How is corporate innovation shaped by scientific discoveries in academia, if at all? While much prior research links academic research and innovation in the industry, how academic science shapes firms’ innovation search and outcomes is less explored. In this paper, we focus on increased exposure to academic science for firms due to the arrival of university scientists in the geographically proximate area and examine how it changes the innovation of local firms. We show that an increase in exposure to academic science is not responsible for a uniform increase in patented inventions at firms, but rather “fattens the tails of innovation,” shifting firms’ R&D efforts from exploitation to exploration. Firms in counties with increased exposure to academic science deliver inventions based on science, and specifically, novel inventions that are at once highly-impactful and zero-impact. This effect can be directly traced via citations to the scientific articles and fields of newly-arrived scientists. Also, firms increase reliance on science originating from a university, rather than non-university organizations such as corporate and government laboratories. Both local established firms and startups are responsible for the effect, yet established firms benefit more from academic science than startups. We provide a causal interpretation via a shift-share instrument for scientist mobility based on the distribution of surnames in the 1940 U.S. Census. Our findings move beyond the notion that universities-matter for innovation, informing R&D of the dual implications of relying on scientific discoveries: outsized breakthroughs may be increased, but so is the probability of failure.