The composite passed inspection. The defect was cured in.

SkyDrive announced two composite aerostructure partnerships in the same week. Kineco brings composite manufacturing heritage from India; ITO brings precision machining credentials from Japan. Both press releases lean hard on precision, quality heritage, and certification readiness. I have read both. Neither mentions the autoclave.

I have qualified composite suppliers across multiple sites. The first thing I ask for is never the CMM report. It is the cure profile logs.

The autoclave decided before the CMM measured

Dimensional inspection confirms geometry. The part conforms to the CAD model within a tolerance band. For a machined aluminium bracket, that is sufficient — the material properties were set at the mill, and your job was to remove the right metal. Composites do not work that way.

The structural properties of a carbon-fibre laminate are created during layup and cure. By the time the part reaches the CMM, fibre volume fraction, void content, degree of cure, interlaminar bond quality, and fibre alignment are locked in. So is the answer to whether that part will delaminate under fatigue loading in service. The CMM can tell you the flange thickness is 2.4 mm ±0.1. It cannot tell you whether there is a disbond between plies 3 and 4.

The defect modes that ground composite aircraft — delamination, porosity, fibre waviness, incomplete cure, kiss-bonding — are baked in during the process. They are not dimensional. No amount of post-cure probing will catch them, because the damage is structural and lives inside the material.

This is why I verify autoclave profiles before I look at a single inspection report.

AS9100 qualifies the supplier. The process window qualifies the part.

There is a difference between QMS maturity and process maturity that most supplier audits never capture. A supplier can have pristine documentation, clean 8D records, and a well-maintained calibration system — and still produce a structurally marginal composite part because their autoclave control is sloppy.

On the shop floor, during supplier qualification, here is what I check before I review any inspection data:

• Prepreg out-time logs. How long did the material sit between freezer and layup? Has cumulative out-time been tracked across work shifts, or reset each morning?
• Layup environment. Temperature, humidity, contamination control. I have seen suppliers lay up prepreg in environments that would fail a paint booth audit.
• Autoclave cure profiles. Not the nominal programme — the actual recorded temperature and pressure curves from the last 20 cycles. I am looking for drift, thermocouple placement discipline, and deviations accepted without re-qualification.
• DSC results on cured coupons. Glass transition temperature consistency is the only honest confirmation that the cure actually happened the way the specification demanded.

At Airbus, the routing verification KPI work that cut internal lead time by 97% was rooted in this same principle: process discipline upstream eliminates the inspection burden downstream. We did not reduce lead time by inspecting faster. We reduced it by building verification into the process route — controlling the variables that determine outcome, so that the outcome stops being a discovery event.

That work was not composites. But the logic is identical, and in composites the stakes are higher because the failure modes are hidden inside the material.

The CMM market grows. The real failures stay invisible.

The coordinate measuring machine market is projected to reach roughly $11 billion. That number tells me the industry is investing in dimensional verification. Some of that investment is justified — composite aerostructures carry tight geometric tolerances, and automated measurement is the right tool for those.

But buying more CMMs does not improve composite structural quality. It improves composite geometric verification. Those are different objectives, and confusing them is expensive.

You can probe a surface to a micron and still miss the void that sinks the aircraft.

What catches composite defects before they reach service is a combination of in-process monitoring, non-destructive testing, and destructive validation — ultrasonic C-scan for disbond and delamination, thermography for near-surface anomalies, dielectric or fibre-optic cure monitoring, and destructive testing on witness coupons from every cure batch.

I built a quality system from scratch at the SNOP greenfield plant — 900+ employees, full programme. When you design a quality system on greenfield, you make resource allocation decisions that shape the plant for a decade. Every euro spent on a CMM is a euro not spent on a cure-monitoring system, an ultrasonic scanner, or a prepreg handling audit programme. The CMM is visible. It photographs well for investor presentations. The cure-monitoring system prevents structural failure. These are different investments with different payoffs, and only one of them keeps aircraft flying.

Suppliers that present dimensional inspection as their composite quality story are telling you something about their process confidence. If the process were controlled, the geometry would follow. The structural integrity — that is where the investment belongs.

eVTOL sharpens the edge

eVTOL makes this problem steeper because the product category is new and the certification framework is still maturing. Traditional aerospace composites benefit from decades of process baselines — autoclave cycles refined across thousands of parts, NDT methods validated against years of fleet data. eVTOL does not have that depth yet. Suppliers are building flight-critical composite structures against specifications that are still being written.

SkyDrive choosing Kineco and ITO is a sourcing decision. Whether those structures are safe is a process decision, and it will be made in the layup room and the autoclave, not at the inspection station.

In composites, the part is already finished when it leaves the autoclave. Everything after that is verification of what was already decided. The autoclave does not negotiate, and the defect that grounds the aircraft was not missed by the inspection — it was cured in.