The 2025 → 2026 cycle

The distinct plan count fell from 5,712 to 5,501 — a net decline of 211 plans, or 3.7%. The gross movement was four times larger: 1,243 plan identifiers from the CY 2025 landscape did not return, and 1,032 new plan identifiers appeared — a gross-churn rate of roughly 40% on the CY 2025 base.

The implication is structural, not cyclical. Every terminated plan forces a member decision; every new plan starts at zero and has to convert. The contestable share of the market at AEP 2026 is therefore much higher than the net plan delta would suggest. For health-plan strategy teams, this is the backdrop against which all other 2025→2026 changes have to be read.

Among MA-PD and SNP plans with a non-zero monthly consolidated premium, the median fell from $37 to $31 — a 5.60-dollar drop (-15%). The mean fell less, from $47 to $44, suggesting the cheap end of the premium distribution compressed faster than the expensive end.

In the same window, the average in-network maximum out-of-pocket rose from $6,803 to $7,111 — a $308 increase (5%). The pattern is consistent with carriers shifting cost-sharing from monthly premium — the most visible signal in a Medicare.gov plan-finder comparison — toward year-end exposure that only triggers for the high-utilization minority. Headlines will read "premiums down again." The structural reality is that the median plan's out-of-pocket ceiling has moved closer to the bid's effective margin.

The Overall Star Rating distribution moved only modestly in aggregate. The 4★+ share of rated plans inched from 54.4% to 55.3%, and the average rating edged from 3.82 to 3.86. The rated-plan base shrank as terminated and too-new-to-rate plans dropped out of the denominator.

The action is in the redistribution. 5 contracts gained at least one full star; 5contracts lost at least one. CVS Health appears three times on the loss list, the only major parent with multiple downgrades. The single most material uplift is Elevance Health's H4004 — moving from 4.0 to 5.0 on a base of 105,914Jan-2026 members — which captures the Quality Bonus Payment increase and stands as one of the few unambiguous bright spots in Elevance's otherwise contracting 2026 portfolio.

On the official MA contract service area files, the number of contracts with a published footprint fell from 730 in CY 2025 to 715 in CY 2026. The average counties-per-contract fell from 594.6 to 580.0 — a 14.6-county net pullback, or roughly 2.5%. The crosswalk file records 733 "Renewal Plan with SAR" (Service Area Reduction) rows against 515 "Renewal Plan with SAE" (Service Area Expansion) rows for CY 2026 — a 1.42-to-1 ratio favoring shrinkage that explains the modest aggregate pullback without any single carrier executing a dramatic retreat.

Aggregate MA enrollment among H-prefix contracts rose by 784.2K (2.9%) between Jan 2025 and Jan 2026. That headline understates the reshuffling underneath it: of the top twelve parent organizations by enrollment, Humana added 869.9K (+17.2%) to 5.9M, Elevance Health lost 308K (-16.8%) to 1.5M, and UnitedHealth Group recorded its first January contraction on record at -84.2K (-1.1%), settling at 7.9M.

A single contract — H7617, Humana — is responsible for most of the headline movement. It went from roughly 8,600 members in Jan 2025 to about 708,000 in Jan 2026, an eighty-fold increase that bins technically as a "new entrant" anomaly because its prior-year base was effectively immaterial. That one contract accounts for roughly 80% of Humana's total parent-level Jan-over-Jan gain. Strip it out and the parent's story compresses to roughly +170K (+2.9%), and the 2026 AEP narrative no longer reads as a Humana-led market rotation. It reads as redistribution from Elevance and UHG flagships into a small number of structurally new vehicles.

CMS publishes a plan crosswalk file each October mapping prior-year plans to current-year plans. For CY 2026 it records 9,066 transitions across seven status categories — 5,064 Renewal Plan, 1,013 Terminated/Non-renewed Contract, 868 New Plan, 746 Consolidated Renewal, 733 Renewal with SAR, 515 Renewal with SAE, 127 Initial Contract. The file is authoritative for HPMS-blessed transitions. It is silent on the moves that matter most.

UCare Minnesota's five MA contracts (H2459, H8070, H2456, H5937, H8783) were all recorded as Terminated/Non-renewed with no successor row. They carried roughly 184,000 Dec-2025 members between them. Those members did not auto-roll. They re-enrolled during AEP 2026, and Minnesota's Aware Integrated (+80K) and Medica (+73K) showed up in the Jan-2026 parent-org leaderboard with the matching gains. The crosswalk file traces none of that.

The practical takeaway for any 2026 competitive analysis: do not rely on the crosswalk alone for member-migration modeling. Tie it to the monthly CPSC enrollment deltas at the (parent_org, state) grain to recover the absorption volume the official file misses. See the migrations view and the anomaly feed for the contract-level detail.

The per-contract forecasting pipeline — Prophet where each contract has enough history, seasonal-naive baseline where it does not, gated to publish only when Prophet beats baseline on a 2024 backtest — projects Jan 2027 MA enrollment at 33.2M, a 17.5% increase over Jan 2026. The gain is concentrated in UnitedHealth (+2.87M projected, +36.6% — a reversion-to-trend recovery from 2026's first-ever January contraction), with Centene (+22.6%), Health Care Service Corp (+22.2%), and Elevance (+16.9%) all projected to rebuild. Humana, this cycle's biggest gainer, is modeled to plateau at +1.7%.

Treat the UHG figure as projection, not thesis. Median Prophet MAPE on the 2024 backtest is 6.2%; the seasonal-naive baseline is 18.3%. The model has skill but its most aggressive recoveries assume the parent re-enters competitive markets at scale, and the AEP 2027 supply-side actions that would underpin that are themselves a forecast. The forecast view carries the full per-contract detail and the method note.