CMRE's +92% one-year run is containership sector beta. A multivariate regression against SPY plus an equal-weight peer basket (DAC, GSL, ESEA, NMM) gives γ_PEERS = 0.94 (t=10.95) over 400 days; β_SPY collapses to 0.16 (insignificant) once peers are in the model. Idiosyncratic variance is 56% over 400 days, 33% over the recent 90. Orthogonal alpha is meaningfully negative. The Q1 announcement of a 16-ship, 15-year COSCO charter program — the largest single transaction in CMRE history — has not yet shown up in price. The thesis is not that the market is wrong today. It is that GSL's CEO publicly described long-tenor charters as "challenging to secure" in his Q2 2025 call; that zero of six US-listed peer lessors matched the pattern in Q1 2026; and that 15 years of locked tonnage will earn through while peers roll lower-rate charters in 2027-2028. The bilateral is conviction premium, not alpha premium — the edge is forward-dated.

What we know

Contracted revenue backlog jumped from $3.4B (end-Q4 2025) to $6.2B (end-Q1 2026), up 82% sequentially. TEU-weighted duration extended from 4.5 to 6.1 years. The driver: 12 × 9,200 TEU vessels at 15-year tenor plus 4 × 3,100 TEU at 8-year tenor, all chartered to COSCO, all financed 1:1 by two Chinese policy banks at maturities matching the charters. Revenue days fixed: 97% (2026), 94% (2027). Twenty-two newbuilds are under construction. Common dividend was raised 8.7% to $0.125/quarter in the same print.

Q1 itself printed soft on the surface — voyage revenue −7.2%, net income from continuing ops −27% YoY — driven by extra dry-docks and peak-cycle charter rolloffs. Operating cash flow $112.4M; liquidity $644.4M; total equity $2.224B. Preferred coverage (Series B/C/D combined) is 16x quarterly on net income and 5.5x annual on operating cash flow against $20.5M annual obligation. No arrears.

Why we think it's non-replicable

The cross-ticker check is the work the market hasn't done. Six US-listed peer lessors checked: GSL, DAC, ESEA, NMM, CCEC, SFL. Zero captured a parallel COSCO program at comparable scale or tenor. DAC has seven legacy COSCO charters and a Q4 2025 newbuild template that tops out at 10-year tenor on smaller vessel classes (5,300 TEU). GSL's anchor counterparty is CMA CGM via founder relationship — not COSCO. ESEA's Q1 2026 newbuild orders went to Chinese yards without announced charters at all. Loadstar's January 2026 reporting and Lloyd's List subsequently characterize the Costamare deal as the third or fourth round of a developing bilateral with COSCO — allocated to a preferred Greek-owner counterparty rather than competitively bid. COSCO is itself ordering 87 vessels at affiliated yards, "stretching beyond self-owned newbuilding orders" via charters with selected independents. CMRE captured allocation; peers did not.

GSL CEO Tom Lister, Q2 2025 (nine months before CMRE's announcement): "still rather challenging get significantly long-term charters... liner operators lid on speculative orders." A peer was actively trying for what CMRE eventually landed and could not secure it. That is the most credible possible source on non-replicability — competition saying out loud what they couldn't do.

What we don't know — the structural unknown

The 15-year charter rates are not disclosed. This is the central open question, not a ranked risk. A $6.2B contracted backlog at below-market rates is a $4-5B backlog dressed up. Until rate quality is verifiable — likely Q3 2026, when first newbuilds approach delivery and segment-level daily-rate disclosure typically appears in MD&A — the thesis has a structural ceiling. The rate question determines whether the moat is durable high-margin cash flow or margin-traded visibility.

This is the single resolution that most narrows the magnitude of the LR.

Counterparty concentration

Twelve of the sixteen vessels are 15-year commitments to a Chinese state-owned counterparty during a US/China port-fee escalation regime ($80/NT in 2026, scheduled to $140/NT by 2028). Even at a conservative annualized 2-3% probability of meaningful charter-impacting US-China escalation, the cumulative 15-year base rate is 25-40%. Median outcome is uneventful; the tail is real and irreducible. The preferred coverage of 16x is meaningless if the primary contracted cash flow becomes contested. Geopolitical exposure is not in the LR — it is a structural caveat the LR sits on top of.

Two instruments, two theses

CMRE common is the thesis vehicle. A separate regression of CMRE-PC against SPY + CMRE common shows β_CMRE = 0.015 (statistically zero), 97.8% idio variance, daily correlation 0.11 to the common. The preferred is a pure rate/credit instrument; it will not participate in any equity re-rating. Anyone wanting the COSCO-bilateral exposure expresses it through the common.

CMRE-PC is a yield instrument. Series C 8.500% cumulative redeemable perpetual at a premium to par yields 8.0% with 16x coverage. It is decoupled from the equity thesis. Trading volume is light by institutional standards — pre-trade liquidity due diligence required for meaningful sizing.

Current spot is at 94% of the 52-week range. Direct equity entry today buys a marginal-Sharpe bet on a forward-dated factor. We assign roughly 25% probability that γ_PEERS compresses below 0.7 by Dec 2026 (the observable materialization signal) and roughly 40-50% probability the thesis materializes in some form by 2027-2028. Sizing under sub-threshold idio (56%) and negative current α_orth implies modest weight, catalyst-gated additions. Pullback to $14-15 (Q2 disappointment or China headline noise) materially improves the entry Sharpe.

Decision points

A second COSCO tranche announcement at any point validates "repeatable" and raises re-rating probability. Konstantakopoulos family Form 4 open-market buying (code P) at current zone is the strongest available management-conviction signal; selling contradicts the structural-moat frame.

Evidence

Memo LR: 1.3 — bullish, the market is underweighting a forward-dated factor decorrelation thesis. Direction is well-corroborated; magnitude is constrained by three things the LR sits on top of: (1) charter rates are not disclosed, capping how much we can claim about backlog quality; (2) γ_PEERS = 0.94 confirms the thesis has not materialized in returns; (3) 15-year duration with a Chinese SOE counterparty carries a 25-40% cumulative geopolitical base rate. The information is most valuable to a 12-24 month horizon with catalyst-gated sizing in the common. The cleanest entries come at $14-15 zone. The preferred is a yield instrument, not a thesis vehicle.