Former Andreessen Horowitz general partner Vijay Pande, who ran a roughly $4 billion biotech practice at the firm before departing last year, said his new fund VZVC is structured to make far fewer investments than his former employer’s portfolio pace allowed, in an interview with TechCrunch published this week.
“We’re not doing 30 bets a year,” Pande told TechCrunch, referring to the cadence of biotech investments during his tenure at a16z. His new firm, VZVC, is positioning itself as an AI-native venture vehicle focused on biology, with a deliberately smaller portfolio footprint. The fund’s structure marks a notable departure from the broad-check-writing style that defined a16z’s bio strategy during one of the most active periods in the firm’s history.
Pande’s move reflects a broader recalibration in how Silicon Valley capital is being deployed into life sciences. Andreessen Horowitz’s bio fund under Pande became one of the most active in the sector, writing checks across dozens of companies working on drug discovery, computational biology, and digital health. The shift to a smaller fund suggests a more concentrated approach at a moment when AI tools are reshaping the cost structure of early-stage biology. It also comes as the broader biotech venture market has cooled from the highs of the early 2020s, with limited partners becoming more selective and the cost of capital pressuring even well-funded managers to demonstrate discipline.
Pande also argued that biology is undergoing a fundamental change in character, moving from what he characterized as a “discovery” science to an “engineering” one — a transition that, if accurate, would have implications for how startups are capitalized and how research translates into therapeutics. The distinction matters for investors sizing risk: discovery-stage biology has historically produced long timelines, high failure rates, and capital-intensive clinical work. A shift toward an engineering paradigm would imply more predictable workflows, modular components, and shorter iteration cycles, the kinds of properties that have historically attracted software-style venture capital.
On clinical trials, Pande acknowledged costs remain punishing even with AI in the mix. Trials continue to dominate biotech burn rates and have not seen the dramatic cost compression that software has experienced. He pointed to shared, open datasets — rather than walled-off proprietary ones — as the structural change most likely to lower the cost of bringing a drug to trial. The argument leans on a recurring thesis in AI-bio circles: that the binding constraint on machine-learning models in biology is not algorithmic sophistication but the availability of high-quality, standardized training data. Pande’s preferred remedy — open, shared infrastructure — places him closer to the open-science end of the spectrum, a position with notable advocates and notable opponents inside the pharmaceutical industry.
Why it matters: Pande’s bet is a test case for whether AI-native biotech investing can be done at venture scale without the spray-and-pray diversification that defined the previous cycle. His track record at a16z, where the bio fund backed a wide range of companies across modalities and stages, gives the new fund credibility but also raises the question of whether a smaller, more concentrated portfolio can outperform a broader one in a field where the base rate of clinical success remains low. The answer will shape how limited partners think about fund construction in life sciences for years to come.
Background and context: Andreessen Horowitz’s bio fund expanded significantly during Pande’s tenure, becoming one of the most visible dedicated biotech vehicles in Silicon Valley. The firm invested across a wide spectrum of companies, from early-stage platform plays to later-stage therapeutics developers, and built a network of operating partners drawn from the pharmaceutical industry. Pande’s departure was part of a broader period of turnover at a16z’s bio practice. His new firm, VZVC, has been positioned as a more focused vehicle, with explicit orientation toward AI-native tooling and computational approaches to biology. The “engineering vs. discovery” framing Pande invoked has become increasingly common in industry discourse, with proponents pointing to advances in protein structure prediction, generative molecular design, and high-throughput experimental workflows as evidence of a more deterministic pipeline. Critics counter that the biology of human disease remains stubbornly empirical, with clinical outcomes continuing to diverge from preclinical predictions.
Analysis: Pande’s framing of biology as transitioning to an “engineering” discipline is a thesis, not an established consensus. While machine-learning approaches have demonstrated measurable gains in protein structure prediction, target identification, and molecular generation, the translation of those gains into reduced clinical failure rates and shorter development timelines remains unproven. Pande’s credibility on the investment side comes from his tenure scaling a16z’s bio practice, but his current fund’s smaller size and concentrated approach is a bet that AI-native tooling will allow leaner teams to compete with well-capitalized incumbents — an outcome the industry has chased for several years without yet delivering it at scale. The argument for open datasets as a cost-lowering mechanism is consistent with broader open-science advocacy but runs against the proprietary-data strategies of large pharmaceutical companies and several competing AI-bio startups, which treat curated datasets as core moats. The success of either model has implications well beyond any single fund, since the data infrastructure built today is likely to shape the competitive landscape of drug discovery for the next decade.
What to watch next: Several indicators will determine whether VZVC’s approach validates itself. The first is the pace and size of VZVC’s announced investments, and whether the firm can attract co-investors at the scale needed to support clinical-stage companies. The second is the clinical progression of any portfolio companies, particularly whether AI-native workflows produce candidates that advance faster or at higher success rates than industry baselines. The third is the response of larger pharmaceutical companies, which have the option to build in-house capabilities, acquire AI-bio startups, or enter strategic partnerships — each of which would shift the competitive terrain for a fund like VZVC. The fourth is the evolution of data infrastructure, particularly whether open datasets gain traction or whether proprietary data sets remain the dominant model.
Conclusion: Pande’s comments point to a venture model that is more selective, more concentrated, and more explicitly AI-native than the one he helped build at Andreessen Horowitz. Whether the new model outperforms the old will depend less on the theory of biology as engineering and more on whether the underlying tooling and data infrastructure have matured to the point where smaller teams can produce clinical outcomes at lower cost. The industry has heard versions of this thesis for several years; Pande’s new fund is one of the more visible attempts to put capital behind it.
Sources
– TechCrunch: “We’re not doing 30 bets a year”: Vijay Pande on betting small after running $4 billion at a16z — https://techcrunch.com/2026/08/29/were-not-doing-30-bets-a-year-vijay-pande-on-betting-small-after-running-4-billion-at-a16z/
Corrections
If you believe this article contains an error, contact Herald Express with the source URL and supporting evidence.
Story synopsis gathered from: TechCrunch — source
