FDA approves 150 novel drugs in a year

Where things stand

• 2026 so far: FDA's Novel Drug Approvals for 2026 table, read on 2026-10-12, lists 46 CDER novel approvals. The latest is Hovilpri on 10/9/2026. BioCentury (Oct 9, 2026) reported 45 through September, with September's 9 the best month of the year. At that pace 2026 ends around 57–62, which could edge past the record. Reaching 150 would take 104 more approvals in about 11 weeks, so 2026 is effectively ruled out.
• History: I counted rows in FDA's year tables: 2021 = 50, 2022 = 37, 2023 = 55, 2024 = 50, 2025 = 46. CDER's all-time high is 59 (2018), and the count has never come close to 100. Hitting 150 means roughly 2.5 times the record.
• Filings set a ceiling: The FY2025 PDUFA Performance Report (fda.gov/media/191112/download) shows new-molecular-entity drug applications and original biologic applications filed in FY2024 at 73 (40 priority + 33 standard). FY2025 had 69 with a review priority, plus 19 that had none yet, some of which are not new molecules. These figures cover CDER and CBER together. About 150 approvals in a year would need filings to more than double, and for drugs approved in 2027–2031 that pipeline mostly already exists.

Base rate (computed)

I fit a log-linear trend to CDER's annual counts for 2005–2025 (values from FDA tables and reports).

• Full period: growth about 5.3% a year, residual standard deviation 0.225 in log terms, projected median about 79 in 2031.
• 2014–2025 only: growth 2.5% a year, projected median about 60 in 2031.
• Tail check: A fat-tailed (Student-t) version puts about 0.2–0.5% a year on reaching 150 by 2030–31, roughly 1% summed over 2027–2031. That overstates the upside: the large historical misses are low years such as 2016's 22, while the best year (2018) beat trend only modestly.
• Other forecasts: The only related crowd forecast I found, Metaculus "Drugs approved by FDA in 2035", sits around 69 with just four forecasters.

Ways it could resolve YES, and why each looks unlikely

1. FDA reforms. Recent changes include a single pivotal trial as the default (Feb 2026), the national priority vouchers that promise 1–2 month reviews, and an expedited pilot for starting clinical trials. These are incremental and might add a few approvals a year. The "plausible mechanism" draft guidance for individualized therapies puts many genetic variants under one application, so it would not create dozens of separate entries. Much of it covers gene editing, which CBER regulates and this series excludes. Leadership is settling under Michael Davis, now permanent CDER director, and nominee Heidi Overton. Analysts don't expect her to remake the agency.
2. Legislation. A reciprocal-approval bill from Rep. Pete Sessions was introduced in March 2026. A conditional or provisional approval law based on Phase 2 data could also pull many drugs forward at once. Neither has passed. Even if one did, it would have to produce more than 100 extra CDER entries in a single year.
3. Artificial intelligence. As of August 2026 no AI-discovered drug had FDA approval. AI's gains so far are mostly before human testing, while trial length, enrollment and review capacity are the real limits. Only a fast, sweeping AI transformation plus a radical change in regulation looks able to triple approvals by 2030–31.
4. Counting changes. Moving CBER products into CDER would add only about 10–20 a year. Totals pooled across years, fiscal-year totals and FDA-wide totals don't count.

My rough split: about 0.3–0.5% from non-AI paths, and about 0.5–1.5% from an AI-driven transformation (roughly a 5–10% chance of such a transformation by about 2030, times a 10–15% chance it produces 150 in one year). I put the total at about 1.6% by end-2031.

How the series is built

I set a chance for each year being the first to reach 150: 2026 = 0.03%, 2027 = 0.12%, 2028 = 0.22%, 2029 = 0.33%, 2030 = 0.42%, 2031 = 0.52%. The probability rises over time because pipelines and any AI effects need years to show up. Within a year I split it 10/20/30/40% across Q1–Q4, since a year's total crosses 150 late unless it far exceeds that. Cumulating gives about 0.15% by end-2027, 0.37% by end-2028, 0.7% by end-2029, 1.1% by end-2030 and 1.6% by end-2031.

Strongest case against this forecast: If transformative AI arrives around 2027–28 and Washington rebuilds drug approval around AI evidence or provisional approvals, a large Phase 2 pipeline could be approved in bulk. That is why the later years are not set at zero. Nothing in current filings, policy or leadership points that way.

Sources: FDA 2026 table (https://www.fda.gov/drugs/novel-drug-approvals-fda/novel-drug-approvals-2026, accessed 2026-10-12); FDA 2025 and 2021–2024 tables; FY2025 PDUFA Performance Report (https://www.fda.gov/media/191112/download); BioCentury, Oct 9, 2026; BioSpace H1 2026 review (Jul 6, 2026); BioPharma Dive on the Davis/Mikhail appointments (Sept 2026); FDA "Accelerate and Modernize Clinical Development" page (Sept 2026); Inside Precision Medicine and BioSpace on the plausible-mechanism guidance (Feb 2026); BioPharma Dive on the single-pivotal-trial default (Feb 19, 2026); Clinical Trial Vanguard on AI drug approvals (Aug 8, 2026); FDA Group news roll (Apr 2026) on the Sessions reciprocity bill.

Question. By the last day of each quarter, has FDA's Center for Drug Evaluation and Research (CDER) publicly reported approving 150 or more novel drugs in a single calendar year?

Definitions.

• Novel drugs: the products counted in FDA's Novel Drug Approvals series, formally headed “New Molecular Entity and New Therapeutic Biological Product Approvals (CDER only).” FDA's classification controls; this is not a count of all new drug applications, all FDA approvals, or new chemical entities eligible for exclusivity.
• Count: count each distinct product once, as FDA counts it in that series. A combination product or co-pack listed as one entry counts as one, not once per ingredient. Diagnostic products and accelerated approvals count if included in the series. No evidence of AI involvement, curative effect, or commercial launch is required.
• Single calendar year: approvals dated January 1 through December 31 of the same year, starting with 2026. A published year-to-date count of at least 150 qualifies without waiting for year-end. The Observation period convention applies; the reported annual or year-to-date period must end after the opening date. Approvals earlier in 2026 can contribute to an eligible 2026 count.

Resolves YES if an eligible FDA publication reports at least 150 novel CDER drug approvals for one calendar year, or its official year-specific table lists at least 150 distinct qualifying products approved in that year. A later FDA revision raising an eligible year's count to 150 qualifies on the revision's publication date. A subsequent downward revision or withdrawal of an approved drug does not reverse a YES, under the Resolution window convention.

Resolves NO if no qualifying publication is available by the quarter-end. Exclude forecasts, annualized approval rates, fiscal-year or rolling-twelve-month totals, and totals pooled across years. Exclude products approved by the Center for Biologics Evaluation and Research (CBER), generic approvals, biosimilars, emergency-use authorizations, and supplemental approvals or new formulations not counted as novel products in the CDER series. Do not count conversion of accelerated approval to traditional approval as another novel approval.

Resolving source. FDA-authored publications only: the year-specific Novel Drug Approvals for [year] tables linked from Novel Drug Approvals at FDA, CDER's annual Advancing Health Through Innovation: New Drug Therapy Approvals reports, or an FDA statement explicitly reporting the same series. The 2026 table identifies the current format. Renamed successor publications count only if they retain an identifiable, comparable CDER novel-drug series; broader FDA-wide totals do not. Apply the Resolution window convention and Resolution date (A3): use the publication or table-update date that first establishes the qualifying count, not the approval date of the 150th drug. For a table update without a verifiable publication date, use the earliest dated public archive capture establishing the qualifying contents. The resolving-source restriction is narrower than Public confirmation (A2).