Setup

Arcosa (ACA, ≈$6.3B mkt cap) divested its barge business April 1, 2026 for ≈$450M and is now a two-segment infrastructure pure-play: Construction Products and Engineered Structures (utility poles, wind towers). The Q1 2026 10-Q filed May 1 is the first post-divestiture quarter and the first reported period under the April 2/6 Section 232 modifications on Mexico-fabricated steel. We're looking at it now because the same filing window also produced VMI's Q1 print — enabling a clean cohort discriminator on cost-structure exposure within the same end market.

What the filing says

Engineered Structures Q1 2026: Utility structures revenue $225.4M (+15.1% YoY), wind towers $70.0M (-21.3% YoY). Segment operating margin 16.9% vs 13.6% prior year (+330bps). Gross margin 25.4% vs 21.8% (+360bps). Cost of revenues decreased 1.0% despite utility volume up 15.1% — the wind cost reduction more than offset.

Backlog: Utility structures $557.6M as of March 31, 2026 — up 35% YoY, +28.2% QoQ. Management commentary: "order and inquiry activity continues to be very healthy, as customers remain focused on grid hardening and reliability initiatives, along with increasing demand for electricity stemming from AI-driven projects." First appearance of "AI-driven projects" in an ACA primary 10-Q.

Capacity: Idled wind tower facility conversion to utility structures expected operational by end of Q2 2026 — pulled forward from prior H2 2026 guidance. MD&A: "we are evaluating our Engineered Structures footprint for additional opportunities to increase capacity to meet elevated demand."

Section 232: zero mention. Generic boilerplate risk-factor language only. No new risk factors.

Construction Products Q1: Revenue $276.3M (+5.1%), operating profit $14.9M (-18.6% YoY). Attributed to planned maintenance downtime and Northeast cold weather. Q1 is seasonally weakest.

Capex guide FY2026: $215-240M. Cash $153M, total debt $1,521M, revolver undrawn ($700M available). $83M of barge proceeds prepaid the term loan in April.

What the market thinks

EV/EBITDA ≈14.5x on a $530M FY26 base — mid-cohort vs VMI 13-14x, HUBB 17x, POWL 25-30x.

Aug 2026 expiry implied vol 25.3% (103d, captures Q2 print) vs 30-day realized 47.4%. IV Rank 20% — vol is crushed. Max pain $95 (-27%). P/C 0.79 mildly bullish but only 52 OI total — illiquid.

The market is treating ACA and VMI as cohort beneficiaries of grid demand. Both at RSI 70+, both +50%+ 1Y. AI-driven language has now appeared in 5 of 7 direct cohort peers' Q1 2026 primary filings — what was novel for ACA last quarter is consensus this quarter.

The forward 12mo standalone EV (probability-weighted across bull/base/sectoral fade/cohort unwind/black swan): roughly +0.4%. Standalone is priced.

The pair (long ACA / short VMI): current ratio 0.2545 vs fair-value ratio estimate 0.300 — about 15% below. Spread EV: ~+10.8% over 12 months, with idio variance estimated at 75-85% (cancels shared XLI/PAVE/momentum exposure).

Why the gap exists

Same filing window, two tickers, opposite Section 232 exposure.

VMI's Q1 2026 10-Q (filed April 28) explicitly disclosed $220M of fabricated steel structures imported from Mexico in fiscal 2025 subject to the new tariff regime, with two formal CBP inquiries pending on valuation methodology. Management's stated assessment: "the majority of our steel poles produced at our Mexico facility will qualify for the 10% tariff rate" — best case, with downside up to 50%.

ACA's Q1 2026 10-Q (filed May 1) — three days later, same end market, comparable product class — said nothing. The carve-out for the 10% rate requires US-melted-and-poured steel content. ACA's converted wind tower facilities (Cherokee, AL and Illinois) are supplied by Nucor — domestic. The silence is consistent with a sourcing structure that doesn't bear the headwind.

Cohort price action treats the two as same-cohort beneficiaries. Forward margin trajectories diverge. Q2 2026 prints are the first quarter where the cost transmission shows in the income statement. The market sees one cohort; the filings show two cost structures.

Risks

  1. VMI passes through Section 232 cost without margin compression. The pair-trade discriminator collapses if VMI's pricing power is stronger than the tariff cost. Utility margins were expanding pre-tariff; demand is strong; pass-through is plausible at the 10% case.
  2. Cohort momentum unwind. RSI 70+/83+ on both names. Industrial sector rollover hits beta-1 names equally regardless of cost structure. Sector beta dominates idio in a 30%+ drawdown scenario; both legs lose absolute, pair may still pay.
  3. Construction Products miss persists. Q1 -18.6% OP was attributed to weather and maintenance. If Q2 doesn't recover, the standalone "another record year" CP guidance breaks.
  4. Illinois facility slips past Q2 end. Pulled-forward timing is now embedded in expectations; slippage compresses Q3 ES margin.
  5. CBP reclassifies VMI to 50%. Inverse risk — bear case fires harder than expected, but creates execution risk if VMI gaps down before pair is established.
  6. Wind-OBBBA cliff steeper than managed-decline scenario. PTC construction-start cutoff July 4, 2026 could create order pull-forward then air pocket. ACA wind backlog $600M provides ≈2.1 years of runway at current rate.

Catalysts

  • Late July 2026: ACA Q2 2026 10-Q. Tests Engineered Structures margin ≥16% (estimate 70% probability) and Illinois facility commencement.
  • Late July 2026: VMI Q2 2026 10-Q. First full quarter of Section 232 cost transmission. Tests Utility segment GM compression ≥100bps YoY (estimate 55%).
  • Mid-to-late July (both prints same week): Single highest-information window for the pair.
  • Q3-Q4 2026: First named DPA Section 303 grid infrastructure contract award. ACA is one of multiple potential beneficiaries.
  • By Dec 31, 2026: Possible second ACA capacity expansion announcement (estimate 45%).
  • July 4, 2026: OBBBA PTC construction-start cutoff — wind backlog pull-forward / air-pocket effects materialize Q2-Q3.

What would change our mind

  • VMI Q2 10-Q showing Utility segment gross margin expansion YoY → Section 232 cost transmission did not occur. Pair discriminator broken.
  • ACA Q2 ES margin <14% → mix-shift accretion thesis broken; standalone re-rates lower.
  • ACA 8-K disclosing material Mexico-fab steel exposure not previously discussed → silence was omission, not insulation.
  • Multivariate regression showing ACA idio variance >50% (vs 31% univariate) → standalone may deserve direct deployment, not only pair construction.
  • DPA Section 303 first contract award naming ACA → standalone re-rates above current fair-value estimate.