The narrative that the desktop is dying resurfaces regularly – yet the same industry pushing SaaS centralisation is quietly rediscovering local hardware as an "edge node" for computationally expensive AI workloads. This contradiction is worth unpacking: it reveals that the question is not desktop vs. cloud, but rather who controls the compute and for whose benefit.

This talk makes the case for QGIS as a first-class architectural component in modern geodata infrastructure. Using a structured SWOT analysis, it compares QGIS against representative cloud-based GIS platforms across four dimensions: processing performance on large vector and raster datasets, extensibility through the plugin ecosystem, deployment and maintenance overhead, and strategic dependency risk (vendor lock-in, licensing, data sovereignty).

Concrete examples ground the analysis: local processing benchmarks for typical geoprocessing workflows (topology validation, large-scale raster analysis), a comparison of customisation depth between QGIS plugins and SaaS extension models, and two deployment patterns that address common IT administration concerns – managed QGIS profiles and MDM-based rollout.

The talk closes with a practical framing for hybrid architectures: where cloud services add genuine value, and where local compute is not a legacy compromise but the technically and strategically correct choice.

Key Takeaways
- The edge-computing trend inadvertently vindicates the local GIS client – and organisations should exploit this rather than defend against it.
- QGIS's plugin ecosystem enables domain-specific innovation that closed SaaS platforms structurally cannot replicate.
- Efficient desktop deployment is a solved problem; the administrative objections are manageable and concrete mitigations exist.