Abstract: I study the strategic investment decisions of competing venture capitals (VCs) in startups, focusing on the impact of uncertainty on investment and startup dynamism. My framework considers multi-round funding requirements and VC optimization based on current information and projections of future success (M&A or IPO). Using a novel dataset on the “life-cycle” of biotech and software startups, I establish that the data identifies model parameters and propose methods to correct for endogenous and dynamic selection to infer true startup values and VC information. Among several others, I find that (1) biotech investors initially possess more information than software investors but learn slower, reflecting sector-specific business models; (2) uncertainty leads to underfunding of promising startups, causing welfare losses of 22% and 21% in biotech and software, respectively; and (3) “dynamic (positive) information externality” from early stage investors to late stage investors causes the former to invest less, leading to welfare losses of $10 billion in biotech and $3 billion in software. I also explore policies to mitigate these losses. Additionally, I estimate that stricter M&A policies reduce VC returns, significantly decreasing startup funding and exacerbating welfare losses.