Arena: Market Conditions Before Oscar Health
People buying health insurance without employer coverage · 2013 · New York first, then exchange states across the US
Regulatory shiftRegulatory constraintsInformation asymmetry
Before 2014, few New Yorkers bought health insurance on their own: the downstate individual market held about 20,000 people, and the 2012 Supreme Court ruling upholding the ACA was expected to grow it toward a million OS31. From October 2013 the new state exchanges let people without employer coverage compare plans from competing insurers, including some of the country's largest carriers, side by side OS4 OS17. Buyers struggled to read their plans and bills, and no one explained what care would cost OS6 OS46. Any new carrier needed a state license and had to hold most of its starting capital in regulatory reserves OS16.
How each step happened
Step 1 of 5 · 2013–18
Own claims and data systems, not vendors
Oscar's founders chose to run the insurance machinery themselves, "a full-stack platform that we control end-to-end," which they called "the only way to create a differentiated member experience" OS4. Schlosser had first planned to assemble the best vendor software, and gave that up after meeting the vendors OSX-1. At its October 2013 launch in New York this meant free phone doctors, free generics and search that matched symptoms to doctors; 15,000 people joined in the first year, at a silver premium 8.6% above the city median OS16 OS4 OS59.
The larger payoff was information: "the bill has to get to the insurance company," Schlosser said, so an insurer that pays its own claims sees costs as they arrive OS31. Oscar replaced its COBOL-based claims vendor with its own engine, in parallel run by 2018 and now paying about 98% of claims automatically OSX-4 OSX-3 OSX-5. The build was costly: in 2016 Oscar spent over $100 per member per month on administration, against a $49 average at other New York plans OS59. Scale later cut the overhead: the administrative expense ratio of Oscar's insurance company fell from over 50% in 2017 to 26.1% in 2020 OS4. It did not touch the biggest cost, though: on a rented, high-rate network Oscar paid over 75% of premiums toward hospital costs, against 63% at UnitedHealthcare, and lost $92 million in New York in 2015 OS59 OS50 OS6. Its data showed which contracts to change; step 2 changed them.
Rivals Bright Health's prospectus promised an "end-to-end intelligent technology platform" that former employees called "more hype than code" BH39. When its membership surged in 2021, it processed 40% to 50% of the year's claims in the fourth quarter and learned its risk scores were wrong BH23 OSX-6.
Vertical Integration Strategy
Step 2 of 5 · 2016–18
Narrow partner networks priced to win
Exchange shoppers sort plans by premium, and a premium follows mostly from hospital prices OS59. From 2016 Oscar gave up broad networks. It dropped Memorial Sloan Kettering and NewYork-Presbyterian affiliates in New York and built its Texas networks around Tenet and Baylor Scott & White OS30. Where a partner gave it low rates, Oscar posted the cheapest plan: lowest-priced in San Antonio, with about 43,000 Texas members in a year, against about 5,000 in Los Angeles, where it priced above Anthem OS30.
The rest of Oscar supported the trade-off. Concierge teams and search steered members to the partner system, which made a narrow network acceptable and gave the hospital a reason to accept lower rates or share risk, as the Cleveland Clinic did in a 50/50 arrangement OS31 OS58. By Schlosser's account the medical loss ratio, the share of premium paid out in claims, fell from 120% in 2016 to 95% in 2017 OS58. In early 2018 Oscar posted its first quarterly profit, with more than 250,000 members OS27.
Rivals Bright Health built the same model: Centura's network was its exclusive provider in Colorado for 2017, and by 2018 it sold with one Care Partner health system per market in six states BH29 BH11. By 2022 it had the cheapest silver plan in cities such as Birmingham and Atlanta, where Oscar matched its $405 price BH41.
Trade-offs & activity fitPenetration Pricing
Step 3 of 5 · 2017–23
Grow where costs are known; retreat elsewhere
Oscar built the business for one segment, people comparing prices for their own coverage on the exchanges, and left markets where its model did not fit. For 2017 it quit New Jersey, where it lacked a narrow network, and Dallas OS47 OS5. Broader lines followed: small-business plans in 2017, Cigna + Oscar small-group plans and Medicare Advantage in 2020, and +Oscar, its technology sold to providers and insurers OS4. None reached scale while the exchange book lost money; in 2022 Oscar lost $609.6 million on $3.96 billion of revenue OS37.
That November it left Medicare Advantage, where it had 4,577 members, and paused new +Oscar deals; Schlosser said Oscar wanted "to be focused" on ACA plans OS7. It paused new Florida sales for 2023 OS20. Mark Bertolini, a former Aetna chief executive, took over in April 2023, paused California and cut costs OS42 OS24, and 2024 brought Oscar's first annual profit, $25.4 million on $9.2 billion of revenue OS11. Paying its own claims (step 1) let Oscar see which markets lost money in time to leave them.
Rivals Bright Health expanded instead, from 8 to 18 of the Urban Institute's 58 sampled markets between 2020 and 2022, adding Medicare Advantage and clinics of its own BH41 BH28. Its 2021 medical cost ratio was 101.3%: claims alone exceeded premiums BH24.
Focus StrategyResource allocation
Step 4 of 5 · 2020–25
Grow on reinsurers' capital (quota share)
An insurer must hold regulated capital against the premiums it writes; at launch, $29 million of Oscar's $40 million round had to stay in reserve OS16. As it grew, Oscar shared its book through quota share: a reinsurer takes a fixed share of the premiums and pays the same share of the claims, so part of the capital requirement sits on the reinsurer's balance sheet. By 2020 its quota share partners were Axa France Vie and Berkshire Hathaway Specialty Insurance Company, and it ceded approximately 55% of its business in 2019 and 77% in 2020 OS4.
In 2022 Oscar ceded $1.49 billion of premium and $1.24 billion of claims to reinsurers OSX-8, and it expected to cede about half its premium for 2025 OSX-9. For 2026 it put the ceding percentage at roughly 55%, in a regulatory rule of thumb of about $50 million of required capital for every $1 billion of premium OS15. Membership could grow faster than Oscar's own equity, and the $1.4 billion raised at its 2021 IPO stayed a cushion: Oscar ended 2022 with $1.56 billion in cash OS26 OSX-8. The retreats in step 3 kept losses small enough for that cushion to last.
Rivals Bright Health paid for its growth with equity: $924.3 million at its June 2021 IPO and $750 million of preferred stock from Cigna and NEA in January 2022 BH6 BH5. By August 2022 a statutory filing raised substantial doubt that it could continue as a going concern, and leaving the exchanges freed about $250 million of regulated capital OSX-7 BH35.
Step 5 of 5 · 2022–26
Survived to take share as rivals exited
Oscar was still selling when enhanced federal subsidies from 2021 nearly doubled exchange enrollment, to over 21 million in 2024 OS55. By 2025 the government paid 93% of Oscar's premiums, and "the vast majority" of members came through independent brokers, whom insurers pay per enrollment OS8 OS57. A broker favors a plan that is cheap for the client and easy to service, which Oscar's narrow-network prices supplied. Oscar fought Florida Blue's broker exclusivity in court and lost OS18 OS36.
Bright left every exchange for 2023, releasing about 970,000 members OS22. Oscar ended 2022 with 1.15 million members OSX-8, grew 37% against the market's 13% in the 2025 open enrollment OS11, and lifted its share of its service area from 17% to 30% as competitors pulled back OS15. When the subsidies expired and Florida's market shrank 13% in 2026, Oscar grew 52% there, to about 40% of the state, and about three million members overall OS14 OS13.
Survival won Oscar the members, not a way to keep them. Buyers re-shop every year through brokers who also sell other carriers, and Aetna and Cigna took the lowest price in many markets the insurtechs left OS8 OS20. Risk adjustment, the ACA program that moves money from plans with healthier members to plans with sicker ones, takes back much of the gain from low-cost members: it took 18.5% of Oscar's premiums in 2025, a year Oscar lost $443.2 million OS15 OS1. Steps 3 and 4 decided the contest with Bright: Oscar pulled out where it lost money and grew on shared capital, so it was solvent when the members became available.
Rivals Bright Health owed $1.9 billion of 2022 risk-adjustment charges, its Texas insurer went into liquidation in November 2023, and NEA took the company private at $7.33 a share in 2025 BH38 BH19. Centene and Florida Blue lost Florida members in 2026 while the gains went to Oscar OS14.
Channel Partners