The $35 monthly cap on out-of-pocket insulin costs in Medicare Part D took effect in January 2023, followed by similar caps voluntarily adopted by the three major insulin manufacturers in commercial markets in 2024. The political achievement was substantial. The patient affordability problem, which had produced harrowing case reports of insulin rationing and preventable diabetic emergencies, was meaningfully reduced. The market structure that originally produced the prices—the rebate-driven formulary architecture, the patent thicket strategies, the parallel pricing of essentially equivalent products—remained largely intact. Whether the cap framework provides a template for addressing other drug pricing crises depends on which problem one understood the insulin episode to be solving.
The insulin pricing trajectory between roughly 2007 and 2019 was one of the most dramatic in modern pharmaceutical history. The list price of Lantus, Sanofi’s flagship long-acting insulin, increased from approximately $96 per vial in 2007 to over $283 per vial in 2018—a roughly threefold increase for a product whose underlying manufacturing cost had not materially changed and whose clinical performance had not measurably improved. Similar trajectories applied to Eli Lilly’s Humalog and Novo Nordisk’s NovoLog. The prices were, in pharmaceutical terms, transparently anomalous—old products with extensive prior price history that should have stabilized or declined had instead accelerated upward in lockstep across three competing manufacturers.
What produced this trajectory was the rebate architecture, which over the same period had grown from a marginal feature of pharmaceutical pricing into the dominant determinant of net price for established products. The three insulin manufacturers competed for formulary placement on the basis of rebate yield to the major PBMs. The PBMs preferred higher list prices because the rebate flowed proportionally; the manufacturers raised list prices in step with each other because the rebate competition required it; and the patient cost-sharing, calculated on list price, rose along with everything else. The mechanism was, in retrospect, an almost perfect demonstration of how pharmaceutical pricing can decouple from underlying value when the parties to the transaction are not the parties bearing the cost.
The political response to the insulin crisis took several forms simultaneously. State-level price caps began with Colorado in 2019 and expanded to over twenty states by 2023. Federal action came through the Inflation Reduction Act’s $35 cap for Medicare beneficiaries. The manufacturers, facing accumulating reputational damage, announced their own price reductions in early 2023—Eli Lilly cut Humalog’s list price by 70 percent in March 2023, with Sanofi and Novo Nordisk following with similar announcements. The combination of state caps, federal action, and manufacturer-led price reductions reduced the patient affordability problem substantially within roughly eighteen months.
What the manufacturer-led list price reductions accomplished, beyond the patient cost reduction, was a partial dismantling of the rebate architecture for insulin specifically. When Lilly cut Humalog’s list price by 70 percent, the rebate the PBMs collected on Humalog necessarily fell. The PBMs, who had been receiving the higher rebates as a meaningful component of their compensation, lost that revenue. The Federal Trade Commission’s lawsuit against the major PBMs over insulin pricing, filed in September 2024, alleges that the PBMs’ rebate-driven exclusion of lower-priced insulins contributed to the original price escalation, and that the post-cap environment has produced corresponding reductions in PBM revenue from this drug class. The PBMs have disputed this characterization, but the structural point—that the rebate architecture that produced the prices was disrupted by the price-cap response—is broadly accurate.
There is a complication in the insulin story that the public conversation has incorporated unevenly. The biosimilar insulin pathway opened in 2020 when the FDA’s regulatory framework was updated to permit biosimilar versions of insulin products. Several biosimilar insulins have entered the market since, with prices substantially below the originator products. Adoption has been slower than the price differentials would suggest, partly because the cap framework has reduced patient incentive to seek lower-priced alternatives, and partly because the formulary placement of biosimilars has been complicated by the same rebate dynamics that originally distorted insulin pricing. Whether the biosimilar insulin market eventually matures into meaningful competition depends on whether the rebate architecture for insulin reasserts itself in modified form.
The CMS price negotiation under the IRA includes Fiasp/NovoLog among the first round of negotiated drugs, with the negotiated price effective January 2026. The negotiation produced a list price reduction of approximately 76 percent against the 2023 list price. Whether this represents meaningful additional savings beyond the manufacturer-led price reductions and the cap framework, or whether the negotiation simply locks in changes that were already happening through other mechanisms, has been actively contested. The CBO’s scoring of the insulin negotiation savings has been criticized by both supporters and critics of the IRA framework as either overstating or understating the genuine savings depending on which counterfactual one uses.
What the insulin episode has revealed about the broader drug pricing reform conversation is that price caps, applied to specific products with specific patient affordability problems, can resolve those problems quickly and with broad political support. Caps have not, however, addressed the underlying market dynamics that produced the original pricing distortions. The three insulin manufacturers retain their dominant market positions. The PBMs remain the dominant intermediaries. The rebate architecture has been modified for insulin specifically but remains operative for most other drug classes. The next pharmaceutical pricing crisis will, in all likelihood, emerge from the same architecture in a different drug class, and the response will, in all likelihood, be product-specific caps that resolve the immediate political problem without addressing the underlying structure.
There is a separate question about whether the cap framework has perverse effects on innovation in drug classes where it applies. Insulin is a specific case—the products are old, the manufacturers have recovered their development costs many times over, and the clinical innovation in the category has been incremental rather than transformative for some time. A price cap on insulin therefore does not meaningfully disincentivize future insulin innovation because there was not much innovation to disincentivize. Applying the same framework to drug classes with active development pipelines would have different consequences. Whether the cap framework can be limited to products that meet specific criteria—old products with mature competition, products whose pricing is documentably anomalous against historical trajectories—or whether it will expand to products where the innovation effects are more substantial is one of the open policy questions for the next several years.
The state-level cap legislation has continued to expand even after the federal action and the manufacturer-led price reductions. Several states have enacted caps that go beyond insulin to include other diabetes medications, EpiPens, and various other commonly used products. The expansion reflects the political appeal of the cap mechanism and its ability to deliver visible patient benefit without confronting the more difficult structural questions about market architecture. Whether this expansion produces incremental patient benefit or simply moves the market dynamics to other drug categories is unclear, and the analysis is not being conducted carefully because the political incentive is to enact additional caps rather than to evaluate the existing ones.
What the insulin episode finally illustrates is that pharmaceutical pricing reform can succeed at narrow patient-facing objectives without succeeding at the broader market-restructuring objectives that some reform advocates have hoped for. The patients who needed insulin and could not afford it are now, in nearly all cases, paying $35 a month and continuing their therapy. This is not a small achievement. The architecture that produced the original prices, however, persists in modified form for insulin and largely unchanged form for other drug classes. The reform that worked for insulin worked because the underlying clinical problem was severe, the patient harm was visible, the political coalition was broad, and the manufacturers eventually concluded that voluntary action was preferable to continued reputational damage. Replicating this combination for other drug classes—where the patient harm is less acute, the political coalition is narrower, or the manufacturer calculus differs—is more difficult than the insulin success suggests. The cap is a tool. The conditions that made the tool effective are not always present. And the larger structural problems remain, mostly, undisturbed.
