<p>The steel industry emits 2,400–2,713 Mt CO₂ and 12 Mt of fugitive methane globally each year, yet progress towards decarbonization has been limited. In this Review, we examine the potential, feasibility and barriers to decarbonization of the steel sector. The blast furnace–basic oxygen furnace (BF–BOF) accounts for ~72% of global steel production and has the highest carbon emission intensity of existing production routes (~2.3 t CO₂ t_steel⁻¹). Conversely, the scrap-based electric arc furnace (EAF) route, which contributes ~23%, has the lowest carbon intensity (~0.68 t CO₂ t_steel⁻¹). China, Japan and others rely heavily on BF–BOF whereas the EAF route is favoured in a limited number of nations such as the USA. Established process-level solutions, such as enhanced energy efficiency and waste heat recovery, can reduce carbon intensities by up to 20%. Whereas reductions of over 80% can be achieved with emerging hydrogen-based and electrolysis-based technologies. For example, the carbon intensity of steel produced with hydrogen-based iron reduction technologies is ~0.4 t CO₂ t_steel⁻¹ at a cost of over US$800 t_steel⁻¹ (compared to ~US$450 t_steel⁻¹ for BF–BOF). These high costs, together with regional resource limitations and low technical readiness, limit widespread near-term deployment. System-wide measures, such as material efficiency, circular economy and industrial symbiosis, could contribute 30–65% of the required total emissions reduction in line with the 1.5°C target. Future work should prioritize implementing a coordinated, multi-scale approach that combines process-level innovations with system-wide strategies and region-specific policies.</p>