Naval-nuclear demand accelerates → BWXT is the priority allocation for scarce inputs → but revenue converts at legacy contract prices → so the margin inflection is dated to the backlog roll-off, not to the demand

BWXT's revenue grew 18% YoY while adjusted EBITDA grew only 7%, and management names the reason — a backlog priced under older customer arrangements that the CFO "fully expect[s]…to be done by the end of 2026" — which reframes the book's second-worst live position from a demand problem into a **dated contract-vintage problem**.

• **Q3 or Q4 FY2026 government-segment margin does *not* progress toward 20.5%** — machine-checkable against the guide. If the segment margin prints below ~20% in Q4 with revenue still growing, step 4 fails and step 3's explanation is probably wrong too.
• **Management re-dates the roll-off** past end-2026 on a subsequent call. A slipping date on a contractual, already-signed backlog would suggest the explanation was a narrative rather than a schedule.
• **The wedge persists into 2027**: revenue growth continuing to outrun EBITDA growth by ~10 points after the stated roll-off completes would falsify the contract-vintage explanation entirely and point at cost or competitive pressure instead.
• **ATI stops prioritizing naval nuclear** (specialty energy shipments recover while naval-nuclear lead times compress) — would remove the step-2 allocation privilege and expose BWXT to genuine input competition.

• **Step 2's evidence cuts against a sibling chain.** russ-jevnin reporting easy access to large forgings is *counter-evidence* to any reading of shared-heavy-forging-capacity-to-specialty-alloy-melt-rent that treats heavy forging as an absolute economy-wide shortage. The reconciliation (allocation, not absolute scarcity) is coherent — but it is a reconciliation, and it should be held as such rather than as confirmation.
• **The commercial-segment margin guide was *cut*** (14% → 13%) in the same quarter the government guide was raised. Two segments moving opposite ways means the consolidated print can stay flat even if this chain is right.
• **The market disagreed on the day**: shares fell 3.23% after hours on a raised guide, which the source attributes to "higher spending plans" and capacity-build dilution. Either the market is discounting the roll-off, or it is pricing the capex.
• **Single-source on the load-bearing quote.** The legacy-backlog statement comes from one call, via a `partial` transcript (investing.com extraction, no verbatim full text). No second source has yet corroborated the end-2026 date.
• **No BWXT transcript existed in this wiki before today**, and BWXT is not on `EARNINGS.md`. This chain rests on a source that was pulled ad-hoc — the same gap flagged on 2026-08-11 for HWM and ATI.

• **BWXT is the tradeable.** The specific asymmetry: a live position sitting at −16.6% sector-excess whose underperformance the wiki had no explanation for now has one, and it is self-limiting with a stated end-date.
• **This reframes a ledger flag rather than opening a new theme.** Per the 2026-08-11 4e read, BWXT was the book's second-worst live position and the wiki's working hypothesis was that naval-nuclear demand might be softer than assumed. The ATI $1B contract (doubling) and BWXT's own +18% revenue both refute that. The lag is priced backlog.
• **A general lens worth reusing:** when a supplier's revenue growth materially outruns its EBITDA growth with no cost shock disclosed, check contract vintage before concluding the demand or the pricing power is weak. This is the mirror image of lta-contract-structure-as-price-insulation — the same contract structure that insulates a buyer from a spike also strands a supplier at stale prices.
• **Commercial nuclear is optionality, not the thesis.** A "credible opportunity" for one new-build order before year-end is real but undated and unsigned; the government segment is ~80%+ of the story.