OSI Systems, Inc.

Rapiscan is one of only a handful of Western vendors qualified to supply certified, non-intrusive inspection (NII) screening systems to sovereign buyers — airport/aviation regulators, customs & border agencies (CBP/DHS), and critical-infrastructure operators. Screening is mandated (not discretionary) and equipment must pass multi-year national certification (TSA/ECAC-equivalent) before deployment; once a screening lane is qualified on a vendor's platform, replacement requires requalification, retraining, and a re-run of the procurement cycle. The moat is regulatory certification + installed-base service network, not raw detector physics.

X-ray transmission & backscatter, 3-D CT, radiation monitoring, metal detection, mm-wave imaging, and chemical trace detection (10-K business description); SaaS screening-adjudication platforms connecting US/EU/ME/LatAm government agencies to "screen millions of containers and vehicles" remotely (10-K). Competitive set is a Western oligopoly (Smiths Detection, Leidos Security & Detection, Rapiscan); Chinese Nuctech is effectively excluded from US critical infrastructure.

OSI's Digital Sovereignty claim is two-legged inside one Security P&L: (1) certified NII screening (Rapiscan) that sovereign buyers cannot swap without requalification, and (2) high-power RF transmit subsystems for homeland-defense OTHR that the 10-K now names as a Security product line next to inspection systems. The quantitative constraint that decides whether OSIS keeps the position is not detector physics — it is how fast funded task orders convert IDIQ/UCA ceilings into backlog, versus how fast the $800M international-contract tail and Mexico product mix roll off.

• NII qualification + installed-base service. Screening is mandated; equipment must pass multi-year national certification before a lane is live. 10-K business description: Security provides "security and inspection systems, turnkey security screening solutions, and high-power radio frequency systems." Competitive set in the 10-K performance-graph peer group is Leidos + (legacy) Smiths Detection, now being replaced by Evolv as Smiths Detection is being divested — i.e. the listed Western oligopoly is itself in flux.
• OTHR / Golden Dome RF. 10-K Item 1: "High-power over-the-horizon radar (OTHR) can provide 24/7 coverage of air, land and orbital space well outside the political borders of most countries. The U.S. has an initiative called Golden Dome…" April 23, 2026 IR: Security awarded a UCA with NTE ~$235M for production and integration of a homeland-defense OTHR transmit subsystem; $46M funded at award. Q4 call (Mehra): this is "our largest RF award to date"; OSIS is also "a participant in the Shield IDIQ which supports much of the Golden Dome related initiatives."
• Sole-source CBP IDIQs (post year-end). Two five-year sole-awardee IDIQs: ~$200M relocatable Rapiscan passenger-vehicle inspection (Aug 17) + ~$85M van-mounted mobile X-ray with ~$21M initial task order (Aug 19). Mehra on the Q4 call: "both the 200 million and the 85 million IDIQ with CBP is we are the only awardee." Ceiling ≠ backlog (Edrick): only firm delivery/task orders book.

A. OTHR UCA already in backlog (funded RF, not a ceiling).

• NTE $235M; Edrick: ~80% booked into Q3 FY2026 backlog = 0.80 × $235M = $188M.
• $188M ÷ FY2026 revenue $1,785.984M = 10.5% of annual revenue sitting in backlog.
• $188M ÷ Security external revenue $1,247.949M = 15.1% of the DS-thesis segment.
• $188M ÷ ~$1.9B consolidated backlog = 9.9% of total backlog.
• Delivery "over the next couple of years" (Edrick) → ~$94M/yr if linear ≈ 5.3% of FY2026 revenue per year — real, but a multi-year conversion, not a 2027 step-change.
• Cost of the moat vs the moat it defends (Security R&D): FY2026 Security R&D $55.021M ÷ Security revenue $1,247.949M = 4.41% annual R&D intensity to defend a $1.25B certified-screening + RF franchise. Consolidated R&D $79.141M ÷ $1,785.984M = 4.43%.

B. CBP IDIQ ceiling is not revenue (multi-vendor-IDIQ materiality inverted — these are sole-award).

• Combined ceiling $200M + $85M = $285M. Even if fully drawn over 5 years: $57M/yr ÷ $1,785.984M = 3.2% of annual revenue.
• Verified first task order: USAspending delivery order 70B03C26F00000549 under parent 70B03C26D00000031, obligated $21,210,786, FFP, signed Jul 30 2026, PoP through Jul 29 2027 (Miami-Dade). $21.2M ÷ FY2026 rev = 1.19%; ÷ Security = 1.70%.
• Mehra: some FY2027 contribution, "significant contributions… in fiscal 28 and beyond." Edrick: FY2027 guide includes "a portion of the delivery orders from CBP already in hand rather than the full ceiling."
• Analyst (Glynn) framing on the call: these two product vehicles sit inside ~$1B of NII funding that still has to be awarded across additional product types — OSIS is on the first two SKUs, not the whole billion.

C. What the replacement has to offset (the actual constraint).

• Three international contracts >$800M (FY23–24 awards): recognized $404M (FY24) → $231M (FY25) → $79M (FY26); majority of equipment already delivered; residual into FY2027 "at relatively lower amounts" (10-K MD&A). FY26 tail $79M = 6.3% of Security revenue.
• Mexico Security product mix: revenue from Mexico Security customers −$148.8M YoY in FY2026 (10-K MD&A) = 8.3% of consolidated revenue; Edrick: FY2027 Mexico headwind moderates to <$25M, concentrated 1H.
• Q4 $50M Middle East delivery slip = 10.3% of Q4 revenue ($50M / $484.058M); deferred, not cancelled; "substantial portion" now guided into 2H FY2027 (Edrick).
• Net: OTHR $188M already-in-backlog + CBP $21.2M firm TO replace a decaying $79M intl tail and a shrinking Mexico product book. The ceiling ($285M CBP + residual $47M unfunded OTHR) is optionality for FY28–30; FY2027 is a conversion year, which is why guide is only +5.0–8.1% despite "highest-ever" RF engagement.

D. Recurring-revenue mix is the margin path, not the bottleneck.

• Services $441.251M / $1,785.984M = 24.7% of revenue, +13% YoY; product GM 29.7% vs service GM 43.8% (10-K statements of operations). Edrick: "strong double digit" service growth in FY2027; FCF "could exceed 100% of net income."
• RPO $1.8B as of Jun 30 2026, 59% expected in next 12 months (10-K Note 5) = ~$1.06B near-term visibility vs FY2027 guide $1.875–1.930B — i.e. ~55–57% of next year's revenue is already contracted if RPO converts on schedule.
• Backlog coverage: $1.9B ÷ $1.786B = 1.06× (5-year definition; not a 1-year book). FY book-to-bill "a little bit north of 1"; Q4 just shy of 1 (Edrick).

• product_service: Security inspection screening systems (Rapiscan), patient monitoring/cardiology (Spacelabs), optoelectronic components and contract manufacturing. (10-K Item 1; accession UNKNOWN)
• technical function: X-ray/gamma/CT imaging, trace/radiation detection, threat-recognition software; patient vital-sign monitoring/cardiology; precision optical/electronic assemblies. (10-K Item 1; accession UNKNOWN)
• why it is hard: Regulatory certifications (TSA/DHS/EU), long qualification cycles, proprietary detection algorithms, global service network, and installed base create switching costs. (10-K Item 1A Risk Factors; accession UNKNOWN)
• substitutes: Manual inspection; competing certified screening/imaging vendors; alternative patient-monitoring incumbents; other contract manufacturers/EMS providers. (10-K Item 1A Risk Factors; accession UNKNOWN)
• pricing power: MED — Government contracts/installed base support pricing, but competitive tenders and hospital capital budgets limit. (10-K Item 1A Risk Factors; accession UNKNOWN)

INTACT — Security is a genuine DS critical-infrastructure toll-road (certification + installed-base moat, 70% revenue, record backlog/FCF), with the fresh CBP $200M/$85M IDIQs a strengthening signal. But the winding $800M international tail, Q4 timing drag, margin compression, and doubled leverage keep this a durable compounder (~5–8% growth) rather than a convex breakout. Research-only; no live capital.

INTACT — the omitted OTHR/Golden Dome/SHIELD leg is a real second DS bottleneck (10.5% of revenue already in backlog) and the CBP sole-award IDIQs confirm the US-border rotation, but both convert on a FY2028-weighted schedule while the $800M intl tail and Mexico product mix are still rolling off. That is why FY2027 is guided +5–8% on record backlog, not a convex breakout. Research-only; no live capital.

INTACT (unchanged) — the RF/Golden Dome leg now has a concrete, high-ROIC origin (~$76M → $235M UCA = 3.1×; 3.3× book-to-bill) and a $151B umbrella, while CBP 40–45% share and a clean insider read (grants only) remove two bear concerns. But it is an inorganic moat whose ceilings convert FY2028+, so the durable-compounder (~5–8% growth) characterization holds — not a convex breakout. Research-only; no live capital.