Bazaz, Ji, Socal, and Kang analyze 28,269 VC deals in biopharmaceutical companies from 2010 to 2024 using PitchBook data, finding just 5.93% of deals, 1,679 total, involved companies disclosing AI as a business focus, with drug discovery the most common AI business area at 38.06% of AI deals, followed by biotechnology and pharmaceutical at 37.22%, discovery tools at 23.88%, and drug delivery at under 1%. AI-related companies were substantially overrepresented in discovery tools relative to non-AI companies, 23.88% versus 3.31% of deals, and underrepresented in drug delivery, and both AI and non-AI VC activity skewed increasingly toward early-stage deals by 2024, 77.42% and 68.21% respectively, though total capital invested grew more evenly across deal stages. AI discovery tool companies saw the fastest growth of any category, with median deal size rising from $0.19 million to $7.50 million between 2010 and 2024 and their share of AI capital jumping from 0.25% to 21.85%, while geographic concentration remained stark: nearly 60% of 2024 US-based AI deals landed in California and Massachusetts alone, versus non-AI deals spread across 42 states.

Founders building AI-driven drug discovery tools get a precise, quantified benchmark for the category's explosive growth, a nearly 36-fold increase in funding share from 2010 to 2024, along with a concrete median deal size trajectory to compare their own fundraise against rather than relying on anecdotal impressions of a hot sector. State policymakers and economic development officials get a specific, evidence-based signal that California and Massachusetts have captured a disproportionate share of AI biopharma investment, nearly 60% of 2024 deals, which the authors explicitly frame as a call for other states to consider emulating those states' research incentives and regulatory pathways to compete for this capital. The paper's methodological caution around AI-washing, companies using the AI label as marketing rather than a genuine core focus, is also a useful reminder for anyone reading VC funding statistics generally, since it flags a real measurement challenge that likely affects many other reported AI investment figures beyond just biopharma.