This talk introduces a new vector tiles format called MapLibre Tiles (MLT), which offers a significant tile size reduction and accelerated decoding performance compared to the de-facto standard Mapbox Vector Tiles (MVT). MLT also adds support for missing features like nested properties, linear referencing and M-values. The design of MLT is influenced by the latest research results on big data analytics formats and adapted for the map visualization use case. Our evaluation against MVT on a OpenMapTiles schema based tileset shows a reduction in tile size of nearly up to 80% with even faster decoding times. Moreover, on a already highly optimized Bing Maps tileset, MLT achieves reductions of up to 40% in size. Based on these results, Microsoft generously donated to MapLibre to integrate MLT into the mapping stack. Additionally, we will explore how next-generation map rendering libraries can leverage SIMD and WebGPU compute shaders for processing to fully utilize the potential of MLT.