US drug manufacturing is in the middle of a building boom driven largely by tariff policy, but most of the plants announced in 2026 will not make a single dose until 2027 at the earliest, and some not until 2034. That gap between a pledge and a producing factory is the most useful thing to understand when reading the headlines. This explainer sets out what the tariffs actually are, which projects were announced, what each will make, and where the supply-chain data say the real weak points sit.

What the tariff policy is, as reported

On April 2, 2026, the administration reinstated tariffs on imported pharmaceuticals at 100% under Section 232, according to BioPharma Dive. The headline rate comes with large exemptions:

  • Drugs from the EU, Japan, South Korea, Switzerland and Liechtenstein, which have trade deals with the US, face 15%. The UK gets a lower rate under a separate agreement.
  • Companies with pricing agreements (most-favored-nation terms) and US manufacturing commitments are exempt. The White House listed 13 at that point, with combined pledges of $400 billion.
  • All generic and biosimilar drugs and their ingredients are exempt, and the administration said it would reassess that in a year.
  • Duties take effect in 120 days for large companies and 180 days for smaller ones, and the most-favored-nation exemption expires January 20, 2029.

By October 6, Fierce Pharma counted 26 large and mid-sized drugmakers that had struck such deals, trading investment and pricing commitments for tariff immunity. The same article reports that the device maker BD pledged $19 billion over several years, $3 billion of it for US manufacturing sites, in exchange for relief from future Section 232 tariffs on its products.

Opinions differ on the policy. PhRMA opposes the tariffs, and RBC analyst Trung Huynh called the threat to the sector low relative to investor expectations while warning that investments could become stranded if a future administration reverses course. I report these as stated, without taking a side.

The projects, and when they would actually produce

The table lists the October 2026 announcements and a few earlier ones, with the timelines as reported.

Company Site Amount What it makes Timeline
Bayer New Albany, Ohio $2.2B Oncology, cardiovascular and renal products Groundbreaking mid-2027; drug substance 2031, drug product 2034
GSK Upper Merion, Pennsylvania More than $800M Biologics plant, pilot plant, quality lab Construction begins 2027
Lilly Lebanon, Indiana Extra $4.5B API for Zepbound, Mounjaro, Foundayo and retatrutide API factory opens 2027
Resilience with Lilly Cincinnati area $750M Injectable device assembly, including KwikPen Full operations early 2027
AbbVie Durham, North Carolina $1.4B Small-volume injectables Site chosen April 2026

The Bayer, GSK and Lilly rows are from Fierce Pharma (Bayer, October 2), Fierce Pharma (GSK, October 9) and BioPharma Dive (Lilly, May 8). Lilly’s Indiana commitments now exceed $21 billion over six years, and Lilly says the API factory will be the largest in US history; that superlative is the company’s CEO’s claim. The Resilience figures come from BioPharma Dive (July 30), which also notes the two companies have produced more than 150 million doses in vial and prefilled-syringe formats since 2023. AbbVie’s row is only from a headline on BioPharma Dive’s manufacturing page, so I give no further detail.

If you read the timeline column closely, you see the pattern: 2027 for the earliest, 2031 and 2034 for module-by-module openings. Large biologics and drug-substance plants are built in stages, and each stage needs inspection before it can ship. Chemistry World made the same point about obesity-drug capacity, reporting that building, installing, validating and qualifying a facility can take three to five years.

The same Lilly report adds that the three newer weight-loss drug facilities, in Houston, Huntsville (Alabama) and Lehigh County (Pennsylvania), are due to be operational by 2031 at the latest, and that more than $16 billion has been committed to them since September. The Indiana advanced-therapies facility, which Lilly said opened that week, is the one item on the list already built. In other words, a single company’s own schedule runs from this year to the end of the decade, and the total of its US capital commitments is not given in the article I read.

Why a plant takes years, not months

Bayer’s schedule is a good illustration. Groundbreaking is expected around mid-2027, a drug-substance module is slated for 2031 and a drug-product module for 2034. Drug substance is the active ingredient made in bioreactors or reactors; drug product is that ingredient formulated and filled into vials, syringes or tablets. Each is a separate regulated operation.

After construction comes equipment installation, validation runs, and regulatory inspection before a facility can ship. Tariff deadlines, by contrast, run in months: the April announcement set effect dates of 120 days for large companies and 180 days for smaller ones. A company exempted through a pledge does not need the plant to exist yet, only the commitment, which is why the announcement date and the production date are so far apart.

Who is exposed and who is not

On the reporting above, exposure splits by company type. Large drugmakers with pledges and pricing deals are largely shielded. Trade-deal regions pay 15%. Generics and biosimilars are exempt for now. That leaves smaller firms that import branded drugs from outside the deal regions, which is why the Mid-Sized Biotech Alliance of America is lobbying for reversal and why RBC’s analyst named smaller companies pursuing large indications as most vulnerable.

The reporting does not quantify how many drugs or how much revenue falls outside every exemption, so I cannot give a share. Nor have I seen a count of how many announced plants are new capacity and how many are existing sites being upgraded. GSK’s Pennsylvania project, for example, involves demolishing and rebuilding at an existing location and is expected to add 300 jobs while helping retain about 4,500.

Where contract manufacturers fit

Not every plant belongs to a drug brand. Contract development and manufacturing organizations (CDMOs) build capacity for hire. Resilience’s Cincinnati expansion is one example, adding at least 400 jobs for a total of more than 1,400 across two facilities.

The CDMO business is not uniformly healthy, though. BioPharma Dive reported on August 24 that Cellares, a cell-therapy manufacturer, would lay off 100 people after an unnamed large pharma customer ended a contract, about two months after a $327 million Series D. The company says it has more than doubled its customer count this year. The same article says Evotec announced 800 job cuts in March. Demand for sterile injectables is strong; demand for cell-therapy contract capacity is more mixed.

What the supply-chain data say

The plants above mostly make branded, higher-margin products. The medicines most exposed to shortages are often old generics. The new National Fragility Index, launched by the API Innovation Center with QYOBO and covered by Fierce Pharma on October 7, examined more than 400 drugs. Fourteen were rated severe and 71 high for supply-chain vulnerability, roughly one in five. The study says more than 550 million prescriptions filled in 2023 were for drugs rated high or severe, and almost six in ten of the drugs analyzed had at least one shortage that averaged over two years.

Two examples show the pattern. Pantoprazole, rated severe, had only 5.6% of domestically dispensed supply from a US finished-dose manufacturer. Metoprolol, rated high, had no domestic API source found by the index. This is a new index from an industry-linked organization, so treat its ratings as one assessment and not an official count.

The generics question

Generics are exempt from the April tariffs, but in a July 21 social-media post the president threatened 100% duties on imported generics starting August 1, 2028, rising to 200% afterward, according to BioPharma Dive. The article describes this as a threat, with no details on legal basis or whether ingredients would be covered. It also notes that a 2025 Brookings report warned tariffs could lead to discontinuations and shortages. Apotex shares fell 9% the next morning, Dr. Reddy’s 7% and Sandoz ADRs 3%. Whether anything is implemented is unknown as of this writing.

The economics matter. A generic sells on thin margins, so a plant costing hundreds of millions is hard to justify unless prices or volumes change. That is an inference from the structure of the business, not a finding of the sources, and the Association for Accessible Medicines pointed to structural problems in generic purchasing and reimbursement.

How to read the next pledge

  1. Look at the date column. A groundbreaking in 2027 and operations in 2031 is not capacity now.
  2. Ask what is made. Drug substance, drug product and device assembly are different steps with different bottlenecks.
  3. Check whether the money is new. Pledges often repackage earlier plans; GSK’s $800 million sits inside a $30 billion US commitment it announced last September.
  4. Separate branded from generic supply. The shortage data point to older generics, not to the products getting the biggest plants.
  5. Remember policy can change. The most-favored-nation exemption ends in January 2029.

This is general information about industry developments, not investment or policy advice.

Sources