The hard cap

$301.2M per club — the first cap above $300M, announced on February 27, 2026, a $22M rise on 2025’s $279.2M. It is the ceiling on what each of the 32 teams may commit to player salary under the cap’s accounting; a further $77.6M per club of league-funded benefits sits on top of it, for $378.8M of total player spending. The number derives from the CBA’s revenue split — roughly 48 per cent of All Revenue, less projected benefits, divided by 32 — which is why every new television deal moves it. Big as the jump was, it was not the record: 2024’s rise of $30.6M remains the largest one-year increase, part of the surge that has carried the cap from $182.5M to $301.2M in the five seasons since the COVID decline.

Yes — the hardest in American sport. The NBA’s cap is soft, riddled with exceptions that let teams spend past it, and MLB has no cap at all, only a luxury tax on payrolls beyond a threshold; the NFL’s is a ceiling no club may exceed for a single day of the league year, with no tax to pay one’s way through. It is also a floor: the league as a whole must spend 95 per cent of the caps in cash over each multi-year period, and each club at least 90 per cent, with shortfalls paid directly to the players. What softens it in practice is timing, not exemption — proration, restructures and void years move charges between league years without ever shrinking them. And it is enforced: the clubs that front-loaded contracts into the uncapped 2010 season were stripped of $46M of cap room between them in 2012.

Almost never — and ‘guaranteed’ itself is a term of art. NFL guarantees come in three prongs — skill, injury and salary-cap termination — and money is ‘fully guaranteed’ only when protected against all three; much of it is guaranteed for injury only at signing, vesting to full protection on a later date that a club can beat by releasing the player first. Kirk Cousins’s three-year, $84M Vikings deal of 2018 was the first fully guaranteed multi-year veteran contract, and Deshaun Watson’s five-year, $230M Browns deal of 2022 remains the largest full guarantee at signing in league history — an outlier no club has repeated. The structural brake is the funding rule, Art. 26 §9 of the CBA: the league may require clubs to escrow the present value of guaranteed and deferred compensation, less a $15M deductible, which owners cite against writing nine-figure promises. In 2025 an arbitrator found the league had encouraged clubs to limit guarantees after the Watson deal — while ruling the union had not proven the clubs colluded in response, a finding upheld on appeal in April 2026.

Cap charges for a player no longer on the roster. When a player is released or traded, the unamortized remainder of his signing-bonus proration accelerates onto the cap at once, joined by any still-unpaid guaranteed salary. It is not a fine but the schedule catching up: every dollar of cash a team pays eventually hits its cap, and dead money is the bill for the dollars the books had not yet charged. The record is Tua Tagovailoa’s $99.2M, set when Miami released him in March 2026 with a June 1 designation that split the charge $55.4M onto 2026 and $43.8M onto 2027 — surpassing Russell Wilson’s $85M of 2024–25. In this guide’s fictional $150M anchor deal, a release the day after signing costs exactly $78M, the fully-guaranteed-at-signing figure to the dollar — which is the point: dead money is the guarantee arriving on schedule.

A contract year that exists only on the books: the player is never under contract to play it, but it counts as a year for signing-bonus proration, so a bonus can spread over up to five league years even on a shorter deal. The trick lowers today’s cap charges by parking them in seasons that will never be played — and when the contract voids, every allocation parked in the void years accelerates onto that league year’s cap at once (Art. 13 §6(b)(ii)(4)). Tom Brady’s exit is the canonical case: his restructures had pushed proration into void years, and his 2023 retirement landed $35.1M on Tampa Bay’s cap for a player no longer in football. Practiced at scale it compounds — New Orleans, the technique’s most famous devotee, carries $112.15M of dead money on its 2026 cap, about 37 per cent of the ceiling, most of it void-year proration coming due.

It decides which season pays the bill. A player cut on or before June 1 sends all his remaining bonus proration onto the current year’s cap at once; cut after June 1, the current year keeps only its scheduled slice and the remainder lands on the next league year’s cap. Because rosters are built in March, the CBA lets each club designate up to two terminations a year as post-June-1 cuts in advance — but the designation exists for terminations only, never trades, and the cap relief does not arrive until June 2, months after the best free agents are gone. Arizona’s 2026 release of Kyler Murray shows the fine print: $47.5M of his $54.7M charge stayed on the 2026 cap anyway, and the roughly $5.1M of relief arrived only in June. Trades can reach the same split only by physically executing after June 1 — which is why Cleveland processed the Myles Garrett trade at 4:01 p.m. on June 1, 2026, turning a $41.09M single-year bill into $15.53M now and $25.56M next year.

Both are one-year tenders a club may apply to a pending free agent — one designation per club per year, franchise or transition, never both. The non-exclusive franchise tag pays the greater of the position’s five-year cap-percentage average or 120 per cent of the player’s prior salary — $43.895M for a quarterback in 2026 — and carries the sport’s heaviest deterrent: another club may sign the player to an offer sheet, but an unmatched one costs two first-round picks. The transition tag is cheaper — computed over the ten largest prior-year salaries rather than the five previous tags, $37.833M for a 2026 quarterback — and buys only a five-day right to match, with no compensation if the club declines: the gap Minnesota drove a poison-pill offer sheet through to take Steve Hutchinson from Seattle in 2006. An accepted tender of either kind is fully guaranteed, and July 15 is the deadline to convert a tag into a multi-year deal. In 2026 four players were tagged — Breece Hall, Kyle Pitts and George Pickens on franchise tags, Daniel Jones on the transition tag — and only Pickens is playing the season on his tender.

At the end of each free-agency cycle the league counts every club’s qualifying unrestricted free agents lost against those signed, and only net losers collect: each net loss returns one extra draft pick, placed at the end of rounds 3 through 7 by where the departed player’s pay, snaps and honors rank him against the whole league — a loss in the top 5 per cent of players returns a third-rounder, down to the top 35 per cent returning a seventh — with a maximum of four picks per club. The cascade has sharp edges: players cut by their old club never count, tag departures are excluded, and minimum-salary veteran signings don’t cancel a club’s losses, which is the loophole disciplined front offices shop through. The NFL has never published the formula — the round bands are CBA text (Appendix V), the dollar thresholds are OverTheCap’s reverse-engineering, and the league simply announces the results each spring. For the 2026 draft that meant 33 picks to 15 teams, with the Ravens, Eagles and Steelers at the four-pick maximum: 32 formula picks plus one special selection, the Rooney-rule pick Detroit earned when the Jets hired Aaron Glenn.