Vijay Pande, the Stanford chemistry professor who built Andreessen Horowitz's healthcare and life sciences practice into a nearly $4 billion portfolio over a decade, has left the firm to launch a radically different kind of venture fund. Co-founded with veteran investor Zach Werner, VZVC operates on a model that inverts traditional venture capital: instead of pursuing dozens of annual investments, the firm concentrates on approximately five highly selective bets each year, with no associate roles and heavy reliance on custom-built artificial intelligence (AI) agents to handle daily operations.

The departure signals a broader philosophical shift in how top-tier investors are approaching AI-driven biotech. Rather than competing in overheated financing rounds, Pande is positioning VZVC as a deeply hands-on partner that secures deal flow through trusted relationships and strategic patience. This concentrated, technology-augmented approach reflects his decade of observations about what actually works in bringing AI-powered therapeutics to market.

What Changed in Pande's Investment Philosophy?

When Pande joined a16z more than a decade ago, the firm had explicitly avoided healthcare and life sciences. His arrival marked a turning point. Over the next ten-plus years, he grew the practice into one of the largest healthcare venture portfolios in the industry. Yet by mid-2025, he decided to walk away from that scale to start something much smaller.

The shift reflects a hard-won lesson about how AI actually accelerates drug development. Pande explained that biology is transitioning from a model reliant on fortuitous discovery to one amenable to precise engineering. Machine learning and automated robotic measurements now enable researchers to identify therapeutic targets, accelerate drug development, and improve the notoriously expensive clinical trial phase.

"For a lot of the way drugs have been developed, there was very much a fortuitous aspect to it. What's shifted is that AI and machine learning allow computers to wrap their type of understanding around something very, very complicated, to try to figure out what targets you want your drugs to hit, for specific diseases, to be able to make those drugs, and now even to help in the clinical trials, which are the most expensive part of the process," said Vijay Pande.

Vijay Pande, Co-founder of VZVC, former Partner at Andreessen Horowitz

However, Pande is cautious about industry hype. AI efficacy is strictly bound by data quality and availability. The technology will not autonomously solve unverified scientific problems, nor will it replace the complex regulatory and commercialization challenges inherent in bringing therapeutics to market.

Why Is Biological Data So Different From Other AI Training Data?

One of the most significant hurdles distinguishing biotech from other AI sectors is data access. Unlike natural language processing, which benefits from vast, publicly accessible datasets scraped from the internet, biological data remains fragmented and proprietary. Each company typically develops isolated datasets, creating silos that mirror traditional medical specialties.

This fragmentation creates a paradox: the same AI breakthroughs that have transformed text and image generation cannot simply be applied to biology without solving the data problem first. Pande noted that biological data cannot be distilled from one model to another in the way that large language models (LLMs) can be fine-tuned and shared across applications.

Despite these constraints, Pande anticipates a market correction toward comprehensive biological atlases and open-source foundation models, which could democratize access to medical intelligence. He observed that open-source LLMs have performed very well against corporate alternatives, and he expects the same pattern to emerge in biology.

How Does VZVC's Operational Model Differ From Traditional Venture Firms?

• Investment Frequency: VZVC makes approximately five highly selective bets each year, compared to traditional top-tier funds that pursue dozens of annual investments.
• Staffing Structure: The firm has eliminated traditional associate roles, relying instead on custom-built AI agents to handle daily operational workflows, allowing principals to maintain deep, long-term involvement with each portfolio company.
• Investment Thesis: VZVC concentrates on two primary domains: artificial intelligence for healthcare delivery and AI-driven clinical trial optimization, rather than pursuing broad sector exposure.
• Founder Selection: The firm prioritizes founder integrity, strategic patience, and collaborative value creation over rapid scaling, seeking relationships expected to last five to ten years or more.

This operational philosophy reflects Pande's broader industry observations. By positioning itself as a deeply hands-on partner rather than a passive capital provider, VZVC seeks to secure deal flow through trusted relationships rather than competing in overheated financing rounds. Pande emphasized that he is looking for founders with high integrity who think long-term and are focused on winning together rather than beating competitors.

What Does This Mean for the Future of AI in Drug Development?

Pande's departure from a16z marks not a retreat from the sector, but a strategic recalibration toward precision, data-driven decision-making, and sustainable innovation in artificial intelligence and medicine. His focus on clinical trials is particularly significant: these trials represent the most expensive phase of drug development, often costing hundreds of millions of dollars and taking years to complete.

The probability of a drug successfully progressing from the first trial through the third trial is only 20 percent, meaning eight out of ten fail. These failures typically occur not because of flawed biology but because the experiments drugs were designed on used animal models like mice, which are poor predictors of human outcomes. AI models, while imperfect, significantly outperform animal trials in predicting human efficacy, which is where the real opportunity lies.

Beyond drug efficacy, Pande is also focused on precision medicine, which tailors treatment to individual patients rather than population averages. Currently, doctors often guess at diagnoses and prescribe drugs sequentially until finding one that works. AI-driven analysis of individual patient data could enable the first drug prescribed to be the right one, reducing trial-and-error treatment and improving outcomes.

"We would all be much better off if the first drug was the right one. Typically, your blood test values are compared to population averages. But really, they should be compared to: is this weird for you?" explained Pande.

Vijay Pande, Co-founder of VZVC

VZVC's concentrated approach signals a broader shift in venture capital, where smaller, specialized funds are leveraging automation to maximize impact while navigating an increasingly capital-intensive biotech landscape. Rather than spreading capital thin across dozens of bets, Pande's model suggests that the highest-impact AI-driven therapeutics will emerge from patient, data-rich partnerships between founders and investors who are willing to think in terms of decades rather than quarters.