Upcoming upgrades to the LHC will significantly reduce statistical uncertainties, enabling measurements at percent-level precision. To fully exploit this level of precision, theoretical predictions must achieve comparable accuracy, which requires incorporating higher-order terms in the perturbative series in the strong coupling constant. Currently, only next-to-leading order terms can be included for arbitrary scattering processes, while higher-order corrections remain a significant challenge. A major obstacle lies in the complexity of two-loop calculations for high-multiplicity scattering amplitudes. In this talk, I will overview the modern analytic techniques that have enabled progress on these complex calculations, leading to the first NNLO-precise predictions beyond two-to-two processes. We will explore the current state of two-loop amplitude calculations with multiple kinematic scales and discuss prospects of extending these techniques to a broader class of processes, including those involving massive electroweak bosons and top quarks.