Ever wondered how modern web maps actually work under the hood—from cloud-native storage all the way to pixels on your screen? This talk takes you on a deep dive through the full vector tile rendering pipeline.

We’ll start at the source, exploring how geospatial data is stored and accessed efficiently using cloud-optimized formats like PMTiles. From there, we’ll unpack how data is encoded into vector tile formats such as Mapbox Vector Tiles (MVT) and emerging approaches like MapLibre Tile (MLT)—what they contain, how they are structured, and the design decisions behind them.

Next, we move into the browser to examine how tiles are decoded and rendered on the GPU using map renderers like MapLibre GL JS, and how they transform raw tile data into interactive maps via WebGL.

Finally, the talk closes with an outlook on the evolving hardware landscape and how technologies like WebGPU and WebAssembly SIMD are opening new possibilities for high-performance browser-based map rendering.