Vijay Pande Shares Wise Investment Strategies Following His Management of $4 Billion at a16z

Vijay Pande gained prominence initially within academic spheres before carving out a notable presence in the investment landscape. This shift occurred nearly ten years ago when Marc Andreessen and Ben Horowitz, initially wary of the healthcare and life sciences sectors, opted to invest in this area and appointed Pande to lead their initiatives. At that time, he was a well-respected chemistry professor at Stanford, recognized for launching Folding@home, a distributed computing project that transformed numerous personal computers into a supercomputer dedicated to disease research. Over the ensuing decade, he scaled a16z’s investment efforts, overseeing nearly $4 billion.

In a surprising development in June of the previous year, Pande chose to step down from this extensive role to create a much smaller venture. His new company, VZVC, co-founded with experienced investor Zach Werner, concentrates on making a limited number of focused investments each year, as opposed to spreading resources across numerous projects, operates without additional team members, and heavily integrates AI into its daily functions.

To explore Pande’s noteworthy transition, we discussed his recent decision to emphasize a select few focused investments rather than diluting resources in today’s market. We also touched upon one of the fascinating challenges in AI-driven biotechnology: unlike textual data, biological data isn’t easily available online, forcing companies to develop isolated datasets. What does this mean for AI’s potential breakthroughs in medicine, and who stands to benefit from these advancements?

This dialogue has been condensed for clarity and brevity. You can also listen to the complete conversation (below).

You mentioned that biology is evolving from a “science of discovery” to a sector that can be engineered. What does this mean?

Historically, drug development included an element of chance. The key change today is that AI and machine learning can analyze complex data to identify drug targets for specific diseases, assist in drug development, and even facilitate clinical trials—the most expensive part of the process.

I thought clinical trials were getting cheaper due to synthetic data, which lessens the need for participants.

That’s a hopeful viewpoint. While the costs and timelines for starting clinical trials have indeed decreased, especially with the aid of AI, executing a trial can still require hundreds of millions of dollars, explaining the high costs associated with new drugs. The likelihood of a drug successfully advancing from the initial trial to phase three is merely 20%. With an 80% failure rate and the significant expenditures linked to each trial, these costs accumulate rapidly. The majority of failures aren’t due to biologists’ mistakes, but often stem from testing drugs on animal models, such as mice, which frequently fail to accurately predict human responses. Although AI models are not infallible, they often perform better than animal models, opening up exciting possibilities.

[The next question is]: Is the drug the right one for me?

You’re alluding to personalized medicine…

The term you’re seeking is precision medicine. When patients consult a doctor for complex conditions, physicians often have to make educated guesses due to insufficient data. They prescribe one medication, and if it fails, they move on to another, and so forth. This is especially common in cancer treatment and other scenarios. It would be extremely advantageous if the first medication prescribed was the appropriate one. Currently, blood test values are compared to general population averages, but they ought to be tailored to judge what is unusual for each patient. We are beginning to enhance our grasp of which therapies are most effective for each individual.

Do you think this progress has been gradual, or has it accelerated recently?

I believe this is a consequence of several intersecting factors. For a long time, precision medicine focused mainly on genomics. However, the human genome serves as the foundational plan for a “structure” that evolves over time. Presently, various other measurable factors, such as proteomics, provide more relevant insights into diseases and health conditions. Additionally, advancements in robotics and automation closely align with AI, fostering natural synergies.

In the last decade, we’ve witnessed consistent advancements in applying AI to biology and chemistry. The biological aspect centers on disease treatment, while the chemistry aspect targets drug development aimed at specific proteins. Indeed, considerable progress has been achieved over this period.

You indicated that biology is one of the few fields where data cannot be easily sourced from the internet. What consequences does this have for the field’s advancement?

It implies a lack of shared datasets that organizations can use collaboratively to train their models. This gap poses unique challenges from an AI perspective.

Doesn’t this underscore a persistent issue in medicine, where practitioners often work in competitive silos?

You’ve highlighted a critical concern. When a patient has a particular type of cancer that falls under both oncology and endocrinology, specialists in each area often fail to collaborate effectively. The exciting aspect of AI is its capability to integrate knowledge across disciplines, simulating the effect of multiple leading physicians working in concert.

Is there sufficient data sharing for this vision to materialize? While I understand the inclination of founders and investors to safeguard their discoveries…

I’m noticing a significant move towards creating extensive biological information atlases, which are frequently developed as foundational models. As these become more common, I expect to see a similar path to what we’ve witnessed with open-source large language models outpacing corporate counterparts: open-source foundational models in biology will generate a considerable influence.

You have ties to Genesis Therapeutics, born from your Stanford lab, and to Insitro, the drug-discovery firm founded by your former colleague, Daphne Koller. Additionally, you’re fostering a venture with a founder you’ve known for twenty years. What traits do you value in founders, and which sectors are you focusing on?

I am focused on two key areas: AI for healthcare delivery, which I investigated extensively at a16z, and AI for clinical trials.

Trust is essential for me when evaluating founders. I seek partners with integrity—individuals who honor their commitments. I aim to build a relationship that spans 5 to 10 years, ideally guiding them into their next entrepreneurial venture. I prefer collaborating with people who regard success as a collective achievement rather than a competitive race.

What victories and obstacles have you encountered in your investment journey thus far?

When I first began discussing AI, machine learning, and their roles in medicine over a decade ago, the prevailing sentiment was skepticism, with many doubting its practicality or benefits. Witnessing the evolution in perceptions has been rewarding.

It took time for me to understand that, regardless of how captivating advanced technologies are, market strategy remains vital. I often advise my founders, especially those from scientific or product-focused backgrounds, to direct their creativity and intelligence into market strategies, as navigating the go-to-market landscape can be as challenging, if not more so, than the technology itself.

Could you share how you’re structuring this new firm differently from your role at a16z?

We’re embracing a notably different approach now… VZ is named using the initials of my name, Vijay, and my co-founder, Zach Werner—he is the “Z.” We intentionally keep our team small… on the investment side, it’s just the two of us. Although we initially intended to hire associates, our developed systems made that unnecessary.

How do you define “concentrated” in your investment strategy?

We are not aiming for 30 investments a year… we are looking to make around five concentrated investments. Bringing a new company into our portfolio is akin to welcoming a new child into a family—it’s a significant obligation on our part.

Given this structure, who are you competing against for deals?

Interestingly, with this model, we often don’t compete for the most desired rounds—investors typically appreciate our involvement. This is in contrast to the usual rush for Series A or B investments. Oftentimes, investors value our participation due to our hands-on approach and the distinctive value that Zach and I provide. I have immense respect for individuals like Antonio Gracias at Valor for his sustained contributions, and I find inspiration in Thrive’s achievements with a more concentrated portfolio as well. Naturally, my experiences at a16z continue to guide my thought process, but these other examples have expanded my perspective.

What aspects of AI and biotechnology do you think are currently overhyped?

The reality is that AI can indeed reveal insights beyond human capacities. The issue arises when claims are made that AI will solve all problems. The hesitance does not stem from doubts about AI but rather concerns regarding data quality. Large language models excel because they have extensive datasets to learn from. When the necessary data isn’t available, AI cannot simply fill that void.