Trapped ions are a leading platform for quantum technologies with long coherence times and high fidelity gates. However, a utility-scale quantum computer will need millions of physical qubits and gate times orders of magnitude faster than currently possible. In this talk I will lay out the problems hindering large-scale trapped ion quantum technologies and how we're working to solve them. This includes better trap design with integrated optics, gates that work even in the presence of experimental imperfections, and active learning for fast, hands- and model-free gate calibration.