Short answer

X-ray CT is often the binding method for complex internal geometry, but CT is not a universal filter: detectability depends on material, thickness, defect morphology, scan/reconstruction settings and validated POD. Monitoring does not replace CT/DT by default.

A finished build is not a released part. Release waits on inspectability, disposition rules and queue time.

1. The bottleneck

Brief #1 framed the industrial metric correctly:

Accepted output ≈ print capacity × availability × print yield × downstream yield × release rate.

This brief isolates bottleneck #5 on the Addithive Atlas: Inspection, NDT and acceptance criteria. After print—and often after stress relief, HIP, machining and cleaning—the route still stops if there is no validated method, no accepted disposition criteria, or approved CT/NDT capacity is scarce relative to build volume.

Failure mode	What breaks	Evidence to track
Method gap	Critical defects not covered	Inspection plan vs FMEA; POD status
Criteria gap	Findings without accept/reject rules	Open MRBs; disposition backlog
Capacity gap	Queue dominates	CT turnaround; scanner utilization
False-call / over-reject	Throughput collapses	False-call rate; repeat scans
Substitution overclaim	Monitoring treated as CT replacement	Approved substitution list

Standing refusal: Voxel size is not the minimum detectable defect size. A successful demo scan is not serial inspection readiness.

2. The evidence

Standards and qualification

Source	What it anchors	Boundary
ISO/ASTM 52908:2023	Post-processing, inspection and testing for metal PBF	Framework ≠ part-specific criteria
ISO/ASTM 52920:2023	Site QA measures across industrial AM	Site QA ≠ flight/clinical approval
NASA-STD-6030 / 6033	Flight hardware + facility control (6033 revalidated Jan 2026)	Agency-specific mental model

Technical evidence

• CT is uniquely useful for 3D internal inspection of complex AM geometry; it is not universally superior to UT, PT, ET or radiography.
• Recent literature supports the detectability boundary: Sun et al. (2024) review XCT across prevention, diagnosis and failure prediction in metal AM; Baig et al. (2025) show XCT performance deteriorates more for irregular lack-of-fusion defects than for near-spherical pores, with reliable capture often needing features several times larger than a voxel. Separately, a 2024 Ti-6Al-4V POD study found high detection rates can require flaw sizes substantially larger than nominal voxel size.

Decision rule: describe detectability with material, thickness, morphology/orientation, scan/reconstruction settings and POD basis—or do not claim a detection limit.

Monitoring does not erase the release gate

NIST framing and Addithive’s substitution brief converge: telemetry ≠ sensing ≠ post-process NDT/CT ≠ destructive testing. A melt-pool image is not a pore measurement. Substitution needs measurand definition, registration, ground truth, POD, false-call behavior, robustness, locked analytics and approval—layered evidence, not automatic elimination of post-build inspection.

Industrial signals (commercial, labeled)

Signal	Status	Why it maps here
Bodycote U.S. east-network HIP/HT/wire EDM for printed metal	~21 Jul 2026	Downstream capacity; inspection release remains separate
Hexagon to acquire Waygate (~$1.45B; 2D X-ray / 3D CT, RVI, UT)	Announced 13 Apr 2026; still pending as of 12 Sep 2026 — expected close H2 2026; no completion announced	Capital to NDT hardware + CT stack beside metrology/CT analysis
America Makes JAQS-SQ; U.S.–UK AAMI	2026	NDT, acceptance criteria, portable evidence as scale constraints

Commercial signal ≠ proof that any vendor resolves AM CT throughput industry-wide.

3. The exposed companies

Mapping only—not investment advice. AM materiality is often emerging or immaterial at group level.

Role	Public exposures (map)	Analyze (not “buy”)
CT / NDT systems	Hexagon (VG; pending Waygate), Nikon, ZEISS; Waygate/Baker Hughes until close	Capacity, AM apps, software attach, integration risk
Metrology / simulation adjacent	Hexagon, AMETEK (FARO/Creaform), Renishaw; Simufact / Ansys / Siemens / Dassault	Whether dimensional or predicted risk sits on the release path
Depends on CT to ship / densification	Aerospace/defense adopters; ATI; Oerlikon; Bodycote (HIP)	Internal-channel families; CT vs print queue; HIP ≠ inspection release

4. The decision

What changed

• Inspection/NDT is the explicit release-rate limiter for complex accepted-part output.
• Hexagon’s Waygate deal elevates NDT/CT strategically (close still pending as of 12 Sep 2026; H2 2026 target unchanged).
• Defense work treats NDT + acceptance criteria + portable evidence as infrastructure.

What remains unproven

• Whether automated CT review materially raises accepted-part throughput without missed-defect risk in regulated apps.
• Whether in-situ monitoring wins approved substitution on flight/clinical hardware at scale.
• Whether NDT M&A shortens customer queue time or mainly consolidates share.

What to watch

1. CT turnaround/utilization at qualified labs (not demo times).
2. POD / locked acceptance criteria for LPBF Ti and Ni.
3. Count of approved monitoring substitutions.
4. False-call and MRB rates after CT.
5. Waygate close and any disclosed aerospace/AM CT backlog (H2 2026).
6. Accepted-part yield vs build-success rate.

Decision rule

Before buying another printer, ask: critical defect population? Validated NDT/CT + POD? Locked acceptance criteria? Approved inspection capacity vs build volume? Approved monitoring substitution—or only complementary evidence? If the queue sits at CT/NDT disposition, more print capacity increases WIP, not shipments.

Sources

• Addithive: Brief #1; CT vs NDT; Monitoring substitution; Aerospace qualification; Bottleneck companies
• ISO/ASTM 52908:2023 (current ed., iso.org stage 60.60); ISO/ASTM 52920:2023; NASA-STD-6030 / 6033
• Sun et al. (2024), Thin-Walled Structures — XCT metal AM review; Baig et al. (2025), Scientific Reports 15:6740 — irregular LoF detectability; 2024 Ti-6Al-4V POD literature — high detection often requires flaws ≫ voxel
• Hexagon / Baker Hughes Waygate announcement 13 Apr 2026; re-checked 12 Sep 2026 — close still pending (H2 2026)
• Bodycote eastern U.S. AM post-processing (Jul 2026); America Makes JAQS-SQ / AAMI (2026)