VZVC: Vijay Pande Limits His Investments to Five Per Year with AI Assistance

After managing nearly $4 billion at a16z, Vijay Pande left in June to co-found VZVC with Zach Werner. Their thesis: a handful of highly focused bets, a lean investment team of two, and AI agents at the helm. Pande also elaborates on his views regarding clinical trials, biological data, and foundation models in healthcare.

Five Investments a Year, No Hunting for Hot Rounds

Vijay Pande clarifies that VZVC does not make 30 investments a year but aims for about five highly concentrated deals. He compares adding a company to their portfolio to the decision to have another child, rather than just adding a contact on a social network. The investment team is limited to two people, Pande and Zach Werner, without associates. The initial intention to recruit more has been abandoned, as the internally developed agents have rendered these positions unnecessary. VZVC does not seek to compete in highly contested investment rounds, and Pande notes that founders generally reserve a spot for them, unlike the competitions for Series A or B funding. According to him, teams want VZVC in their capital for the expertise and involvement of the two co-founders.

Clinical Trials: High Costs and the Promise of AI Against Animal Models

The reduction of clinical trial costs through synthetic data remains, according to Vijay Pande, more of an aspiration than a reality. He observes a decrease in the cost and time to access trials, particularly thanks to AI, but emphasizes that a trial can still cost hundreds of millions of dollars. The probability of success for a drug from Phase I to the end of Phase III is 20%. Pande attributes many failures to the low predictive capacity of animal models for humans and argues that AI models, while not perfect, should perform better. He adds that AI is also beginning to assist directly in trials.

From Genomics to Proteomics, Towards More Personalized Medicine

Vijay Pande advocates for precision medicine where the first treatment would be the right one, rather than a succession of drug trials. He points out that blood test results are generally compared to population averages and argues for personalized benchmarks, a capability he believes is emerging on the medical side. Historically focused on genomics, the approach should, in his view, integrate proteomics and other measures closer to the actual state of the body, with a contribution from robotic automation naturally linked to AI. Over the past decade, he describes constant advancements in AI in biology and chemistry, with the latter helping to design molecules targeting specific proteins. He considers these advancements highly significant and credits AI and machine learning with their ability to illuminate complex biological systems and select therapeutic targets.

Biological Data: From Silos to Open Foundation Models

Pande emphasizes that, unlike online texts, biological data cannot be extracted from the internet to train models and do not easily distill from one model to another. Many companies are therefore building isolated datasets. However, he argues that AI can, in principle, cover multiple specialties and identify patterns imperceptible to humans, akin to a real-time expert consultation. He observes a push towards biological information atlases that are materializing as foundation models. He anticipates that open-source foundation models in biology will have a broad impact, similar to the performance observed with open-source LLMs compared to enterprise models.

Investment Areas and Founder Criteria at VZVC

VZVC focuses its work on two axes: AI for healthcare delivery and AI for clinical trials. Pande is involved with Genesis Therapeutics, which originated from his lab at Stanford, and with Insitro, launched by Daphne Koller, a former colleague. He is also incubating a company with a founder he has known for 20 years. Regarding founders, he seeks mutual trust and high integrity for relationships lasting 5 to 10 years or more, with a shared success mindset. From his experience, he notes that the go-to-market strategy is at least as challenging as the technology itself and advises investing as much creativity in it as in the product.

Background and Influences: From Stanford to a16z, Then VZVC

Coming from an academic background, Vijay Pande was chosen by Marc Andreessen and Ben Horowitz to launch a16z's health and life sciences practice after an initial five-year period without investments in this sector. A chemistry professor at Stanford and the initiator of Folding@home, he grew this practice to nearly $4 billion under management in a decade. He left a16z in June to co-found VZVC, named after his initials with Zach Werner. Among his inspirations, he cites Antonio Gracias at Valor, known for a deal with SpaceX and active for 20 years, as well as Thrive for its concentrated portfolio, while claiming the DNA of a16z with the infusion of new influences. Pande believes that the past resistance to AI in medicine has largely dissipated. He distinguishes VZVC's approach from hot market rounds and positions his project within this personal and sectoral trajectory.