“We’re not doing 30 bets a year”: Vijay Pande on betting small after running $4 billion at a16z

For years, Vijay Pande was a figure more synonymous with the ivory tower than the venture capital boardroom. A Stanford chemistry professor, he gained global recognition for spearheading Folding@home, a pioneering distributed-computing project that harnessed the idle processing power of millions of home computers to unlock the mysteries of disease.

That trajectory shifted dramatically over a decade ago when Marc Andreessen and Ben Horowitz—who had famously steered their firm away from the life sciences for its first five years—decided the sector was finally ripe for disruption. They tapped Pande to lead the charge. Over the subsequent decade, Pande transformed that initial foray into a powerhouse biotech practice managing roughly $4 billion.

Yet, in a move that caught the industry off guard, Pande stepped away from the a16z juggernaut last June. His new venture, VZVC, co-founded with longtime investor Zach Werner, represents a radical departure from the "spray and pray" model of modern venture capital. Eschewing the associate-heavy, high-volume approach of his previous firm, Pande is now focused on a hyper-concentrated portfolio, leveraging AI-native operations to redefine how biotech startups are built and funded.

From Discovery to Engineering: The Biological Shift

Pande views the current moment as a fundamental transition in how we approach medicine. For decades, drug development was a game of chance, relying on serendipity and trial-and-error. Today, he argues, biology is evolving into an engineering discipline.

"I think 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."

Despite the promise of AI, Pande remains grounded in the harsh realities of the industry. Clinical trials remain the most expensive and risky bottleneck in the pharmaceutical pipeline. With a success rate of only 20% from Phase 1 to Phase 3, the financial burden of failure is immense. Pande notes that these failures often stem from the reliance on animal models—like mice—which frequently fail to predict human physiological responses. He believes AI models, while not perfect, offer a significantly more predictive alternative, potentially saving billions in wasted R&D.

The Precision Medicine Conundrum

Beyond drug discovery, Pande is deeply invested in the promise of precision medicine. He points out that current medical practice is often reactive and generalized, comparing individual patient data to broad population averages rather than understanding the unique biological context of the person.

• The Genomic Limitation: While genomics provides a blueprint, it is static. Pande emphasizes that proteomics and other dynamic measurements are now essential for understanding the body’s current state.
• Automation and AI: The integration of robotic measurement systems with AI is creating a feedback loop that allows for faster, more accurate biological insights.
• The "Right Drug" Goal: The ultimate objective is to move away from the trial-and-error cycle of prescribing medications, ensuring the first treatment a patient receives is the one that actually works.

The Data Wall: Why Biology Isn't Like Text

One of the most significant challenges in applying AI to biology is the nature of the data itself. Unlike the vast, scrapable expanse of the internet that fueled the rise of Large Language Models (LLMs), biological data is fragmented, proprietary, and often locked behind corporate firewalls.

Pande acknowledges that this creates a unique hurdle: because biological data cannot be easily distilled or scraped, every company is effectively building its own walled garden. This siloed approach mirrors the broader, often dysfunctional structure of modern medicine, where specialists in oncology, endocrinology, and other fields rarely share data effectively.

"What is really intriguing about AI is that it can, in principle, be a specialist in everything, and it can start to see things that really any single human being couldn’t. It would be equivalent to having a team of the very best doctors all clamoring together in that moment."

Pande is optimistic that the industry is moving toward foundation models—shared atlases of biological information. He believes that just as open-source LLMs have challenged corporate giants in the tech world, open-source biological models will eventually drive broader, more impactful innovation.

VZVC: A New Philosophy of Investing

At VZVC, Pande and Werner are intentionally keeping their footprint small. The firm operates without associates, relying on AI agents to handle the heavy lifting of day-to-day operations. This allows the partners to remain deeply involved in the companies they back.

Key Principles of the VZVC Model:

• Extreme Concentration: Rather than chasing 30 deals a year, VZVC focuses on a handful of high-conviction bets.
• The "Child" Analogy: Pande describes adding a company to their portfolio not as a casual transaction, but as a long-term commitment, akin to raising a child.
• Founder Integrity: Pande prioritizes founders with high integrity who are focused on long-term value creation rather than short-term competitive wins.
• Go-to-Market Focus: Having learned from his own career, Pande now emphasizes that for scientists and product-focused founders, the go-to-market strategy is often the most difficult hurdle to clear.

Looking Ahead: What’s Real vs. What’s Hype

When asked about the current AI frenzy, Pande remains pragmatic. He warns against the narrative that AI will "cure everything," noting that the efficacy of any AI model is strictly bounded by the quality and availability of the underlying data.

"The reason for hesitance there is not because of any doubt about AI—it’s about doubt of the data," Pande explains. "LLMs work because there’s so much data to learn from. When the data is just simply not there, then AI can’t magically solve that problem."

As he looks back on his career, Pande finds fulfillment in the fact that the skepticism he faced a decade ago—when he first championed the intersection of AI and biology—has largely evaporated. By betting small and staying hands-on, he is betting that the next decade of medical breakthroughs will be defined not by the sheer volume of capital deployed, but by the precision and intelligence of the engineering behind it.