Researchers agree that alliance networks can be an important instrument in a firm's innovation process, but there is limited empirical evidence on actually how they facilitate the creation of new knowledge for exploratory innovation. The research question is what alliance network configuration is optimal for exploratory innovation. The present study investigated the interaction between a firm's alliance portfolio structure (the micro-level) and the industry alliance network structure (the macro-level), and it empirically tested how their interaction may be affecting the exploratory innovation outcome of network participating firms in the biotechnology industry. The paper uses data from exploratory patents filed by 455 dedicated biotechnology firms in 1986–1999 and an overall network comprising 2,933 technological alliances over the same period. The results indicate that, in the case of biotechnology, firms with high exploratory innovation output have short path indirect access to many other firms (micro-level), and operate in dense industry alliance networks centralized around a few key firms (macro-level), and that these effects are curvilinear.