Oscar Health, Inc. (NYSE: OSCR) — The Individual Market’s Last Believer, Priced for a Cliff It Just Climbed

⚡ Claude’s Take

This is the author’s own independent opinion and general information — not investment advice. The analytical sections that follow carry no position and no price target.

Verdict: HOLD / AVOID adding here — a genuinely improved, well-run business at a valuation that has already priced the good outcome and left the biggest risk (2027) unpaid for. Not a short (the tape and the fundamentals are both against you), but I would not chase $32. Constructive re-entry zone ~$18–24 (≈0.4–0.5× forward sales / ≈8–10× a de-risked, mid-cycle owner-earnings number); accumulate-on-weakness below ~$20.

Oscar is the rare pure-play bet that the subsidized individual (ACA) insurance market is a permanent, growing pillar of U.S. healthcare — and, uniquely among the marketplace carriers, it walked toward the enhanced-subsidy cliff instead of away from it. Management (CEO Mark Bertolini, ex-Aetna) built its plan two years ago assuming no extension of the enhanced premium tax credits, engineered cheaper products and broker tools, and grew membership +56% to 3.2M into 2026 while Centene’s Ambetter book fell from 5.6M to 3.6M. The Q1’26 print — revenue $4.6B (+53%), MLR 70.5%, a record $679M net quarter, full-year guide reaffirmed at $18.7–19.0B revenue and $250–450M operating earnings — is a real inflection, and the market has rewarded it: the stock roughly doubled off its February low ($13→$32) to its richest-ever valuation (price/tangible-book in the 99th percentile of its own history; price/sales in the 81st).

That is exactly the problem. What you are buying at $32 is a thin-margin (~2% operating), zero-moat, government-subsidy-dependent underwriter priced at ~26–45× a set of 2026 earnings that management concedes hinge on a single unknowable input — risk-adjustment/market-morbidity, which is what blew up the entire sector in 2025 and turned Oscar’s first-ever profit ($25M in 2024) into a −$443M loss. The reported “$1B of free cash flow” is reserve/float build from membership growth, not earned profit — it reverses if the book shrinks. And the true test is 2027: the full price shock of the expired subsidies hits the next open enrollment, and a company that is ~100% individual-market has no diversified book to absorb it. The framing is momentum/quality-inflection-at-a-price, not value — a survivor re-rating that has run to the point where the reward is capped and the 2027 tail is free to the seller. Consistent with that, insiders have been distributing into the spike (June 2026 Form 4/144 cluster). Conviction: medium. Flips bullish: Congress restores enhanced subsidies and the 2026 Wakely risk-adjustment reports confirm favorable morbidity (removes the cliff and the earnings-quality overhang at once) → the multiple is defensible. Flips bearish: the mid-2026 Wakely reports show adverse morbidity (a 2025 repeat) or 2027 membership/pricing guidance signals a shrinking book → thin margins invert and the 99th-percentile multiple has a long way to fall.

One-line tag: “The best house in a market the government might stop subsidizing.”

📈 Stock Price Action — Five-Year Event Map

Oscar has completed one full boom-bust-boom round trip since its March-2021 IPO. From a $34.80 IPO price (intraday all-time high ~$37 on 3/10/2021), the stock collapsed ~94% to an all-time low of $2.05 (Dec-2022) as underwriting losses and a rising-rate de-rating of unprofitable growth names converged; it then recovered in three legs to a 52-week high of $33.10, closing $32.18 on 2026-07-02 — ~13% below the 2021 all-time high but up nearly 16× off the 2022 trough. The 52-week range is $10.69 – $33.10. Price moves below are FACT (daily adjusted price history); attributed drivers are INTERPRETATION.

Cycle narrative. (1)–(2) The 2021–22 collapse was a textbook de-rating of a cash-burning, pre-scale insurer as rates rose and risk appetite vanished — Oscar lost roughly $1.2B across those two years and the market questioned solvency. (3) The 2023–24 recovery was a genuine operational turnaround: Oscar exited unprofitable Medicare Advantage and small-group lines, repriced, drove MLR down, and posted its first-ever annual profit (+$25M) in 2024. (4) 2025 was a two-sided year — the profit was real, but the sector-wide ACA risk-adjustment/morbidity shock (the same Wakely re-measurement that forced Centene to withdraw guidance) pushed Oscar back to a −$443M loss, and the enhanced premium tax credits expired 12/31/2025, capping the stock in a $13–20 range. (5) Early 2026 marked the low on cliff-year fear. (6)–(7) The violent H1’26 rally is the “survivor” re-rating: Oscar grew membership through the cliff (+56%) by taking share from retreating competitors, printed a record Q1, reaffirmed a near-doubling of 2026 revenue, and drew a Street upgrade — carrying the stock to its richest-ever valuation. The move is a fact; whether the fundamentals justify a 99th-percentile multiple with 2027 still ahead is the subject of the memo.

1. Executive Summary

Oscar Health is the largest U.S. health insurer dedicated exclusively to the individual (ACA marketplace) and small-group markets — ~3.4M members at 2026 open enrollment, ~3.0M paid members entering Q2’26, on FY2025 revenue of $11.7B and a FY2026 guide of $18.7–19.0B. Founded in 2012 as a “tech-first” insurer and public since March 2021, Oscar sells subsidized individual health plans in ~20 states and is building an adjacent technology/distribution layer (+Oscar, and new 2026 launches Lucie — a carrier-agnostic ACA marketplace — and ICHRA X). CEO Mark Bertolini, the former chairman/CEO of Aetna, has run the company since 2023.

The investment debate is unusually clean. The business has genuinely inflected. After burning ~$1.9B cumulatively from 2020–2023, Oscar posted its first-ever annual profit in 2024 (+$25M), and — critically — grew membership +56% into 2026 through the expiration of the enhanced premium tax credits, precisely the event that shrank Centene’s Ambetter book by a third. Q1’26 was a record: revenue +53%, MLR down 490bps to 70.5%, net income $679M ($2.07 diluted). Management guides FY2026 to $250–450M of operating earnings and “meaningful profitability.”

But three things temper the story. First, there is no moat. This is a thin-margin (~2% operating), annually-repriced, government-subsidy-dependent underwriting business; members can switch every January at zero cost, and the “tech advantage” has not (yet) produced a structurally lower cost base — Oscar’s SG&A ratio (~15–16%) is double Centene’s ~7.4%, inflated by ~9–10% of fixed ACA taxes/fees on a smaller book. Second, 2025 proved how fragile the margin is: a single mis-estimate of market morbidity (the zero-sum ACA risk-adjustment transfer) flipped a profit into a −$443M loss, and that same variable is the explicit swing factor for 2026. Third, the reported ~$1B of “free cash flow” is float — reserve build from a growing membership base — not owner earnings; it reverses when the book stops growing.

The stock has already paid for the good outcome. At $32.18 it trades at its richest-ever valuation (price/tangible-book ~99th percentile of its own history; price/sales 81st) after a ~140% five-month rally, on ~26–45× a 2026 earnings number whose quality is subsidy- and morbidity-contingent. The unpriced risk is 2027: the full consumer price shock of the expired subsidies lands at the next open enrollment, and a 100%-individual-market carrier has no diversification to cushion a shrinking, adversely-selecting pool. This report takes no position and sets no price target; it lays out the embedded expectations and the falsification tests for both sides. The bull must believe subsidies are restored (or the market proves resilient without them) and that Oscar’s tech eventually earns a real cost moat; the bear must believe 2027 is a 2025 repeat with no diversified book to hide in.

2. Business Overview

Oscar Health is a tech-enabled health insurer operating a single reporting segment, ~98% of revenue from premium underwriting of subsidized individual (ACA marketplace) health plans. FY2025 revenue of $11.70B breaks down as premium $11.47B + net investment income $0.20B + other revenue $0.03B (FY2024: $8.97B / $0.19B / $0.02B). The economic model: Oscar collects a per-member-per-month premium — of which ~93% is funded directly by the federal government (CMS) via Advance Premium Tax Credits and ~7% by the member — bears the medical-cost risk, and earns the thin residual between premium and claims after acquisition and administrative costs. It is, in essence, a distributor and underwriter of federal subsidy dollars.

A defining structural feature: Oscar enrolls a healthier-than-market pool and is therefore a large net payer of ACA risk-adjustment transfers — $2.60B in 2025 (~18.5% of $14.0B direct premium, up from 14.8% in 2024, growing ~70% YoY). Every year Oscar hands ~$2.6B to competitors with sicker books. This zero-sum transfer, calculated after the fact on market-wide morbidity, is the mechanism that turned 2025 into a loss and is the single least-controllable line in the model.

Oscar has narrowed to a pure individual-market carrier: it exited Medicare Advantage and dropped its Cigna+Oscar small-group co-brand, and now sells almost exclusively individual ACA plans across ~18 states. Revenue is contract-recurring but annually re-enrolled and re-priced — members select plans each open-enrollment season and can leave every January at zero cost, so the recurrence is real but conditional on winning the annual price competition and on federal subsidy policy. Adjacent to the insurance book, Oscar is building a technology/distribution layer — +Oscar (its “Campaign Builder” platform, ~0.6M “client lives” but revenue still buried in the immaterial ~$29M “other” line, <0.25% of total) and 2025 tuck-in acquisitions (Lucie/INSXCloud, a CMS-approved enrollment platform; Trove, a brokerage; Healthinsurance.org) that back a push into ICHRA (employer-funded, employee-chosen individual coverage). These are real strategic optionality but financially immaterial today.

Membership & geographic concentration (effectuated members; the base is small, fast-growing, and dangerously concentrated):

The ~86% concentration in three states — with Florida alone at 58% — is a critical vulnerability the growth headline obscures: a single adverse state regulatory action, benchmark-premium shift, or morbidity surprise in Florida or Texas would swing the whole company. Year-to-year state share is also wildly unstable (Texas +155%, Georgia −42% in one year; full exits of California and Connecticut), confirming a low-loyalty, price-driven book rather than an entrenched franchise.

3. Industry Dynamics

Structure — a scaled oligopoly on subsidized government dollars. U.S. managed care is dominated by six scaled payers (UnitedHealth, Elevance, CVS/Aetna, Cigna, Humana, and government-program specialists Centene and Molina), plus a fringe of individual-market and insurtech names (Oscar, Alignment, Clover). For the ACA individual market Oscar serves, the economic reality is stark: the government sets the subsidy that funds most of the premium, the product re-prices and re-enrolls every single year, and margins are thin by design. Barriers to entry are real but low-return — state-by-state licensure, statutory capital, actuarial scale, network contracting — while the price and the demand are both heavily determined by federal policy the insurer does not control. (Framing draws on the broader managed-care peer set — Centene, Molina, UnitedHealth, Elevance.)

The ACA individual market: large, growing — and subsidy-dependent. Enrollment roughly doubled from ~11M (2020) to a peak of ~24M (2025) on the back of the enhanced Advance Premium Tax Credits (eAPTCs) enacted in 2021 (ARPA) and extended through 2025 (IRA). Management frames the market at ~23M lives and “a fundamental pillar of American health care.” The bull secular case — consumers want to shop for coverage like any other product; the small-group employer market is contracting into individual coverage (ICHRA); gig/part-time/early-retiree cohorts need portable coverage — is directionally credible and is Oscar’s entire strategic thesis.

The enhanced-subsidy cliff is the defining industry event (FACT / structural). The eAPTCs expired December 31, 2025, and Congress did not extend them (per Centene’s FY2025 10-K and Oscar’s Q1’26 commentary). The mechanical consequences: consumer out-of-pocket premiums jump sharply, the healthiest marginal enrollees are the most likely to drop coverage (adverse selection), and industry ACA enrollment fell ~21.5% for 2026. Oscar’s own estimate is a 20–30% market contraction; its differentiated bet was to plan for zero extension since its 2024 Investor Day, engineer cheaper plan designs and broker tools, and gain share as competitors retreated — which is exactly what the 2026 membership numbers show. But the cliff is a two-year event: 2026 was the first repricing; 2027 open enrollment is when the full price shock and any second-order adverse selection compound, and it is the single biggest unpriced industry risk in this name.

ACA risk adjustment — the mechanism that broke the sector in 2025. ACA risk adjustment is a budget-neutral, zero-sum transfer: plans with lower-acuity (healthier) members pay into a pool that compensates plans with higher-acuity members. The transfer is calculated after the plan year, depends on every other insurer’s measured acuity, and is (in Centene’s words) “subject to a high degree of estimation and variability.” In mid-2025 an independent actuarial study (Wakely) showed the marketplace pool was materially sicker than carriers had priced; healthier-than-market books (including Oscar’s) flipped from net receivers to net payers of risk-adjustment, gutting revenue and driving Oscar’s −$443M FY2025 loss. For 2026, risk-adjustment transfer as a percent of premium is guided to ~20% (tracking ~24% in Q1) — Oscar’s single most important, least controllable revenue variable. The 2026 Wakely reports (first meaningful one due Q2’26) are the year’s key swing datapoint.

Regulatory overhang beyond the cliff. The One Big Beautiful Bill Act (July 2025) tightens ACA eligibility verification and shortens enrollment windows, adding friction to the individual market; state-level dynamics (ICHRA incentives in Mississippi/Illinois; Essential Plan changes in New York) cut both ways. Net, the policy backdrop is a structural headwind with idiosyncratic pockets of opportunity (ICHRA adoption).

Marathon capital-cycle read. The insurtech/individual-market cohort attracted enormous capital in 2020–2021 (Oscar, Bright Health, Clover, Friday Health), most of which destroyed it — Bright Health and Friday Health exited/failed, validating the supply-side thesis that capital flooded a low-return, regulation-distorted pool. The survivors (Oscar, Alignment) are now consolidating share as capital exits — a constructive capital-cycle setup on the supply side, but one where the demand (subsidized enrollment) is set by Washington, not the market. Verdict: a structurally difficult, low-return, policy-driven industry with a real secular growth vector (individual/ICHRA) sitting on top of a fragile, subsidy-dependent, zero-sum-risk-adjustment margin. Attractive only for the lowest-cost, best-executing operator — and even then, cyclically and politically exposed.

4. Competitive Position

There is no durable competitive moat — this is the memo’s central business conclusion. Run through Greenwald’s taxonomy:

• Customer captivity: none. ACA plans re-price and re-enroll annually; federal subsidies make members hyper-price-sensitive at the benchmark; switching cost is effectively zero. Oscar’s own book demonstrates the churn — it attrites through the year (3.4M at OE → ~3.0M paid by Q2), and its state-level share swings violently (Texas +155%, Georgia −42% in a single year; full exits of California and Connecticut to zero).
• Network effects: none. Health insurance is not a two-sided network in any economically meaningful sense.
• Economies of scale: weak / subscale. Against Centene’s Ambetter (the #1 ACA carrier at ~3.6M even after the cliff) and Molina, Oscar is a smaller player with correspondingly weaker provider-contracting leverage — the opposite of a scale advantage.
• Cost advantage: not visible in the P&L. FY2025 MLR of 87.4% (up 570bps from 81.7%) is worse than well-run peers, and the headline SG&A ratio (~15–17%) is roughly double Centene’s ~7.4%. Crucially, though, that gap is channel-driven, not a tech penalty: it decomposes into broker commissions (~$975M, ~8% — the individual market is broker-distributed, unlike zero-acquisition-cost Medicaid) + ACA exchange fees and premium taxes (~$446M, ~4%) + “all other” true administrative cost of only ~$628M (~5.4%). So Oscar’s actual admin efficiency is competitive-to-good — the tech does show up in the true-admin line — but the overall cost structure carries the individual market’s inherent distribution load, and none of it constitutes a barrier a competitor could not replicate.

The market-share-stability test fails outright: carriers enter and exit ACA states freely (Oscar included), state share swings 30–100% YoY, and the government sets both the subsidy and the benchmark price against which Oscar competes. Oscar is a price-taker in a commodity market with a technology veneer — competent, occasionally best-executing, but with no captive customer, no scale edge, and no cost barrier that would deteriorate a rival’s economics. Verdict: a crowded, annually-contested, subsidy-driven market with weak differentiation; the “tech-first” identity improves execution at the margin but does not create a durable advantage.

5. Growth History and Forward Opportunities

Oscar’s growth is spectacular in magnitude and low in quality. Revenue compounded at ~57% from ~$1.9B (2021) to $11.7B (2025), with effectuated membership rising 1.04M (2023) → 1.68M (2024) → 2.04M (2025) → 3.17M (Q1’26, +55% YoY), and 2026 revenue guided to ~$18.8B (+61%). Essentially all of this is organic (the 2025 acquisitions were immaterial tuck-ins). But the quality problem is stark: 2025 delivered record revenue and a −$443M net loss — growth outran profitability, and the growth itself is policy-manufactured, riding the enhanced subsidies that inflated the entire ACA market and then expired at end-2025.

The 2026 share gain is genuinely impressive — Oscar grew +56% into a market that shrank ~21.5%, by preparing cheaper plan designs and broker tools for a no-subsidy world while competitors hesitated. That is real execution and the core of the bull case. But it must be read against three cautions: (i) the growth is concentrated in three states (FL 58%, TX 18%, GA 11%), a fragile base; (ii) it is front-loaded seasonally — Oscar earns in H1 (Q1’26 MLR 70.5%) and gives much of it back in H2 (full-year MLR guided 82.4–83.4%, and 87.4% actual in 2025); and (iii) the forward TAM, while real (23M+ individual lives, plus the secular small-group-to-ICHRA shift), is entirely hostage to federal subsidy policy in the near term. Forward opportunities — deeper ICHRA penetration, the Lucie carrier-agnostic marketplace, +Oscar technology licensing — are credible and potentially high-margin/capital-light, but today are optionality, not earnings. Verdict: high-magnitude, low-quality, policy-dependent growth with genuine execution alpha layered on top; the durability of the growth will be tested at the 2027 open enrollment, not before.

6. Financial Quality

The multi-year record lays the story bare — explosive top-line, a single year of profit, and a margin that snapped back to a loss:

Revenue: hyper-growth off a subsidized base. Revenue compounded from $590M (2020) → $1.92B (2021) → $4.13B (2022) → $5.86B (2023) → $9.18B (2024) → $11.70B (2025), a ~5-year CAGR near 80%, and is guided to $18.7–19.0B in 2026 (+~60%). The table’s message is the thesis in miniature: revenue up ~6× in four years, exactly one profitable year, and a 570bps MLR blowout in 2025 that erased it — all while the share count grew ~55%. This is almost entirely premium revenue from subsidized individual plans plus federal/consumer risk-adjustment and reinsurance flows; the +Oscar/services line is still immaterial. The growth is real but low-quality in the Greenwald sense — it is volume of a commodity, subsidized product, not pricing power. Two-thirds of the 2026 revenue jump is membership (+56%) and rate; the rest is metal-mix and risk-adjustment optics.

Margins: structurally thin, and they broke once already. The two ratios that matter are MLR (medical loss ratio — claims ÷ premium) and the SG&A ratio. FY2025 consolidated results produced a −$443M net loss (−3.8% net margin); the FY2026 guide is MLR 82.4–83.4% and SG&A 15.8–16.3%, leaving operating earnings of $250–450M (~1.5–2.4% operating margin). Q1’26 flattered the full year — MLR 70.5% is seasonally low (deductibles unmet early in the year; bronze-heavy new-member mix) and will rise sharply into Q4; management explicitly guides MLR “lowest in Q1, highest in Q4.” The decisive fact about margin quality: a ~1-point miss on MLR (~$185M on 2026 premium) roughly wipes out the midpoint of full-year operating earnings. That is exactly what 2025 demonstrated — Oscar’s own utilization was in line, but a change in market morbidity (risk-adjustment) alone swung the year to a loss. The margin is real but wafer-thin and exposed to a variable Oscar does not control.

The SG&A “tech” question. Oscar markets a technology-driven cost advantage, yet its SG&A ratio (~15–16%) is roughly double Centene’s ~7.4%. Part of the gap is structural — ~9–10 points are fixed ACA taxes/fees that hit a smaller premium base — but even adjusting for that, Oscar has not yet demonstrated a durable unit-cost moat; the record-low 15.2% Q1’26 ratio is progress (60bps YoY of fixed-cost leverage plus AI-driven service automation), and the trajectory is the bull’s best moat evidence, but it remains a claim in progress, not a proven structural edge. Watch this ratio: it is the single number that would convert “tech veneer” into “real advantage.”

Cash flow — float, not owner earnings (critical QoE flag). Oscar reported operating cash flow of ~$1.09B and “free cash flow” of ~$1.06B in 2025 — a year it lost $443M. The entire gap is working-capital/reserve build: claims-payable and unearned-premium reserves grow mechanically as membership grows (accounts payable rose ~$1.2B in 2025). This is insurance float, not earned profit; the correct reading is that Oscar’s economic free cash flow tracks net income, and the reported ~$1B is a growth artifact that reverses if membership stops growing or shrinks (as it may in 2027). Any valuation anchored on “3× FCF” (as some screens show) is anchoring on float and is misleading. SBC was $88M (2025), down from $160M (2023) — a positive de-escalation, ~0.7% of revenue, modest for the sector.

Balance sheet — well-capitalized, but the cash is regulated. At Q1’26 Oscar held ~$8.1B of cash and investments against $431M of debt and $1.67B of total equity (accumulated deficit −$2.6B; tangible book ~$5.59/share). It looks like a fortress — but only ~$279M sits at the parent; the rest is statutory capital trapped in the regulated insurance subsidiaries (with ~$809M of excess statutory surplus above minimums). So the “$4.4B net cash” is largely not distributable to shareholders and exists to backstop claims. Solvency risk is low; the capital is a regulatory buffer, not dry powder for buybacks or M&A. Verdict: economics do improve with scale (fixed-cost leverage is visible), but from a structurally low base, with a margin that is thin, seasonal, and hostage to an exogenous risk-adjustment variable — and cash flow that must be read through the float distortion. Quality is improving, not high.

7. Capital Allocation

A founder/VC control wedge. Oscar is a self-identified controlled company with a dual-class structure: Class B shares carry 20 votes each vs. 1 for Class A. Per the April-2026 proxy, Thrive Capital (Joshua Kushner) controls ~68% of the combined vote on just ~14% of the economics; departed co-founder Mario Schlosser adds ~13% of the vote; together the two Class B holders hold ~81% of voting power. CEO Bertolini owns ~4.5% economically but only ~1.2% of the vote (no Class B). Vanguard (~6.9%) and BlackRock (~5.3%) own more economics than management but are voting non-entities. Public Class A holders are effectively disenfranchised on board composition and any change of control — a structural governance discount that also, in fairness, insulates a long-horizon strategy from activist pressure. Thrive’s decision not to sell into the H1’26 spike is the single most constructive ownership signal.

Steady dilution, zero buybacks. Shares outstanding rose from ~210M (2021 IPO) to ~299M (Q1’26), +42% (~7%/year), entirely via SBC and equity issuance — no buybacks, ever. SBC ran $86M (2021), $112M (2022), $160M (2023, larger than that year’s entire operating result), $110M (2024), and ~$88M (2025, ~0.75% of revenue). The declining SBC trajectory is a genuine positive, but the base case remains one-way dilution — every year the per-share claim on Oscar’s earnings shrinks ~7%, a meaningful drag against the growth.

Capital is deployed defensively, not for shareholders. The ~$8.1B cash-and-investments headline is misleading: it is statutory capital trapped in the regulated insurance subsidiaries (~$1.0B surplus, ~$809M above minimum), with only ~$279M of liquidity at the parent. Debt is ~$445M of convertible notes ($410M 2030 notes issued Sept-2025 with capped calls to blunt dilution, plus $35M 2031 notes) and a new Feb-2026 revolver; Oscar uses quota-share reinsurance for capital efficiency and downside protection; there has been no M&A. This is prudent, not value-destructive — but it means the fortress balance sheet is a regulatory buffer, not distributable dry powder, and shareholders should expect neither buybacks nor dividends for the foreseeable future.

Incentives are refreshingly per-share/TSR-aligned. Bertolini’s 2025 total comp was just ~$1.15M (no new equity in 2024 or 2025 — his 2023 sign-on was front-loaded), and that sign-on grant is stock-price-hurdle PSUs vesting at $11 and $16 (both since hit) and an unearned $39 hurdle — pure total-shareholder-return alignment. The annual cash metric was shifted from Adjusted EBITDA to Operating Margin in 2025; other-executive PSUs use a cumulative Adjusted EBIT target with a relative-TSR modifier. Kushner takes zero compensation. This is a margin- and price-oriented incentive design — the opposite of membership-at-any-cost — and is the strongest governance mark in Oscar’s favor.

Insider signal — distribution into strength, no conviction buying. Across the 15 most-recent Form 4s there is not a single open-market purchase (code P). Into the H1’26 run from ~$14 to ~$30, insiders sold ~3.8M shares (~$110M gross): Bertolini sold ~2.45M shares (~$71M, largely tax-withholding on the vested price-hurdle PSUs — bullish that the $11/$16 hurdles cleared — while retaining ~7.75M shares); Schlosser exercised and sold ~1.06M shares to zero (a genuine founder cash-out); the CFO sold ~242k on a routine 10b5-1 plan. All sales are planned or sell-to-cover — no red-alert discretionary dump — but the absence of any open-market buying, even at $14, alongside ~7%/yr dilution, is a mild negative. Verdict: capital allocation is disciplined and incentives are well-aligned, but the controlled-company structure, perpetual dilution, non-distributable cash, and one-directional insider selling collectively make governance/capital-allocation a modest net negative — competent stewardship of a business whose cash the outside shareholder cannot touch.

8. Changes and Headwinds — Last Two Years

The last 24 months contain the three developments that define the thesis:

1. The 2024 profitability inflection (positive). After ~$1.9B of cumulative losses (2020–2023), Oscar posted its first-ever annual profit (+$25M) in 2024 on $9.2B revenue, validating the 2022–23 turnaround (market exits, MLR discipline, repricing). This is what re-rated the stock from ~$2 to ~$19.

2. The 2025 risk-adjustment/morbidity shock (negative). Mid-2025, an independent actuarial (Wakely) re-measurement showed the ACA pool materially sicker than priced; as a healthier-than-market book, Oscar flipped from net receiver to net payer of ACA risk-adjustment, and the year swung to a −$443M loss despite in-line own utilization. This is the sector event that also forced Centene to withdraw guidance — and it is the clearest evidence of how little control Oscar has over its own margin.

3. The enhanced-subsidy cliff (structural, ongoing). The eAPTCs expired 12/31/2025 and were not extended. Oscar’s differentiated response — planning for zero extension since 2024, engineering cheaper plans and broker tools — let it grow membership +56% (to 3.4M at OE, ~3.0M paid) while the market shrank ~21.5%. 2026 is being navigated; 2027, when the full consumer price shock lands at the next open enrollment, is the unresolved risk.

Other developments: the September-2025 convertible note issuance (capped calls) and Feb-2026 revolver shored up parent liquidity; new 2026 product launches (Lucie carrier-agnostic marketplace, ICHRA X data exchange) extend Oscar toward an asset-light, unregulated distribution/technology model (management pitches higher-margin, no-risk-capital economics — unproven, detail promised at the September 16, 2026 Investor Day); Bertolini’s contract was extended through ≥2029. Verdict: the changes net to a business that is materially stronger operationally than two years ago, but whose central risk (subsidy-dependent, single-market, morbidity-exposed) has not diminished — it has merely been deferred to 2027.

9. Risk Analysis

10. Valuation Discussion (embedded expectations)

Oscar is hard to value on earnings because GAAP net income is (a) negative or near-zero on a trailing basis and (b) violently seasonal (Q1’26 alone was +$679M; the full year guides to $250–450M of operating income). The market is therefore pricing it on sales, tangible book, and a forward view of “normalized” earnings — and on all three it sits at the top of its own historical range.

Diluted market cap ~$10.6B (329.75M diluted shares × $32.18); on-balance-sheet cash is mostly regulated, so EV is only modestly below market cap on a distributable basis (~$10.3B if you credit only parent cash). Embedded expectations: at ~0.7× sales on a ~2% operating-margin business, the market is underwriting that Oscar (i) sustains ~60% 2026 revenue growth into durable double-digit forward growth, (ii) expands operating margin from ~2% toward the 4–6% “meaningful profitability” management implies over time, and (iii) does not suffer a 2025-style risk-adjustment reversal or a 2027 membership contraction. Put differently: to justify $32, you must believe Oscar reaches ~$500–700M of sustainable operating earnings within a few years (≈4% of a ~$18–20B book) and that the subsidy/enrollment base holds. That is a coherent bull outcome — but it is the base case already in the price, not an upside scenario.

Scenario sketch (illustrative, not a target):

• Bear (subsidies not restored; 2027 membership −25–35%; a morbidity/risk-adjustment reversal): revenue falls back toward ~$13–14B, operating margin compresses toward breakeven, and a thin-margin, shrinking insurer re-rates toward ~0.25–0.35× sales — the low end of its own history. Downside is large because both the numerator (earnings) and the multiple compress together.
• Base (2026 guide met; 2027 book roughly flat-to-modestly-down; ~2–3% margin): ~$18–20B revenue, ~$400–600M operating income, and a ~0.4–0.5× sales / ~15–20× normalized-earnings multiple — i.e., roughly today’s revenue base but a lower multiple: a stock that grows into a valuation that stops expanding.
• Bull (subsidies restored or market proves resilient; margin marches to 4–6%; +Oscar/Lucie/ICHRA add a real high-margin, capital-light layer): ~$700M–$1B operating income and a defensible ~0.6–0.8× sales — the current multiple justified and growing.

The asymmetry at $32 is unattractive: the bull case is in the price, the base case implies multiple compression, and the bear case (a repeat of 2025’s exogenous shock, which is not remote) is a large drawdown. No price target; no recommendation.

Price Action, Momentum & Factor Positioning

Oscar is, quantitatively, a high-beta, high-idiosyncratic-risk momentum name in a violent re-rating. A standard equity factor model shows a Market beta of ~1.1–1.2 and an Industry (Healthcare Providers) beta ~1.2, but with R² of only ~0.19–0.21 — i.e., ~80% of Oscar’s variance is idiosyncratic (company/subsidy-policy-specific), not factor-driven. This is not a stock the factor models “explain”; it trades on its own policy and print narrative. Relative strength is extreme: rs_6m ~122, rs_ytd ~124 (an 80% YTD move), and the leaderboard shows a 1-year return near +94% — but sitting on a five-year record scarred by a −90.4% max drawdown and only +7.7%/yr over five years. The short-horizon Sharpe figures are enormous (m3/m6 annualized), which is the mechanical signature of a low-base recovery inflection, not a durable compounder — de-annualize and the message is “a stock that fell 94% and has doubled twice off the bottom.”

Read for the thesis: the tape is a one-way street up right now — a crowded, high-momentum survivor re-rating, not a falling knife. That argues strongly against shorting (you would be fighting both the fundamentals and the trend). But it equally argues against chasing: extreme relative strength into the richest-ever multiple, with insiders distributing and a known 2027 policy risk unpriced, is the classic profile of a name where the easy money — the trough-to-fair re-rating — has already been made. The factor read supports Claude’s “HOLD / don’t-chase, accumulate-on-weakness” framing: the positioning is late-momentum, and the next leg is a fundamental (subsidy + morbidity) coin-flip the model cannot price. Factor-similar names are thin (closest common-stock peer is Alignment/ALHC — another unprofitable, richly-valued insurtech), underscoring that Oscar is a special situation, not a factor basket. (Facts — loadings, returns, drawdowns — are reportable; “continues/mean-reverts” is interpretation. No price target, no support/resistance levels.)

11. Variant Perception

Consensus view. The Street has swung constructive in 2026: Oscar is “the individual-market winner” that navigated the subsidy cliff by taking share, printed a record Q1, and is on a credible path to “meaningful profitability” — hence the Barclays upgrade to Overweight (PT $35) and the 80% YTD move. Consensus underwrites 2026 guidance and extrapolates the share-gain and margin-leverage trend forward.

Strongest bull case. Oscar is the low-cost, best-executing pure play on a large, secularly-growing individual/ICHRA market that big diversified payers are de-emphasizing. Its tech/AI stack is finally delivering visible SG&A leverage (record-low 15.2% in Q1’26); management is uniquely credible (Bertolini) and incentivized on margin/TSR, not volume; and the Lucie/ICHRA/+Oscar layer could add a high-margin, capital-light, unregulated distribution business on top of the underwriting book. If enhanced subsidies are restored (a live 2026 political possibility) or the market proves resilient without them, the cliff risk evaporates and Oscar compounds share and margin from here — the current multiple is then cheap for the growth.

Strongest bear case. This is a no-moat, ~2%-margin, single-market, subsidy-dependent underwriter at its richest-ever valuation, whose margin is hostage to an exogenous, unforecastable variable (market morbidity/risk-adjustment) that already produced a −$443M loss once. The “$1B FCF” is float that reverses when growth stops; the cash is regulated and non-distributable; dilution runs ~7%/yr; insiders are selling. 2027 is the reckoning — the full subsidy price shock hits the next enrollment with no diversified book to cushion adverse selection — and none of it is priced at $32. Multiple compression is the base case even if operations merely hold.

The 3–5 assumptions that matter most:

1. Do enhanced subsidies get restored, and how does 2027 enrollment hold? (Policy — the single biggest swing; unforecastable.)
2. Is 2026 market morbidity favorable or adverse? (The Wakely reports; determines whether risk-adjustment is a tailwind or a 2025 repeat.)
3. Does the SG&A “tech” advantage become a real, durable cost moat (ratio structurally below peers, ex-taxes), or stay a claim?
4. Can +Oscar/Lucie/ICHRA become a material, high-margin business, or remain immaterial optionality?
5. Does membership growth prove profitable at maturity (mid-cycle MLR + operating margin), or is it share bought at breakeven?

Falsification. Bull is falsified if 2027 membership/pricing guidance signals a shrinking book, or the H2’26 Wakely reports show adverse morbidity. Bear is falsified if subsidies are restored and Oscar sustains growth with a stable/expanding operating margin and a visibly peer-beating cost structure. The factor read (extreme relative strength, 99th-percentile valuation, insider distribution, ~80% idiosyncratic variance) says consensus is now positioned long the good outcome — offsides to the downside if 2027 disappoints, with little cushion.

12. Fact vs. Interpretation

13. Open Questions

1. Will Congress restore the enhanced premium tax credits for 2026/2027, and on what terms? (The dominant swing variable.)
2. What do the 2026 Wakely risk-adjustment reports (Q2 and H2) show on market morbidity — tailwind or a 2025 repeat?
3. What is 2027 membership and pricing guidance — does the book shrink when the full price shock lands?
4. What is Oscar’s true normalized (mid-cycle) operating margin once MLR seasonality and risk-adjustment normalize — 2%, 4%, or 6%?
5. What are the actual economics and scale of Lucie/ICHRA X/+Oscar (revenue, margin, capital) — to be detailed at the Sept-16-2026 Investor Day?
6. How much of the record Q1’26 was pull-forward/seasonality vs. a sustainably higher earnings base?
7. What is the distributable cash at the parent over time, and does Oscar ever return capital, or is it perpetual dilution?

14. What Must Be True

Bull case — what must be true: (a) enhanced subsidies are restored, or the individual market proves durably resilient without them (2027 enrollment roughly holds); (b) 2026–27 market morbidity is neutral-to-favorable, so risk-adjustment does not reverse; © Oscar’s SG&A/tech leverage continues, lifting operating margin from ~2% toward 4–6%; (d) Lucie/ICHRA/+Oscar add a real high-margin, capital-light layer. Falsification test: if 2027 membership guidance is down materially, or any 2026 Wakely report shows adverse morbidity, or the operating margin fails to expand in 2027, the bull thesis is broken — a thin-margin, shrinking underwriter cannot support a 99th-percentile multiple.

Bear case — what must be true: (a) 2027 enrollment contracts sharply as the subsidy shock hits, with adverse selection lifting MLR; (b) a morbidity/risk-adjustment reversal (2025 redux) compresses the thin margin toward or below breakeven; © the tech advantage never becomes a structural cost moat; (d) the rich multiple compresses toward the low end of Oscar’s own history. Falsification test: if subsidies are restored and Oscar prints two-plus consecutive years of growing membership with an expanding operating margin and a visibly peer-beating cost ratio, the “no-moat, cliff-doomed” bear thesis is broken and the business is a genuine, if cyclical, franchise-in-the-making.

APPENDIX A — Standard Diligence Questionnaire — Oscar Health, Inc. (NYSE: OSCR)

Report date: 2026-07-03. Grounded in the underlying analysis; Fact / Interpretation / Assumption labeled where it matters. Where a question does not map to a subsidized individual-market insurer, the correct sector analog is given.

General

What thoughtful questions have other investors asked? (1) Is the 2024 profit / Q1’26 record repeatable, or is it seasonal front-loading plus a favorable risk-adjustment swing? (2) What happens to membership at the 2027 open enrollment when the full enhanced-subsidy price shock lands? (3) Is the “tech advantage” real, or is Oscar just a normal ACA underwriter with a better app? (4) Is the ~$1B of reported free cash flow real or float? (5) Can +Oscar/Lucie/ICHRA ever be material? (6) Given a controlled-company structure and ~7%/yr dilution, does the outside shareholder ever see cash?

Cyclicality & Earnings Nature

Cyclical high or low? Interpretation: Oscar is at a cyclical/seasonal high in perceived earnings (record Q1’26) but a structurally early point in normalized profitability — it has been profitable in only one full year (2024). Earnings are intra-year seasonal (H1-heavy: low early-year MLR as deductibles are unmet; H2 gives it back) and policy-cyclical (subsidy regime). External vs. internal? Both: the 2022–24 turnaround was internal (market exits, repricing, MLR discipline); the 2025 loss was external (market-wide morbidity/risk-adjustment). Revenue stability? Contract-recurring within a year but fully re-priced and re-enrolled annually and subsidy-dependent — low durability. Market size/direction? The individual ACA market is ~23M lives and secularly growing (individual/ICHRA shift), but near-term shrinking post-subsidy-cliff (−21.5% industry in 2026). Domestic only.

Business Quality & Competitive Moat

Industry more or less competitive? Structurally competitive and low-return; carriers enter/exit freely; the government sets subsidy and benchmark price. Profitability (ROIC/ROE)? Poor and volatile — ROA −7.9% (2025), +0.6% (2024); ROE meaningless on an accumulated-deficit book; normalized operating margin only ~2% at the 2026 guide. Industry profitability / barriers? Thin margins by design; barriers (licensure, statutory capital, actuarial scale) are real but produce low returns on capital. Easily understood? Yes — a subsidized-premium underwriter. Undermined by low-cost foreign labor? No (domestic regulated service). Do brands matter? Marginally — price and benchmark position dominate; “Oscar” has consumer awareness but members are price-driven. Nature of competition? Annual price/benefit competition on the exchanges vs. Centene/Ambetter, Molina, Blue plans. Switching costs? ≈zero — members can switch every January at no cost. Verdict: no durable moat.

Financial Condition & Balance Sheet

Assets not fully recognized? The +Oscar/Lucie technology and enrollment platforms carry little balance-sheet value relative to their strategic pitch (optionality). Off-balance-sheet liabilities? None material flagged; quota-share reinsurance shifts risk contractually. How conservative is the accounting? The key judgment area is risk-adjustment and IBNR reserve estimation — inherently uncertain and the source of the 2025 miss; management took a “cautious” Q1’26 risk-adjustment posture. CapEx-hungry? No — asset-light (~$36M capex/yr, ~0.3% of revenue); the capital intensity is statutory surplus to back claims, not physical capex.

Capital Allocation & Management

FCF generation / use / philosophy? Interpretation: reported operating cash flow (~$1.09B in 2025) is largely float/reserve build from membership growth, not owner earnings; true economic FCF tracks net income (negative in 2025). No capital return. Recent acquisitions? Small 2025 tuck-ins (Lucie/INSXCloud, Trove, Healthinsurance.org) for ICHRA/distribution — immaterial. Buying back shares? No — never. Issuing shares to insiders? Yes — steady dilution ~7%/yr via SBC ($88M in 2025) and equity; ~210M → ~299M shares since IPO. Director/management comp? Bertolini 2025 total ~$1.15M with price-hurdle PSUs ($11/$16 hit, $39 unearned) and an operating-margin cash metric — well TSR-aligned; Kushner takes zero. Management motivations? Margin/price/TSR-oriented (a positive), but within a controlled-company structure (Thrive ~68% vote on ~14% economics) that disenfranchises public holders.

Valuation & Market Data

ADR/MLP/K-1? No — a Delaware C-corp, common stock (Class A on NYSE). Dividend policy? None; none expected. How profitable? Marginally and only recently; ~2% operating margin at guide. Net income vs. cash from operations diverging? Yes, sharply and importantly — 2025 showed −$443M net income against +$1.09B operating cash flow; the divergence is reserve/float build, a growth artifact that reverses if the book shrinks — do not read it as earnings quality.

Risks & Downside

What would cause the stock to decline? A 2027 membership contraction as the subsidy shock lands; an adverse 2026 Wakely/risk-adjustment print (2025 redux); an MLR miss (a ~1pt miss ≈ full-year operating income); multiple compression from the 99th-percentile valuation; policy tightening. Catastrophic-loss risk? Low near-term — well-capitalized (statutory surplus + ~$809M excess); a total loss would require a multi-year morbidity spiral or a wholesale collapse of the subsidized individual market. Total loss chance? Low, but the equity is more exposed than a diversified payer given single-market concentration and thin margins.

Recent News & Events

Environment changed recently? Yes — the enhanced premium tax credits expired 12/31/2025 (the defining change), and Oscar grew share through it (+56%). Acquisitions? Minor 2025 tuck-ins. Accounting changes? None material; ongoing risk-adjustment estimation is the key judgment. Recent changes — markets, facilities, management? New 2026 product launches (Lucie marketplace, ICHRA X); Sept-2025 convertible notes + Feb-2026 revolver; Bertolini contract extended through ≥2029; Investor Day scheduled Sept 16, 2026. Stock at a 52-week high (~80% YTD) after a Q1’26 record and a Barclays upgrade.

APPENDIX B — Source Appendix — Oscar Health, Inc. (NYSE: OSCR)

Report date: 2026-07-03. Primary sources first. Facts reconciled to filings where possible; third-party aggregated data (fundamentals aggregators, public market data, factor models) labeled and used as cross-checks, not primary authority.

Primary — SEC filings (EDGAR, CIK 0001568651)

• Form 10-K, FY2025 (filed 2026-02-13; oscr-20251231) — revenue segmentation, premium/investment/other split, MLR, SG&A composition, membership, state footprint, risk-adjustment payable ($2.60B), dual-class structure, statutory surplus, reinsurance, convertible notes, risk factors. Primary source of record.
• Form 10-K, FY2024 (filed 2025-02-20; oscr-20241231) — prior-year comparison (first profitable year, +$25M).
• Form 10-K, FY2021–FY2023 (oscr-20211231 / -20221231 / -20231231) — multi-year revenue, loss, share-count, SBC history.
• Form 10-Q, Q1 2026 (period ended 2026-03-31) — Q1’26 income statement, balance sheet ($8.1B cash/investments, $1.67B equity, ~299M shares), membership 3.17M.
• DEF 14A proxy (April 2026) — dual-class voting (Class B 20 votes), Thrive/Kushner ~68% vote on ~14% economics, Schlosser stake, Bertolini comp (~$1.15M, price-hurdle PSUs), operating-margin cash metric, controlled-company status.
• Form 8-K, 2026-06-08 (Item 7.01, Reg FD) — mid-year business update reaffirming FY2026 outlook (drove +12% on 6/9).
• Form 8-K, 2026-06-09 (Item 5.07) — annual meeting voting results.
• Form 8-K, earnings releases (Q4’25, Q1’26 results of operations).
• Form 4 / Form 144 corpus (Jun–Jul 2026) — insider transactions: Bertolini ~2.45M shares sold (~$71M, tax-withholding on vested PSUs); Schlosser option-exercise + ~1.06M share sale to zero (~$31M); CFO Blackley ~242k on 10b5-1; no code-P open-market purchases; Thrive no sales.

Primary — Company disclosures

• Oscar Health Q1 2026 earnings call transcript (2026-05-06; CEO Mark Bertolini, CFO Scott Blackley) — FY2026 guidance ($18.7–19.0B revenue, MLR 82.4–83.4%, SG&A 15.8–16.3%, operating earnings $250–450M, adj. EBITDA +~$115M over EBIT); risk-adjustment ~20% (tracking 24% Q1); MLR 70.5%; membership 3.2M (3.0M paid Apr-1); Wakely market-contraction 20–30%; Lucie/ICHRA X launches; Investor Day Sept-16-2026; “planned for no eAPTC extension since 2024 Investor Day.”
• Oscar Health Investor Relations (ir.hioscar.com) — earnings releases, supplemental data.

Industry / regulatory (public)

• Centene Corp. FY2025 10-K (filed 2026-02-17, CIK 0001071739) — used for ACA industry framing: enhanced APTC (eAPTC) expiration 12/31/2025 and non-extension; ACA risk-adjustment mechanism and the 2025 Wakely morbidity shock; 2026 industry ACA enrollment −21.5%; Ambetter membership 5.6M→3.6M; OBBBA provisions. (Public filing; used for industry cross-read.)
• Wakely (independent actuarial) marketplace morbidity reports — referenced via Oscar Q1’26 call and Centene 10-K (the 2025 sector risk-adjustment catalyst; 2026 reports pending Q2/H2).
• CMS ACA marketplace enrollment and APTC program rules (public).

Third-party quantitative (cross-checks; reconciled to filings)

• ROIC.ai (aggregated fundamentals) — income statement, balance sheet, cash flow, per-share, profitability ratios, enterprise value, valuation multiples (annual/quarterly). Used for multi-year trend and ratio cross-checks; EDGAR/10-K remain primary.
• Public market data — daily adjusted price history (five-year event map); own-history valuation percentiles (P/B 99.7th, P/S 81st, composite 90.5th pctile); public news (June-2026 catalysts: Reg-FD business update; Barclays upgrade to Overweight, PT $30→$35).
• Factor model data — factor loadings (Market beta ~1.1–1.2; Healthcare Providers ~1.2; R² ~0.19–0.21), leaderboard (y1 +94%, y5 +7.7%/yr, max DD −90.4%), relative strength (rs_6m ~122), related stocks (ALHC closest peer). Statistical estimates, not primary.

Notes on authority

Where third-party aggregated figures diverge from the filing, the filing governs. Management commentary (transcript) is treated as hypothesis and validated against filings and external data. No analyst price target or third-party rating was adopted as a view; all valuation framing is embedded-expectations/scenario analysis. This report carries no recommendation and no price target outside the clearly-labeled Claude’s Take.