The market prices a 38% probability that Modine Manufacturing completes its industrial spin-off and re-rates as a DC cooling pure-play. We price 60%. The mechanism is fixed-schedule — segment recast July 2026, spin close H2 CY2026 — not open-ended execution. At $295.88, the 22-percentage-point gap is +31% probability-weighted return over 12 months, Sharpe 0.81.

Setup

Modine Manufacturing is a thermal management company separating its automotive segment (Performance Technologies) from its Climate Solutions business via Reverse Morris Trust into Gentherm. The Q4 FY2026 print disclosed the only dollar-specific, single-customer binding capacity commitment in the DC cooling cohort: a $4 billion chiller agreement with an unnamed hyperscaler covering CY2027–2029. The spin leaves RemainCo as a DC cooling pure-play currently priced at an industrial multiple.

What the filing says

CEO exact language: "a landmark $4 billion long-term agreement for chiller sales with a major hyperscale customer, cementing Modine's position as a critical partner for data center cooling." Slide 4 confirms "more than $4 billion of cooling products during calendar years 2027 through 2029" — ≈$1.33B/year from one customer.

The deposit confirms it is binding: contract liabilities increased $159M in FY26 vs a $44.5M decrease in FY25. Q4 FCF of $152.8M was driven by "a single DC customer upfront cash payment." The counterparty wired money; this is not backlog language.

DC revenue grew 158% YoY in Q4 to >$400M, despite losing 20 production shifts to weather. FY27 guidance: DC +60–80% on an ≈$1.15B FY26 base, implying $1.84–2.16B. The $2B target management gave for FY2028 in February has been pulled forward a full year.

July 2026: MOD publishes recast FY26 financials in three segments — Data Centers, Commercial HVAC, Performance Technologies. First standalone DC P&L. RMT spin confirmed on track for CY2026.

What the market thinks

Current price $295.88 (+13.57% on the print). Enterprise value $16.3B.

RemainCo (Climate Solutions) FY27 EBITDA estimated $505–540M (midpoint $520M). PT disposal value ≈$1.6B ($210M RMT cash + 40% THRM est. value). Net debt ≈$392M allocated to RemainCo. VRT trades 52.5x trailing EV/EBITDA.

Back-solving: at $295.88, market implies 38% probability of pure-play re-rating. We estimate 60%. Sharpe 0.81 — above the 0.5 investable floor, below 1.0 target; justifies a starter position, not full weight.

Why the gap exists

MOD co-moves with the industrial sector (XLI) and does not co-move with DC cooling (VRT). Factor regression: market β=2.59, industrial sector β=2.63 — both dominant. DC theme β≈0, indistinguishable from noise. PT's automotive thermal exposure is dragging an industrial multiple onto a business growing 158% YoY in data centers.

The counterparty is XLI-classified institutional holders priced at industrial multiples because no standalone DC P&L exists. The spin creates forced re-rating: not waiting for holders to update their models, but a structural event that removes PT and requires the market to value DC independently for the first time. The July segment recast is the mechanism.

Cohort check confirms the $4B disclosure is idio, not sector repricing: VRT's $15B is cancellable orders with no single-customer binding disclosure. Trane confirmed "long-term capacity master purchase agreements" — zero dollar specificity. CARR and AAON have nothing comparable. META ($107B step-up in contractual commitments Q1 2026, capex raised to $125–145B) is the strongest hyperscaler candidate.

Risks

1. $4B cancellation mechanics — primary underwriting gap. ≈$1.33B/year from one hyperscaler is ≈60–70% of the FY27 DC guide. The $159M deposit is partial protection, but the contract structure has not been disclosed: we do not know whether the deposit is a minimum purchase guarantee, partially refundable on cancellation, or has specific termination triggers. The July recast or a direct management call should clarify the floor.

2. PT spin timing. IRS private letter ruling is the bottleneck. Announced January 2026; PLR range is 9–12 months, meaning arrival October 2026–January 2027. A CY2026 close requires the PLR to land at the faster end and Gentherm shareholder vote, SEC registration, and closing to clear before year-end. P(CY2026 close): 65%.

3. Factor beta amplifies drawdowns. β_SPY = 2.59, β_XLI = 2.63. A 10% market correction delivers ≈26% drawdown from factor exposure alone before any thesis change. The high beta persists until PT separates.

4. TT disclosure risk. Trane Technologies Q2 2026 (August) may disclose comparable single-customer binding agreements with dollar specificity. If it does, the idio premium predicated on disclosure uniqueness compresses. P(TT discloses): 20%.

5. Margin recovery assumption. CS adj EBITDA margin was 18.7% Q4 FY26 vs 21.4% prior year. Management attributed headwinds to temporary capacity expansion, tariffs, and weather. If DC segment margin in the July recast comes in below 20%, RemainCo EBITDA misses $505–540M and the gap math narrows.

Catalysts

Sizing

Idio variance at 54.5% (below the 75% target) reflects unintended industrial sector beta that resolves post-spin — a sizing discount that narrows as the July recast and spin close remove the conglomerate structure. Edge erodes above $325; at that price the market-implied rerate probability reaches 53% and the gap to our 60% estimate no longer justifies the factor risk.

What would change our mind

• 8-K announcing PT spin termination. $45M break fee from Gentherm and a 3-month extension exist, but IRS PLR denial collapses the RMT structure. Exit immediately.
• DC guidance miss >20% at August earnings. Implies execution failure or contract delivery delay — the contracted floor becomes uncertain.
• TT discloses a $1B+ single-customer binding agreement. Comparable TT disclosure compresses the idio premium predicated on disclosure uniqueness.
• July recast: DC segment EBITDA margin below 20%. RemainCo intrinsic value falls ≈15% from base; Sharpe drops to ≈0.6 — borderline investable.
• Hyperscaler capex cut >20%. Deposits provide some floor, but undisclosed cancellation mechanics mean the height of that floor is unknown.

On LR 1.6

Evidence stack is strong: $4B contract (LR 3.0), deposit confirmation (LR 2.2), guidance (LR 2.0), cohort uniqueness (LR 1.5). But memo LR reflects net signal after discounts: the three LR 2.0+ items are from the same 8-K — correlated, not independent. Single-customer concentration with undisclosed cancellation mechanics adds structural uncertainty the filing doesn't resolve. And idio variance at 54.5% means 45% of the position's variance comes from market and industrial sector factors, not the thesis — the thesis could be right and the position still underperforms if industrials sell off pre-spin. Net discounted signal: 1.6.

Evidence

LR signal: 1.6