The production stops. The obligation doesn't.

Tesla has put a public death date on the Cybertruck: production winds down by mid-2027. The coverage has been a product story – sales, styling, reputation. Fine. What nobody asked, anywhere between the investor deck and the news cycle, is the question that decides the next two decades: who writes the control plan for the last ten thousand units? In two decades of automotive and aerospace quality I have never seen anyone write it. Wind-down gets treated as a finance memo – provisions, severance schedule, a decommissioning cost centre. It is not. The most expensive quality phase of a programme often starts the day the end is announced, because nothing loosens a torque wrench like a death date.

Ramp-up has a manual. Ramp-down has a memo.

Launch quality is ritualised to the point of liturgy. APQP gives us phase gates. PPAP forces the process to prove itself before a saleable part ships. Capability studies, MSA, run-at-rate and safe-launch containment all exist because we know young processes fail young. I built a greenfield plant's QA/QC department from an empty office – 900-plus people, IATF 16949 from a standing start – and I wrote or signed off on most of the launch artefacts myself: control plans, PFMEAs, reaction plans, traceability matrices, escape-point analysis. That left me with an inconvenient piece of knowledge. I know exactly which documents a wind-down never creates.

No standard writes the exit. IATF 16949 assumes production continues; the APQP manual ends at ongoing production. So the exit gets improvised, and the artefacts that should exist do not. A decommissioning control plan. An end-of-production containment protocol – safe launch, run in reverse. A defined audit frequency for final builds, and a frozen-change board with authority to say no. A launch gets a phase-gate review. An exit gets a Tuesday email.

The last builds are the risky builds

Walk a dying line and you feel the risk before you can measure it. The best operators leave first – they have options – and their replacements inherit tribal knowledge nobody wrote down. Engineering changes freeze, which means every "temporary" deviation from the past two years quietly becomes the permanent final configuration. Suppliers deprioritise you in real time: your volumes are falling and so is their attention, and nobody funds a PPAP for a change on a part with eighteen months of life left. I have watched sub-tier substitutions slip into final builds precisely when attention had moved to the wind-down project plan.

The countermeasures are not exotic; they are launch countermeasures, reapplied. Double the layered process audits in the final months. Put full containment on the last lots and treat the final thirty days of production with the suspicion you applied to the first thirty. Then build exit traceability: every unit mapped to build date, shift, torque records and material lots, so that when a field failure surfaces in year ten, triage takes an afternoon instead of an archaeological dig.

At WITTE Automotive I cut failure costs substantially with QRQC, A3 and a Q-Wall – fast problem-solving at the source, with data. That reflex does not retire when the line does. It matters more, because the feedback loop that lets you fix things is dying at the exact moment your obligations are not. The standard is unchanged: the quarter with zero critical customer escalations is what a dying line must still deliver.

The fleet outlives the line

The line stops; the fleet does not. Hundreds of thousands of vehicles keep driving for two decades on whatever you bought, tooled and stored before the end. Service-part maths is where wind-downs are actually lost. Under-buy a bracket nobody can form anymore and you face tooling recreation at a multiple of the original cost, or low-volume remanufacture at unit costs nobody budgeted for. Unique panels and exotic materials are the worst case, and a stainless-steel exoskeleton qualifies on both counts.

Aerospace learned this decades ago and encoded it. Under AS9100 and EN 9100, configuration management and obsolescence planning are not optional softness; aircraft programmes routinely outlive their production lines by thirty years, so the configuration baseline stays alive, the tooling preservation plan stays funded, and suppliers owe discontinuation notices long after final assembly closes. Automotive should steal this wholesale.

Concretely: a tooling asset register with defined storage conditions, contractual end-of-life notification from every supplier, an obsolescence register reviewed annually against fleet age. Cheap now. Ruinous to reconstruct later.

The launch makes the promise; the wind-down is when someone finally audits it.

Give the exit its own 8D

Retirement is a problem definition, not an event, so run 8D logic on the retirement itself. Form the exit team while the line still runs. Define the problem in measurable terms: quality decay during wind-down, an unsupported fleet thereafter. Containment is the final-build audits. Root cause is the disengagement you can already predict. The corrective action is the decommissioning control plan. And define the exit criteria before the exit begins – audit performance, containment escapes, last-time-buy coverage. It is the last document anyone will write about the product.

The Cybertruck announcement is a product story for the business pages. For those of us who answer for field failures, it is the starting gun for the least-planned phase of the product lifecycle. The production stops; the obligation does not – and it will not for twenty years. Ramp-down deserves its own control plan, and the week a death date goes public is the week to write it. A launch is judged by its first parts. A programme is judged by how it leaves.