Scenarios play a central role in helping policymakers envisage pathways to limit global warming to well below 2°C. We demonstrate that the most recently assessed set of climate stabilization scenarios still favors fossil fuels, and in particular coal, and bioenergy. In contrast to insights from empirical innovation studies, scenarios are optimistic on deployment of lumpy, energy-systems technologies, such as carbon capture and storage, while insufficiently reflecting innovation dynamics in granular technologies. Our analysis shows that two pathways for rapid decarbonization remain systematically undersampled in models that underpin IPCC scenarios: A) strong growth in intermittent renewables, in particular solar PV, together with electrification of sectors; and B) widespread adoption of efficient end use technologies, digitalization, and new service provisioning systems enabling low energy demand. A combination of continued PV growth and sector coupling with low to medium energy demand (a corridor of 250 to 500 EJ of primary energy) would make fossil fuels obsolete by 2050, thus enabling near-term cost effective climate change mitigation and reducing the need for carbon dioxide removal in the 2nd half of the century. These pathways are realistic, target inclusive well-being, but remain underrepresented in the modelling literature. We see three modeling innovations that would improve resolution of near and mid-term dynamics: 1) updating of renewable energy cost assumptions and fuller representation of technological learning curves, 2) more explicit modelling of sector coupling and specifically power-to-X technologies, and 3) including insights from hourly resolution modelling of energy systems.