Vijay Pande, who previously led a massive four-billion-dollar biotech investing team at the famous venture firm Andreessen Horowitz, has left to launch a much smaller investment firm called VZVC. In a recent interview, he explained why he is deliberately keeping his new fund small and focusing heavily on artificial intelligence. The move signals a shift toward quality over quantity in health technology investing.
Pande believes biology is changing from a field of slow discovery into one of fast engineering. In the past, scientists mostly stumbled upon findings through years of painstaking trial and error, often relying on luck to find useful molecules. Now, he says, advanced AI can help design experiments and predict outcomes, turning drug development into something closer to building software or hardware. However, he warns that clinical trials remain brutally expensive and time-consuming, so even the best AI cannot instantly make medicine cheap.
A key part of his vision is data sharing. Pande argues that the biggest medical breakthroughs will come from open datasets that researchers everywhere can use, rather than from information locked inside private company vaults. He thinks walled-off data slows progress, while shared data lets AI models train on far more examples and become more useful for patients everywhere. This philosophy could reshape how pharmaceutical research gets done.
By running a smaller fund, Pande plans to make fewer but more careful bets. He says he will not invest in thirty companies a year. Instead, he wants to work closely with a handful of startups that truly understand how to merge AI with biology. This approach reflects a growing belief in Silicon Valley that the next wave of health breakthroughs will come from specialized, AI-first companies rather than traditional pharmaceutical giants.
If his strategy works, it could speed up the creation of new treatments and lower costs over time. But success will depend on whether startups can actually deliver on the promise of engineering biology with code, and whether the industry agrees to share the data needed to train the next generation of medical AI. Patients and doctors will be watching closely to see if this model produces real cures faster than the old system.