a16z Podcast: The Science and Business of Innovative Medicines

Episode Summary

Executive Summary: In this A6NZ conversation, Novartis CEO Voss talks about how pharma is being reshaped by disciplined portfolio focus, rising scientific complexity, and new platforms like cell and gene therapy, RNA, and digital therapeutics. He argues that success now requires early reimbursement thinking, strong data foundations, operational rigor, and an engineering mindset that can turn biology into scalable platforms.

Main Topics: Portfolio strategy and selective exits: Novartis is narrowing focus to medicines powered by advanced therapy platforms and data science, exiting areas where it cannot be a long-term leader or is subscale. R&D productivity, attrition, and the economics of drug development: Despite scientific progress, attrition remains stubbornly high and costs keep rising; discipline and clear standards are needed to avoid crowding out stronger programs. Emerging modalities: cell/gene therapy, RNA, and regenerative medicine: The discussion covers CAR-T, gene therapy, RNA interference, xenotransplantation, and tissue regeneration as areas moving from science fiction toward real therapeutic categories. Technology, data science, and AI in pharma operations: AI is most effective where data is clean and structured, such as pathology imaging, trial operations, and forecasting; broader unstructured-data promises remain difficult. Clinical trials, reimbursement, and market access: The CEO argues that reimbursement should be considered early in development, not only after FDA approval, because it is the true proxy for value. Build vs. buy, partnerships, and talent ecosystems: Novartis balances internal R&D with acquisitions and partnerships, keeps some acquired tech independent, and builds distributed talent hubs to attract data and digital expertise. Culture shift toward agile, unbossed execution: The company is trying to operate more like a knowledge organization and less like a traditional industrial giant, emphasizing curiosity, empowerment, and rapid failure.

Key Arguments: Pharma companies must concentrate capital and talent in areas where they can sustain long-term leadership, and should be willing to exit subscale therapeutic areas. Scientific progress has accelerated, but the industry's human success rate has not meaningfully improved; bringing a molecule into humans still succeeds only about 5%-10% of the time. Because me-too products are increasingly unattractive, Novartis now requires candidates to show they can replace the standard of care rather than offer incremental benefit. Reimbursement is a value test, not just a late-stage commercial hurdle; founders should work backward from payer value, not only from FDA approval. Technology can materially reduce clinical trial costs, but pharma has been slow to adopt it because the industry is high-margin and complexity has historically been tolerated. AI works best where there is high-quality, structured data and clear outcomes; its promise is strongest in imaging, operational forecasting, and finance rather than vague discovery claims. Cell and gene therapy are becoming platform businesses, where core manufacturing and delivery infrastructure can support multiple products across diseases. Partnerships with startups and universities are essential because large pharma lacks the agility and external talent flows needed to build all capabilities alone. A company like Novartis must combine open debate with disciplined decision-making so scientific conviction does not turn into personal attachment or NIH syndrome.

Data Points: Novartis yearly doses: 70 billion doses a year - Described as the scale of Novartis medicines across therapeutic areas Patients reached: about 1 billion patients a year - Reach of Novartis, largely through its generics/access business Generics share of doses: roughly 80% - Approximate share of the 70 billion doses that are generics Internal R&D employees: 17,000 R&D people - Size of Novartis R&D organization Annual R&D spend: $9 billion plus a year - Novartis investment in research and development Company size: 120,000 people in 150 countries - Scale of Novartis as a global organization Research hubs: 3 main research hubs - Cambridge, Basel, and Shanghai Development centers: 5 major centers mentioned - Cambridge, Basel, Shanghai, Hyderabad, and East Hanover Clinical trial volume: 400 to 500 clinical trials a year - Scale used to train operational AI models Trial patient base: 2 million patients in clinical trials in the last 10 years - Dataset used for imaging and analysis efforts Human success rate: about 5% to 10% - Likelihood that a medicine entering humans will work Company success rate: 8% to 10% - Novartis internal average mentioned for that metric Industry success rate: about 5% - Industry-wide average success once a drug reaches human testing Potential trial cost savings: 20% - Estimate cited for possible clinical trial cost reduction through technology at scale 2018 deal activity: around $50 billion - Portfolio reshaping transactions used to reposition the company Recent acquisitions in advanced therapies: $15 billion - Acquisitions in the prior year in advanced therapy spaces Millennial workforce share: 50% - Approximate share of knowledge workers who are millennials FDA new molecular entities since creation: about 1,500 - Used to emphasize how hard drug discovery is Druggable proteins: 1,200 - Approximate number of druggable proteins referenced

Pivotal Quotes: "you have to be prepared to exit, I think, areas where you're going to be subscale" — Voss: On portfolio discipline and deciding which therapeutic areas to leave "we say to ourselves, it has to replace the standard of care" — Voss: On the bar Novartis now sets for innovative medicines "Software is a drug." — Vijay/host discussion: On digital therapeutics and the expanding definition of medicines

Implications: The episode suggests pharma’s future belongs to companies that combine platform biology, data discipline, and commercial realism. For startups and investors, the winning play is clear value, scalable infrastructure, and partnerships that fit the new engineering-driven model.