- GE cut shop visit turnaround by about a week and logged a ninth straight quarter of supplier input growth, while past-due spare parts grew 20% sequentially.
- Turnaround and output measure departure rates. Past-due backlog measures arrivals minus departures, and orders rose over 30% while shipments rose 25%.
- Pratt routed material to its MRO shops and took weaker equipment sales for it, so the split between new build and aftermarket is a live choice.
The monthly operations review runs green down the page. Turnaround time down. Output at a record. Supplier receipts up again. Then someone reads out a note from a customer whose parts were promised in April and still have not shipped, and there is no line on the scorecard that accounts for it.
GE Aerospace held that review in public on 16 July. Larry Culp told analysts the company had "improved total shop visit turnaround times by about a week since the end of 2025", that LEAP turnaround times were "now around 100 days, down over two weeks year over year", and that he was "thrilled to see not only nine consecutive quarters with increases from our critical suppliers", with priority supplier material input "increasing double digits sequentially and year over year again in the second quarter". Grounded LEAP-powered aircraft were at "nearly zero". Internal shop visit revenue grew 25 per cent. Spare parts sales rose more than 25 per cent, and CFO Rahul Ghai credited that to "improved material availability".
Then Ghai read out the other number. "Even with strong revenue growth, given robust orders, spare parts delinquency, which represents shipments that have been delayed due to material availability constraints, grew 20% sequentially in the second quarter." Culp came back to it later without an analyst asking: "As much as we're pleased with the delivery increases, our spare parts delinquencies are up, unfortunately."
That clause of Ghai's is the entire published definition of the metric. It behaves as a level rather than a rate. In April Culp called it "that overdue, that delinquent backlog" and said it was going to "take us a while yet here to get to zero delinquency", and the same call put delinquency "up roughly 70% since the end of '24". Compounding April's level by June's 20 per cent puts the backlog at roughly twice its end-2024 size. That multiplication is an inference drawn from two published growth rates rather than a figure GE reports. The company has never disclosed the absolute level, the units, or an aging profile in any period, and while a delinquency chart appeared on slide 5 of the first-quarter deck, none appears in the second-quarter deck. The number now lives only in the spoken remarks.
Every green metric in that first list measures a departure rate, meaning how fast work leaves a station once it has arrived. Past-due backlog measures something different: arrivals minus departures, accumulated and never reset. In April Culp said spare parts orders were "up over 30% year-over-year" since the beginning of March. Shipments grew 25 per cent. Arrivals grew faster than departures, so the queue lengthened, and not one of the rate metrics is wrong or even flattering. Asked in April why the chart kept climbing, Culp gave the mechanism plainly: it was "just a function of demand outstripping supply". He added, "Operationally, it is a number we are not proud of, right? Because we are holding, we're failing to meet customer expectations in that regard."
The simple interpretation is that the engine makers are feeding new build at the aftermarket's expense. The evidence points the other way. Pratt and Whitney made the opposite call and said so on 23 July, when Chris Calio explained weaker equipment sales as "mix as, again, we continue to ensure that we have strong material flow into our MRO shops", with the priority being "making sure that they have the material that they need in order to do what they need to do". Pratt reported PW1100G-JM output up 43 per cent year on year, enabled by a 23 per cent cut in turnaround time. Further upstream, Howmet's total spares revenue rose 37 per cent against 28 per cent growth in commercial aerospace, at a 32.1 per cent EBITDA margin, while John Plant described having "one casting pit left" and "a 3-year lead time for some of that very exotic equipment". Nobody in that chain is rationing spares to protect new engines. The capacity is sold, and the clock on adding more of it runs to 2028.
So far the queue is commercial rather than physical. Airbus said on 29 July that it had "no gliders in the sense of aircraft not being delivered solely because of engines", and GE puts engine-driven groundings near zero. The queue is still growing, though, and Ghai gave the figure that shows where it sits: "Engines already off-wing and the pipeline of planned removals in the third quarter exceed our full-year shop visit guide by over 40%." Those removals have already happened or are already booked, ahead of the capacity GE guided to for the year.
Two things worth checking. First, what the on-time number is measured against. If the clock resets every time a date is re-promised, on-time delivery improves each time a job slips, and the wait the customer actually experiences never appears anywhere. Second, whether anything on the scorecard is a level at all. Rates cannot show a growing queue, so carry past-due as an absolute, with an aging profile, beside the turnaround chart. When every rate improves and the level keeps climbing, what changed is the arrival side, and that is a capacity conversation rather than a shop floor one.
