The conventional Hamiltonian H = p 2 + V N (x), where the potential V N (x) is a polynomial of degree N , has been studied intensively since the birth of quantum mechanics. In some cases, its spectrum can be determined by combining the WKB method with resummation techniques. In this paper we point out that the deformed Hamiltonian H = 2 cosh(p) + V N (x) is exactly solvable for any potential: a conjectural exact quantization condition, involving well-defined functions, can be written down in closed form, and determines the spectrum of bound states and resonances. In particular, no resummation techniques are needed. This Hamiltonian is obtained by quantizing the Seiberg-Witten curve of N = 2 Yang-Mills theory, and the exact quantization condition follows from the correspondence between spectral theory and topological strings, after taking a suitable fourdimensional limit. In this formulation, conventional quantum mechanics emerges in a scaling limit near the Argyres-Douglas superconformal point in moduli space. Although our deformed version of quantum mechanics is in many respects similar to the conventional version, it also displays new phenomena, like spontaneous parity symmetry breaking.where V N (x) is a polynomial of degree N . When N = 2 (the harmonic oscillator), the spectral problem was solved in the 1920s, but for potentials of higher degree there is no obvious analytic solution. On the contrary, the Hamiltonian (1.1) with N ≥ 3 has become the textbook testing ground for approximation techniques, like stationary perturbation theory or the WKB method. Starting in the 1970s, it was realized that these approximation methods can be sometimes upgraded by using resummation techniques, providing in this way a close analogue of exact solutions. In particular, exact or uniform versions of the WKB method lead to quantization conditions for anharmonic oscillators [3,4,10,25,61,96,102,103,109,110,111]. These conditions require Borel-Écalle resummations of the WKB periods and include non-perturbative corrections to the perturbative quantization condition of Dunham [33], which in turn is a generalization to all orders in of the Bohr-Sommerfeld quantization condition.1 There are some peculiar quantum mechanical models with periodic potentials where spontaneous parity symmetry breaking is possible, see for instance [107]. However this is a very different setup from the one considered in this paper. 2 We would like to warn the reader that in this work "instantons" refers to both quantum mechanical instantons and gauge theory instantons. We hope this will not lead to confusion.