Canberra Pays to Fix an Autopilot Whose Code It Does Not Hold

Analysis as of 14 August 2026.

Among the contract actions the United States Department of War published on 31 July 2026, one small order carries an outsized structural lesson. Boeing received a cost-plus-fixed-fee order, N0001926F1107, worth $11,381,402, to develop, test and deliver loadable flight-control software addressing six named deficiencies in the P-8A's autopilot: altitude-acquire behaviour, a placard-speed exceedance, pitch oscillations in climb and descent, an uncommanded switch from altitude-acquire to vertical-speed mode, sluggish autothrottle response, and bank-limit-margin tuning. The whole sum comes from Royal Australian Air Force funds; the work happens at Tukwila, Washington, under NAVAIR administration, without competition, with completion expected in February 2028. The official wording promises only that the updates should improve or resolve the listed behaviours — not that each will disappear.

The order is a rounding error against its neighbours. Under the same basic ordering agreement, a 22 June 2026 order for nine retrofit A-kits came in at $121,195,041 and a training-systems vehicle announced four days earlier at $880,000,000; back in 2016 the Defense Security Cooperation Agency priced the non-recurring-cost recoupment for one P-8A at $63,720,841 — nearly six times this entire software effort. On the Australian side, the AIR 7000 project carries approved equipment-acquisition spending of A$7.409 billion, of which A$5.580 billion had been consumed by 30 June 2026, and the country only completed its fourteen-aircraft fleet on 27 May 2026, when the final Poseidon landed at RAAF Base Edinburgh. An eleven-million-dollar line does not even surface individually in the budget papers. Its interest lies elsewhere: it is the point where national money, foreign contract administration and shared design authority visibly intersect.

That intersection is crowded with separate authorities. Boeing holds the industrial design capability but cannot authorise military flight with the modified load. NAVAIR issues and administers the order; the US airworthiness system will judge the design inside its own framework, currently complicated by a version-control wrinkle in which the governing instruction cites one revision of the certification handbook while the active edition, dated 29 May 2026, is another. In Australia, the Surveillance and Response Systems Program Office holds the Military Type Certificate — AUS.DASA.MTC.A47 — and the Defence Aviation Safety Authority must classify the change as major or minor before approving it, with a lawful route to reuse American evidence where configuration, role and environment align. One technical fix can therefore travel through two distinct approval tracks, acquire national limitations, and enter service on different dates in different fleets without the codebase ever formally forking.

What the notice conspicuously withholds is the ownership question. Paying the full amount gives Australia no disclosed possession of source code, no unlimited licence, no authority over the master baseline and no power to make the US Navy adopt the result. Nor does any public document show that the other operators — the United States, the United Kingdom, Norway, New Zealand, Korea or Germany — even experience the same six behaviours, so the premise that this is a fleet-wide problem awaiting a common cure is itself unproven. The history counsels patience about closure, too: Australia declared its Operational Capability 2 milestone in February 2019 with two caveats, and its audit office noted that the department had never defined what a caveat or a deficiency formally means.

The single question that organises everything downstream of this award is whether an allied nation that finances a correction to a jointly operated design can convert its money into durable programme influence — whether, that is, the software, the test evidence and the configuration decision bought with Australian funds will flow back into the common baseline as assets every operator can draw on, or whether they will harden into a national release with its own certification artefacts and operating limits, leaving Canberra to fund, one order at a time, improvements to an aircraft whose code, baseline and release calendar remain in other hands?