We develop a model of two-stage startup financing with signaling under ambiguity. In our model, the nature determines the ability of technology entrepreneur and he strategically chooses a costly patent level as a signal to inform his ability to potential investors. Angels participate in investments for seed money in the first stage after observing the patent level and then venture capitals offer investments to the entrepreneur in the second stage based on angels’ behavior. We provide three different financing models in view of the degree of ambiguity: (1) no ambiguity; (2) only investors face ambiguity; (3) all agents face ambiguity. In each signaling game between the entrepreneur and investors who may have ambiguous beliefs about the types of the entrepreneur, we find a unique perfect Bayesian equilibrium by imposing the Intuitive Criterion of Cho and Kreps (1987) and characterize the refined equilibria. In particular, angels ask the highest level of patents in order to ensure the ability of the entrepreneur when only investors have ambiguous information. We also find that the entrepreneur can obtain a higher utility under ambiguity than without it if his project is sufficiently overvalued. On the other hand, when investors have ambiguous information, the entrepreneur can be better off by resolving ambiguity if the project is sufficiently undervalued.