Last reviewed · How we verify
How much does it cost to develop a drug?
The headline number: $2.6 billion
The most-cited figure is the Tufts Center for the Study of Drug Development estimate published in 2016: $2.558 billion (in 2013 dollars) for a single drug to reach approval. The number includes (a) out-of-pocket costs, (b) the cost of failures (most candidates don't reach approval), and (c) cost of capital (capitalising at 10.5% over the development period). Without the cost-of-capital adjustment, the figure falls to $1.4B. Without including failures, it falls further to ~$300-500M per "successful" drug.
Why estimates vary so widely
Lower estimates (DNDi $130-250M; Light + Warburton $43-125M for some classes) exclude either failures, the cost of capital, or both. They also use different denominators — drugs that reach Phase 1, drugs that reach Phase 2, drugs that reach approval. The choice of base population dramatically changes the per-drug cost.
Component costs
Preclinical research: $100-300M (3-6 years). Phase 1: $5-20M (~$50K per patient, 50-100 patients). Phase 2: $20-100M ($30-50K per patient, 100-500 patients). Phase 3: $100-500M+ ($40-60K per patient typical, but can be $200K+ for oncology with PFS endpoints). NDA/BLA filing: ~$4M user fee + regulatory consulting. Post-approval Phase 4: variable.
Why oncology costs more
Phase 3 oncology trials cost $200-500M+ because: longer follow-up (survival endpoints), more imaging, more biomarker testing, more comparator drug supply, and bigger sample sizes. By contrast, antibiotic Phase 3 trials run $50-100M. Rare-disease orphan trials can be cheaper ($20-80M) due to small patient populations but the per-patient cost is higher.
Implications for pricing
Pharma companies argue high US drug prices reflect the high cost of R&D + the failure rate. Critics note that (a) much R&D is funded by NIH + academic grants before pharma's involvement, (b) marketing spend often exceeds R&D, (c) other countries pay less for the same drugs. The cost-of-development debate is foundational to drug-pricing reform discussions.
FAQ
Is the $2.6B figure credible?
It's the most cited but contested. The Tufts methodology is sound but the inputs (which trials, which failures, what capital cost) are debatable. Independent estimates range from $300M to $3B+.
Does the government pay for drug development?
NIH funds ~$45B/year in biomedical research, much of which underlies later commercial development. A 2023 NIH study found that NIH contributed to 99.4% of new drugs approved 2010-2019. But the final clinical-trial costs are almost always borne by industry.
Why are rare-disease drugs so expensive?
Small patient populations mean R&D costs must be recouped from fewer prescriptions. A drug for 5,000 patients globally needs to charge $200K+ per patient per year to recover even modest R&D investment.
Is drug development getting cheaper or more expensive?
Per-trial costs have risen ~10% annually due to greater protocol complexity, more biomarker testing, and longer follow-up. But platform improvements (single-cell sequencing, AI-driven target ID, biomarker stratification) may flatten this in the next decade.