As high-performance computing, artificial intelligence, and power-electronic devices continue to evolve toward higher integration and power density, localized chip hotspots impose increasingly severe thermal constraints. Conventional air cooling and heat-spreading-only solutions are becoming insufficient to simultaneously remove high heat fluxes and maintain temperature uniformity. Microchannel liquid cooling offers a large surface-area-to-volume ratio, short heat-transfer paths, and high volumetric heat-transfer capability. In particular, manifold microchannels shorten the effective flow length and reshape coolant distribution, providing a promising route for cooling chips under high and spatially non-uniform heat loads. Based on a thorough revision of the original manuscript, this review discusses the development of microchannel heat sinks, geometric parameters and flow-distribution mechanisms of manifold structures, hotspot-oriented cooling strategies, field-synergy and secondary-flow mechanisms, thermal response under dynamic boundary conditions, multiphysics simulation, and advanced manufacturing. Current studies indicate that manifold structures can reduce streamwise pressure losses while improving local heat transfer; however, their performance strongly depends on the coupled manifold-microchannel geometry, flow uniformity, and operating boundary conditions, and no universal optimum geometric parameter exists independently of the specific application. Future work should focus on adaptive flow control for dynamic hotspots, chip-cooler co-design, two-phase and advanced coolants, data-driven optimization, and integrated additive manufacturing to move manifold microchannel cooling from high-performance prototypes toward reliable and scalable engineering systems.