The Supply-Chain Bill: 1,513 Missing Jets, and What the Filings Say They Cost

Airbus and Boeing have delivered 1,513 fewer aircraft since 2022 than their own 2018 output would have produced. The cost of that gap is not a single number — but it is itemised, in public, in the filings.

Between 2022 and 2025 — four years in which demand was not the problem — Airbus and Boeing delivered 1,513 fewer aircraft than they would have had each simply matched its own 2018 output. That is our arithmetic on the two manufacturers' own reported delivery figures, and the split is the part nobody says out loud: 84% of the shortfall is Boeing's.

The obvious next question is what that gap costs. The tempting answer is a single headline number, and you will see several in circulation. None of them survives contact with the filings, because the bill does not add up in the way that framing implies — it lands on different companies, in different currencies of pain, under accounting rules that do not permit summing. What the filings do give you is an itemised invoice. It is enormous, it is specific, and almost all of it is public.

How many jets are actually missing?

Counterfactuals invite cheating, so here is ours in full, with its weaknesses stated.

We take each manufacturer's 2018 output — Airbus 800, Boeing 806, the last year both were running clean — and ask what 2022–2025 would have produced at that rate. We deliberately exclude 2019–2021: Boeing's 2019 collapse was the MAX grounding, not the supply chain, and in 2020–21 demand genuinely vanished, so "they could have built more" is not a claim worth making. From 2022, demand was back and the constraint was physical.

Every delivery figure below is the manufacturer's own: Boeing's from the delivery tables in its Forms 10-K, Airbus's from its annual orders-and-deliveries releases.

Airbus is 245 jets light over four years — a real number, but a rounding error against Boeing's 1,268. Boeing delivered 600 aircraft in 2025; in 2018 it delivered 806. Seven years on, it remains 25% below its own high-water mark, while Airbus finished 2025 at 793 against a 2019 peak of 863 and has already guided to around 870 for 2026.

This is the correction the "industry-wide supply-chain crisis" framing needs. There is a genuine, universal supplier problem — engines, castings, structures — and we come to it below. But in pure delivery terms, five-sixths of the missing metal has one address.

Where did the chain actually break?

Two places, and both are now documented rather than rumoured.

The fuselage. Spirit AeroSystems built the 737's fuselage, and it was losing money doing it. Per its own SEC filings, Spirit lost $2,139.8m in 2024 and a further $1,968.2m in the first nine months of 2025 — about $4.11bn over 21 months — driven by forward-loss charges as production costs on 737, 787, A220 and A350 work ran past contracted prices. A supplier that loses money on every unit cannot fund a rate increase, which is why Boeing's ramp and Spirit's solvency became the same problem.

Boeing's answer was to buy it back, closing on 8 December 2025. The composition of the price, straight from Boeing's FY2025 10-K, is the single most eloquent line in this entire story:

Read the middle two rows again. Of the $8.37bn Boeing paid, $3.5bn was not buying a company — it was settling money Boeing had already advanced to keep its own supplier alive, and repaying that supplier's debt. Boeing's 10-K says it has been funding Spirit since 2023 to support "liquidity, rate readiness, and 787 tooling and capital expenditures." The vertical integration everyone reported as strategy was, in the ledger, a rescue that had already happened.

The engine. Pratt & Whitney's GTF powder-metal contamination — turbine and compressor discs made with contaminated powder between 2015 and 2021 — forced roughly 600–700 incremental engine shop visits between 2023 and 2026 and grounded a large slice of the GTF-powered A320neo fleet. RTX took an approximately $3bn pre-tax charge in Q3 2023, and here the crucial detail is the footnote: that figure represents Pratt's net 51% programme share. The gross bill across the PW1100 partnership is therefore on the order of $5.9bn — our arithmetic, not RTX's disclosure — for one engine variant.

It is not over. In its FY2025 results Airbus stated plainly that Pratt & Whitney's "failure to commit to the number of engines ordered" was hurting both its guidance and its ramp-up. An airframer publicly naming its engine supplier in a results statement is not routine; it is what happens when the problem outlives the polite phase.

So what does the bill actually say?

Here is what is disclosed, by whom, with the periods attached — and, critically, why you must not total this column.

Why these do not add. Spirit's losses partly become Boeing's after 8 December 2025, so counting both double-counts. The $3.5bn Boeing settled in the purchase price overlaps the funding that produced Spirit's reported losses. BCA's operating loss already contains the 777X and 767 reach-forward charges. RTX's charge is a different company in a different year on a different programme. Anyone handing you one tidy total has added at least two of these together.

What survives all of that is a floor, not a total: on the 777X and 767 alone, in two years, Boeing has told the SEC it expects to lose $9.36bn. The 777X launched in 2013 and now expects first delivery in 2027 — fourteen years, before a single revenue aircraft.

Who eats it?

Not evenly, and this is where the story connects to the rest of the market.

The OEMs eat the charges. That is what a reach-forward loss is: an admission that the programme will lose money over its accounting block, recognised now.

Airlines eat the delay, and get paid for some of it. Boeing's 10-K is explicit that customers hold "contractual remedies, including compensation for late deliveries or rights to reject individual airplane deliveries" — and that delays on the 737, 777X and 787 have already produced termination rights and compensation. The $443m of 737-9 customer considerations after the January 2024 grounding is what that looks like when it reaches the income statement. Airlines also adapt in ways no one invoices: SMBC Aviation Capital notes Frontier terminating 24 leases and deferring 69 Airbus deliveries — a carrier resizing itself around metal that will not arrive.

Lessors do not eat it. They collect. This is the part worth sitting with. The same delivery gap that produces Boeing's $9.36bn of losses is what pushed used-aircraft lease rates up 60–75% since 2021 — we traced that repricing through the lessors' own filings. The shortage is a cost to whoever must build or buy aircraft and a windfall to whoever already owns them. The bill is not paid into a void; a large part of it is a transfer.

Does it get fixed?

Partially, and slower than the targets say.

Boeing cleared the FAA's post-door-plug cap of 38 737s a month in October 2025, reached 42, and passed a rate-47 review in May 2026, with 53 as the year-end goal. Airbus is targeting 70–75 A320s a month by end-2027 — against an actual 2025 rate of 50.6, about two-thirds of the target. SMBC expects both manufacturers to lift output 12–15% in 2026 and exceed 1,500 aircraft.

The reason to discount all of it: an RBC supplier survey cited by SMBC has Boeing building 55 737s a month in 2028, against Boeing's own 63/month goal. The people who must actually make the parts are planning for less than the people selling the aircraft are promising. That gap — between OEM target and supplier plan — is the next instalment of this bill, already being written.

Why it matters

The backlog is the industry's most-quoted number and its least-priced one. Sixteen thousand undelivered jets reads as a demand triumph; the filings show what sits underneath it — a manufacturer 25% below its own 2018 output, a fuselage supplier that lost $4.11bn and had to be rescued by its customer, an engine programme whose fix cost its majority partner $3bn, and $9.36bn of losses booked on two Boeing programmes in two years. None of that is forecast or estimate. It is all disclosed, and almost none of it is ever added up — because, done honestly, it cannot be.

Sources & methodology

Delivery figures and our own arithmetic. Every delivery number in this piece is taken from the manufacturer's own reporting, not from our internal dataset.

Boeing — from the "deliveries, including intercompany deliveries" tables in Boeing's Forms 10-K: 2018 806, 2019 380 (FY2019 10-K); 2020 157, 2021 340, 2022 480 (FY2022 10-K); 2023 528, 2024 348, 2025 600 (FY2025 10-K).

Airbus — from Airbus's annual orders-and-deliveries releases: 2018 800 (A220 20, A320 Family 626, A330 49, A350 93, A380 12; release, 10 January 2019); 2022 661 (release, January 2023); 2023 735; 2024 766 (release, January 2025); 2025 793 (FY2025 results).

The 1,513-jet shortfall is our calculation: 2018 output (Airbus 800, Boeing 806) held flat across 2022–25 (Airbus 3,200, Boeing 3,224) against actual deliveries (Airbus 2,955, Boeing 1,956). We exclude 2019–21 deliberately — the MAX grounding and the COVID demand collapse are not supply-chain effects, and including them would inflate the gap to 3,414 jets on a basis we do not consider defensible. The implied ~$5.9bn gross GTF programme cost is our arithmetic on RTX's disclosed ~$3bn charge at Pratt's stated net 51% share; RTX does not publish that gross figure in the release cited.

Boeing's two sets of books, disclosed. This analysis first returned 1,510, not 1,513, and running down the three-jet difference turned up something worth stating plainly: Boeing's own two public delivery records do not agree.

The delivery tables in Boeing's Forms 10-K — audited and filed with the SEC — report 806 deliveries in 2018, 380 in 2019, 340 in 2021 and 480 in 2022. Boeing's public orders-and-deliveries database, the commercial record it publishes and updates, yields 805, 379, 338 and 479 for the same years. For 2023, 2024 and 2025 the two agree exactly (528, 348, 600), as does every Airbus year against Airbus's own releases. The gaps are small — one or two aircraft — but they are real, and they sit in the 737 line.

We have not been able to establish the cause and do not speculate; the most it is safe to say is that a 10-K is frozen at filing while a live commercial database can be restated. We have published the 10-K figures throughout, on the principle that an audited SEC filing outranks a marketing-side database. Readers preferring the commercial record should read 1,510 for 1,513 and 1,265 for 1,268 — the finding does not move.

We note this rather than bury it because our own delivery dataset is built from Boeing's public database and reproduces it faithfully, including these figures. That is not an error in our pipeline; it is a discrepancy at the source, and anyone citing "Boeing delivered N aircraft in 2018" should know which N they have. It is logged permanently on our corrections and data notes page. See also Orders & Deliveries and The 12-Year Wait.

Boeing. The Boeing Company, Annual Report on Form 10-K for the fiscal year ended 31 December 2025, filed with the SEC (filing) — 777X additional reach-forward losses of $4.9bn (2025) and $3.5bn (2024), programme launched 2013 with first delivery expected 2027; combined 777X and 767 reach-forward losses of $5,283m (2025) and $4,079m (2024); BCA loss from operations $7,079m (2025), $7,969m (2024), $1,635m (2023); $443m of 737-9 customer considerations related to the January 2024 grounding; customers' contractual remedies for late delivery; Spirit Acquisition completed 8 December 2025 with total consideration of $8,371m (stock $4,704m; settlement of loans, advances and other payments $2,571m; debt repaid on Spirit's behalf $948m; premium on assumed exchangeable notes $109m; share-based awards $39m); Boeing funding to Spirit since 2023 for liquidity, rate readiness and 787 tooling.

Spirit AeroSystems. Net loss figures taken directly from the SEC's XBRL company-concept API for CIK 1364885, tag `us-gaap:NetIncomeLoss`: FY2024 −$2,139.8m (Form 10-K, filed 28 February 2025); nine months to 2 October 2025 −$1,968.2m and Q3 2025 −$724.3m (Form 10-Q, filed 31 October 2025).

RTX / Pratt & Whitney. RTX Corporation, "RTX provides update on Pratt & Whitney GTF fleet," 11 September 2023 (release) — approximately $3bn pre-tax charge in Q3 2023 reflecting Pratt & Whitney's net 51% programme share of the PW1100 GTF programme; RTX pre-tax operating profit impact estimated at $3–3.5bn over several years; approximately 600–700 engines removed for shop visits between 2023 and 2026; elevated AOG expected through 2024–26.

Airbus. Airbus SE, "Airbus reports Full-Year (FY) 2025 results," February 2026 (release) — 793 commercial deliveries (A220 93, A320 Family 607, A330 36, A350 57); revenues €73.4bn; EBIT Adjusted €7.128bn; FCF before customer financing €4.574bn; record backlog of 8,754 aircraft; 2026 guidance of around 870 deliveries; A320 rate target of 70–75 a month by end-2027 stabilising at 75; and the statement attributing damage to its guidance and ramp-up to Pratt & Whitney's "failure to commit to the number of engines ordered."

Market context. Shane Matthews, Darren Naughton and David Griffin, Push and Pull Factors on Aircraft Lease Rates 2026, SMBC Aviation Capital, March 2026 (PDF) — 2026 output expectation of 12–15% growth and over 1,500 aircraft; the RBC supplier survey (Boeing 55/month in 2028 against a 63/month goal); Frontier terminating 24 leases and deferring 69 Airbus deliveries. SMBC is a lessor and an interested party; see the disclosure in our lease-rate piece.

Boeing 737 production rates. The FAA capped 737 output at 38 a month following the January 2024 door-plug accident; the cap was lifted in October 2025, with output reaching 42 a month through Q1 2026, a rate-47 review passed in May 2026 and a 53/month year-end 2026 goal. Rate and cap history is corroborated by Boeing's Form 10-Q for the quarter ended 31 March 2026 and contemporaneous reporting; the 10-K confirms the programme may only raise rates with FAA concurrence.

What we did not publish. Our own topic brief carried an unsourced "$11bn-class" figure for the total supply-chain bill. We could not trace it to any filing, regulator or company disclosure, and it is omitted rather than repeated — the itemised table above is what the primary documents actually support. We also make no claim about the total number of GTF-powered aircraft currently grounded: the figures in circulation (commonly around 600–720) trace to secondary aviation press and consultancy commentary rather than to a disclosure by RTX or Airbus that we could verify, so only RTX's own shop-visit range is cited. Finally, the "at 2018 output" baseline is a counterfactual, not a forecast: it measures the two firms against their own demonstrated capability, and nothing more.