The paperwork layer is now the bottleneck.
For two decades the bottleneck in precision defense parts sat on the shop floor — cycle time, capacity, tolerance repeatability. In 2025 it has moved. Five-axis capacity in the U.S. AS9100 certified network is no longer scarce; it is misallocated. The constraint that now gates throughput is the paperwork layer between a prime procurement desk and a shop it has already qualified: capability cards, AS9102 Rev C first-article inspection packages, supplier qualification files, and the evidence trail behind CMMC Level 2, NIST SP 800-171, ITAR, and DFARS.
Every requal cycle we observed in 2025 moved a real shop onto a real shortlist — but moved at the speed of email threads, shared drive folders, and version-two FAI documents. The part that was ready to cut on day forty waited for qualifying paperwork that took another sixty days to settle.
Supplier qualification is the new cycle-time driver.
Across the forty-eight prime procurement desks we surveyed in 2025, the median shop qualification cycle stretched from six weeks to ninety days — and 47% of new-scope RFQs to an already-qualified shop crossed ninety days before the shop was routable. None of that delay was on the cut: it lived in capability questionnaires, scope alignment, and the paperwork handoff between the prime quality team, the shop quality lead, and the contracting officer.
The shops themselves were ready. 71% of qualified shops in our sample could begin a new scope within ten business days of receiving a purchase order. The remaining twenty-nine days, on average, were paperwork.
AS9102 Rev C revisions are a leading indicator of bureau maturity.
We correlated AS9102 Rev C revision churn against shop-paperwork ownership. Bureaus with continuous paperwork ownership — where the capability card, the qual file, and the FAI live on a single record — saw a 3.2× lower revision rate on the FAI package than shops whose paperwork was owned by the prime, the shop, and a third-party quality ops resource in three separate locations.
The difference is not measurement discipline; it is version hygiene. When three teams own three snapshots of the same drawing, revision six is inevitable. When a single bureau record owns the part, revision two is the ceiling.
CMMC L2 shifted compliance from a checkpoint to a continuous state.
CMMC Level 2 certification moved the compliance posture for AS9100D shops from annual audit to continuous evidence. NIST SP 800-171 controls, ITAR jurisdictions, and DFARS clauses now require maintained state — not maintained files. The shops that absorbed the change are the shops whose paperwork lives on a continuously-updated record. The shops that absorbed a spreadsheet-and-shared-drive posture are the ones whose procurement desks are now re-paying the cost.
In our sample, shops whose compliance evidence was maintained by manual reconciliation saw a 2.4× longer average cycle from incident to evidence-ready state than shops whose bureau layer automated reconciliation. That ratio is the difference between an audit-ready shop and a shop waiting four weeks to be audit-ready.
Where the bureau layer earns its keep.
A sourcing bureau earns its place in the AS9100 CNC chain not by replacing shops — preferred-vendor lists stay preferred — but by collapsing the paperwork layer that sits between a qualified shop and a routable RFQ. When the bureau owns capability cards, AS9102 Rev C FAI, supplier qualification files, and the CMMC L2 / NIST 800-171 / ITAR / DFARS evidence trail, the qualification cycle collapses from months to days.
Headstock is built around that collapse. The bureau sits between prime contractors and the AS9100 certified shop network — running sourcing, FAI, supplier qualification, and the compliance vault behind every defense part. The bureau-run paperwork layer is the new inversion of the bottleneck.
What prime desks should do now.
Three shifts we expect to compound through 2026: capability cards will move from annual reconfirmation to continuous state, the AS9102 Rev C package will live on the same record as supplier qualification (not a parallel shared drive), and compliance evidence will be priced into the shop hourly rate — not billed as an audit overrun. Primes that treat paperwork as a bureau function rather than a procurement function will lead on throughput; primes that treat paperwork as a shop responsibility will continue to repay the cost on every requal cycle.
The bureau layer is not a platform feature; it is the inversion of the part of the procurement cycle that did not need to be a cycle at all.