; as governance, the file reads:
- Cash behind the promiseArt. 26 §9
- The NFL may require each club to deposit into a segregated account the present value of the deferred and guaranteed compensation it owes, less a $15M deductible (2020–28 league years; $17M for 2029–30). Permissive on its face — “the NFL may require” — and in practice required.
- The 75% ceilingArt. 26 §9
- For guaranteed contracts, the amount a club must fund is capped at 75% of the total contract compensation — the rule stops short of demanding a full dollar-for-dollar escrow.
- Injury guarantees waitArt. 26 §9
- Future salary owed under an injury-only guarantee is not treated as owed — and so not funded — until the club acknowledges the player’s injury qualifies him for the payments.
- Escrow as excuseArt. 26 §9
- Agents and the union say the rule is wielded against fully guaranteed deals (“we’d have to put it all in escrow”); the union calls escrow an anachronism given owners’ liquidity. Colts owner Jim Irsay, March 2023, mid-Lamar-Jackson standoff: “I do not believe in fully guaranteed contracts… I just don’t see it as positive competitively.” The 2022 collusion grievance grew from exactly this ground.
The enforcement file
The cap polices paper; audits police reality. Compliance with the ceiling is checked when contracts are filed — what the league’s auditors hunt is the money that never touched the paper: side letters, undisclosed deferrals, cash stuffed into a year the rules weren’t watching. In three decades the file holds three major entries, every one final.
San Francisco 49ers
2000What was alleged
Undisclosed side agreements from the DeBartolo/Policy years — including a promised off-cap bonus to TE Brent Jones.
What was decided
Fines and stripped picks (a 2001 fifth-rounder and a 2002 third-rounder).
Club fine · $300kPolicy personal fine · $400kClark personal fine · $200kDraft picks stripped · 2case closed
The case that established personal liability: ex-president Carmen Policy was fined $400,000 and executive Dwight Clark $200,000 on top of the club’s $300,000 — side letters promising money the cap never saw. The pick-year split is occasionally reported reversed; the ESPN retrospective’s version is used here.
Denver Broncos
2004What was alleged
Roughly $29M of improper deferred, interest-bearing salary to John Elway and Terrell Davis in 1996–98 — effectively undisclosed loans shifting cap charges.
What was decided
Two sanctions for the same era: Dec 2001 ($968K + a 2002 third-rounder) and Sept 2004 ($950K + a 2005 third-rounder).
2001 fine · $968k2004 fine · $950kThird-round picks forfeited · 2case closed
Punished twice for the same title-era conduct as further violations surfaced — including a side deal not to waive a player. Both championship rosters were implicated, which is the “asterisk” jab Denver fans still hear from rivals.
Washington & Dallas
2012What was alleged
Front-loading contracts into the uncapped 2010 season “to gain an unfair competitive advantage.”
What was decided
Cap room stripped over 2012–13 at each club’s election; the $46M redistributed to other clubs.
Washington cap stripped · $36MDallas cap stripped · $10Mcase closed
The strangest case on the sheet: teams punished for spending in a year with no cap. Washington had taken Albert Haynesworth’s $21M option bonus entirely in 2010; Dallas paid Miles Austin a $17M base then restructured. Giants owner John Mara: the clubs “attempted to take advantage of a one-year loophole… quite frankly, I think they’re lucky they didn’t lose draft picks.” The redistribution pointedly excluded also-implicated Oakland and New Orleans — and the NFLPA answered with a May 2012 collusion claim alleging a secret $123M limit had governed the “uncapped” year.
The file, in words
- 2000, San Francisco: side letters from the DeBartolo/Policy years promised money the cap never saw; the club paid $300k, Carmen Policy $400k and Dwight Clark $200k personally — the case that established personal liability — and two draft picks, a 2001 fifth and a 2002 third, were stripped.
- 2004, Denver: roughly $29M of undisclosed deferred, interest-bearing pay to John Elway and Terrell Davis in 1996–98 — effectively off-books loans shifting cap charges — was sanctioned twice as violations kept surfacing: $968k and a third-rounder in 2001, $950k and another third in 2004, both title rosters implicated.
- 2012, Washington & Dallas: the two clubs front-loaded contracts into the uncapped 2010 season and were stripped of $36M and $10M of cap room over 2012–13, the $46M redistributed to their rivals — punished, as the NFLPA’s May 2012 collusion claim pointed out, for out-spending a secret $123M limit in a year that officially had no limit at all.
The rules above the cap
The cap governs what clubs may pay players; a second rulebook governs the clubs themselves — who may own one, with how much borrowed money, at what price, and with whose permission.
- Debt limits
- Standard club debt is capped at $700M; a new buyer gets an acquisition allowance on top, to $1.4B total (raised May 2024 from $1.2B, itself set October 2023). The cap on borrowing is the quiet enforcer of the all-cash club sale.
- The 30% rule
- A controlling owner must hold at least 30% of the club’s equity (family-aggregation exceptions exist; the league has studied relaxing it). At a ~$7B average valuation, the entry ticket to control is ~$2B of personal money.
- Private equity, on a leash
- Resolution JC-7 (August 27, 2024, passed 31–1 with the Bengals’ Mike Brown dissenting): approved funds — Ares, Arctos, Sixth Street, plus a Blackstone/Carlyle/CVC/Dynasty/Ludis consortium — may hold 3–10% of a club, passive and non-voting, at most six teams per fund, with a six-year minimum hold. Expected to deploy ~$12B.
- What a franchise is worth
- The 2025 lists average $7.1B (Forbes, +25% year over year) to $7.65B (CNBC) per club — cite as a ~$7.1–7.7B range, with the Cowboys on top around $12.8B. Estimates, not audited numbers.
- The $23B league
- League revenue cleared $23B for the 2024 fiscal year (Goodell’s stated target: $25B+ by 2027), split roughly 60/40 national/local — $13.84B of it shared national money. Of each revenue dollar, about 48 cents reaches players through the cap-and-benefits formula.
- Selling a team takes 24 votes
- Any transfer of a controlling interest needs an affirmative vote of three-quarters of the clubs — 24 of 32 — after Finance Committee vetting; the same supermajority governs debt-limit changes.
One quieter rule polices the rest of an owner’s portfolio: the league’s cross-ownership policy has long restricted a controlling owner from also owning major-league franchises in other NFL markets — relaxed only at the margins over the years, and one more reason the ownership class remains a closed club that admits members 24 votes at a time.
The 2030 horizon
The current agreement expires after the 2030 season, and that horizon is already inside contract design. Deals signed today reach past it — void years and prorated bonuses scheduled into league years nobody has bargained yet — and as expiry approaches the CBA’s endgame plumbing returns: the 30% rule, which caps how steeply a contract’s salaries may escalate into years beyond the deal, exists precisely to stop clubs stuffing money past a ceiling that might not be there. The last expiry wrote the syllabus: in uncapped 2010, spending did not explode — free agency tightened in compensation and most clubs behaved as if the cap still existed — and the two that treated the loophole as real were stripped of $46M of room in 2012 for “an unfair competitive advantage” in a year with, officially, no rule to break.
As of August 2026 no formal talks on a successor deal — or on an 18th game, which the CBA bars the league from adding unilaterally — had begun; the union elected JC Tretter its fifth executive director in March 2026. Until a new bargain exists, every restructure is a wager on caps that haven’t been negotiated and every void year a bill postdated into another CBA — and how those bills come due is the dead-money section’s whole subject.