Vijay Pande on Leaving a16z's $4B Fund to Bet Small on AI and Biology
Summary
Vijay Pande left a16z's roughly $4 billion biotech practice to launch the smaller, AI-native fund VZVC. He argues that biology is transitioning from a discovery science to an engineering one, and that open, shared datasets — not proprietary ones — are what will truly enable AI to transform medicine.
Vijay Pande, one of Silicon Valley's most seasoned biotech investors, made a surprising move when he departed from andreessen horowitz (a16z) and its nearly $4 billion biotech fund. In its place, he founded VZVC, a leaner, AI-native venture firm with a fundamentally different investment philosophy. 'We're not doing 30 bets a year,' Pande explains, signaling a deliberate shift away from the wide-net approach that characterizes many large funds in favor of deeper, more focused conviction bets.
At the heart of Pande's thesis is a bold claim about biology itself: it is undergoing a historic transformation from a 'discovery' science — driven by serendipitous findings — into an 'engineering' science, where outcomes can be systematically designed and predicted. Advances in AI and machine learning are making it possible to model biological systems with unprecedented precision, opening the door to a new era of medicine that is faster, smarter, and more targeted than anything we have seen before.
Yet Pande is candid about the obstacles that remain. Clinical trials, he acknowledges, are still brutally expensive and slow — a bottleneck that no amount of algorithmic sophistication can fully bypass. Regardless of how powerful AI tools become, new therapies must still be validated in human beings under rigorous regulatory oversight. This reality demands patience and long-term thinking from investors, and it is one reason Pande prefers making fewer, higher-conviction bets rather than spreading capital thin across dozens of companies.
Perhaps most strikingly, Pande challenges the conventional wisdom that proprietary data is the ultimate competitive moat in biomedical AI. He argues passionately that open, shared datasets are what will actually unlock AI's transformative potential in medicine. When researchers and companies pool their data, AI models can be trained on broader, more diverse, and more representative samples — ultimately producing better, fairer, and more generalizable medical tools. It is a vision that puts collaboration above competition, and one that could reshape the economics of the entire sector.
Need IT help in Stockholm?
Book Sovin IT from 499 SEK