Feasible supply of steel and cement within a carbon budget is likely to fall short of expected global demand

Steel and cement underpin modern economies but their production has more than doubled over the past two decades and now contributes roughly 15% of global CO₂ emissions. Achieving net-zero around mid-century is therefore critical but challenging within limited timeframes. Existing decarbonization scenarios for these sectors emphasize rapid, large-scale deployment of supply-side technologies, particularly carbon capture, utilization and storage (CCUS) and hydrogen-based processes. Recent analyses also point to potential competitiveness for hydrogen-based steel under favorable conditions. However, both CCUS and hydrogen approaches critically depend on extensive infrastructure for CO₂ transport and storage, renewable electricity, and green hydrogen, whose deployment is uncertain. Key feasibility concerns include: (1) long preparation periods for scaling mature yet not widely deployed technologies (pilot projects, permissions, social license, finance), which historically take decades; and (2) the achievable rate of deployment, with evidence suggesting that even aggressive growth trajectories for electrolysis may not deliver sufficient green hydrogen in time. Given these uncertainties, this study quantifies the feasible global supply of steel and cement under Paris-aligned carbon budgets while explicitly representing constraints from CCUS and non-emitting electricity infrastructure. Using a stochastic, physically based mass-balance optimization framework, we link plausible infrastructure build-out pathways to decarbonized steel and cement production and assess the resulting supply relative to expected demand.

The paper situates its contribution within a literature that predominantly relies on accelerated supply-side technology deployment for industrial decarbonization. IEA Energy Technology Perspectives scenarios assign large 2050 emissions reductions in steel and cement to CCUS. Academic studies often emphasize CCUS for heavy industry and increasingly explore hydrogen direct reduced iron (DRI) pathways enabled by low-cost renewable electricity. However, historical deployment of CCUS has lagged far behind projections and future scenarios envision scaling to levels unsupported by current construction plans. Prior critiques of integrated assessment modeling highlight feasibility risks of assumed infrastructure scale-up. This study addresses the under-emphasized risk of deployment failure by quantifying feasible supply given uncertain infrastructure build-out.

The authors develop a global optimization model grounded in physical mass balancing to estimate the maximum feasible production of steel and cement under Paris-consistent carbon budgets while constrained by infrastructure deployment and scrap availability. The model maximizes total materials production subject to: (a) cumulative CO₂ emissions budgets; (b) available CCUS capacity and its energy penalty; (c) availability of non-emitting electricity (which also proxies the feasible uptake of green hydrogen); (d) process-specific capacities and scrap flows; and (e) yields and recovery rates. End-of-life material generation is included to capture recycling flows. Infrastructure uncertainty is parameterized using ranges derived from IEA reports and current project data. For CCUS, the upper bound follows ambitious IEA 2023 projections (steel 670 Mt CO₂ yr⁻¹ and cement 1355 Mt CO₂ yr⁻¹ by 2050), while the lower bound linearly extrapolates from current operating capacity and the 2030 construction plan (steel 15 Mt CO₂ yr⁻¹ and cement 50 Mt CO₂ yr⁻¹ by 2050). For non-emitting electricity, upper and lower bounds follow ranges across IEA scenarios for future total supply and emission intensity. Given cross-sector electrification, electricity available for materials production is assumed to grow proportionally with total supply. A Monte Carlo approach samples uniformly within each bound for CCUS, electricity supply, and electricity emission intensity, with 1,000 iterations. For each iteration, the optimization solves for maximum feasible supply of steel and cement under the sampled infrastructure and the selected carbon budget. Carbon budgets: two global pathways are used—1.5 °C with 50% probability (approximately −420 Gt CO₂) and a well-below 2 °C budget represented by a 1.7 °C pathway with 50% probability (approximately −770 Gt CO₂). Sectoral budgets for steel and cement are allocated proportional to current emissions and global mitigation rates, consistent with prior studies. The budgets do not assume large-scale carbon dioxide removal, implying more rapid near-term reductions than IEA Net Zero. Technology and process assumptions align with industry-accepted best practices and linear improvements to 2050. For steel, process routes include BF-BOF, fossil DRI-EAF, H₂ DRI-EAF, and scrap-EAF. Hydrogen-based DRI assumes 45 kWh per kg H₂ and a hydrogen mass fraction of 1.5, reflecting one-step full conversion with oversupply; no further electrolyzer efficiency improvements are assumed. For cement, clinker substitution is assumed to increasingly rely on alternatives to fly ash and GBFS (e.g., calcined clay, agricultural by-products, by-product ash, end-of-life binders). Waste-derived thermal fuels are assumed to expand using agricultural, chemical, and food-sector wastes. Feasible supply outcomes are compared to expected demand from three scenarios: IEA Baseline (Stated Policies), IEA Net Zero, and the Low Energy Demand (LED) scenario, to quantify supply-demand gaps and the implied need for demand-side resource efficiency.

- Feasible supply shortfalls by 2050 under 1.5 °C: Steel can meet 55–85% (interquartile range) of expected baseline demand; cement can meet 22–56% (interquartile range). - Under a well-below 2 °C budget (1.7 °C, 50% probability), steel feasible supply is estimated around 57–84% of expected demand (as indicated in the results narrative), still below baseline expectations. - Process shifts in steel: BF-BOF production declines markedly to meet the 1.5 °C budget. Hydrogen-based DRI-EAF and scrap-EAF expand substantially using non-emitting electricity. By 2050, these two routes together could exceed 60% of today’s total steel supply, with interquartile ranges of 100–200 Mt (H₂ DRI-EAF) and 1100–1120 Mt (scrap-EAF). Scrap-EAF growth is more robust (lower electricity per unit) but ultimately constrained by scrap availability; fossil DRI-EAF is regionally limited by natural gas access. - Infrastructure feasibility gaps: IEA 2010 projected ~195 Mt CO₂ yr⁻¹ capture for steel and cement by 2021, whereas actual is under 1 Mt CO₂. IEA scenarios continue to assume ~2000 Mt CO₂ capture in these sectors by 2050, versus a 2030 construction plan of ~19 Mt CO₂—implying two orders of magnitude scale-up beyond current plans. The study’s 2050 CCUS bounds span steel 15–670 Mt CO₂ and cement 50–1355 Mt CO₂. - Basic needs coverage: Even within the 1.5 °C budget, feasible supply suffices to cover minimum material requirements for basic human needs by 2050: these needs represent 13–14% (steel) and 52–63% (cement) of feasible supply (interquartile ranges). Under the well-below 2 °C budget, the shares fall to 11–12% (steel) and 39–45% (cement). - Sectoral implications: Allocating 1.5 °C feasible supply by current shares suggests median coverage of only ~40% of expected 2050 baseline demand in construction and ~60% in manufacturing—highlighting substantial gaps without demand-side measures. - Equity considerations: Current per-capita stocks in high-income countries exceed feasible 2050 levels, while many low-income countries remain below. Over 90% of additional material needed to meet basic needs arises in lower-middle and low-income countries, implying equity and allocation challenges within constrained feasible supply.

Findings indicate that, despite plausible technological advances, uncertain deployment of enabling infrastructure (CCUS, non-emitting electricity, green hydrogen) restricts the feasible supply of steel and cement below expected demand under Paris-aligned carbon budgets. This underscores a dual imperative: (1) governments must accelerate infrastructure deployment dramatically; and (2) industries must prepare for the possibility of deployment shortfalls rather than relying solely on future large-scale infrastructure. The analysis reframes mitigation strategy by prioritizing a forecasting-supply/backcasting-demand approach. Establishing what supply is feasible clarifies how much demand must be reduced through resource efficiency. To remain within a 1.5 °C budget, the construction sector would need to deliver the same services with roughly 60% less material, while manufacturing would need about 40% less, achievable via design optimization, right-sizing, material utilization improvements, product life extension, reuse/remanufacturing, and modal shifts in mobility. Equity considerations are central: aligning high-income stock replacement with feasible global limits while expanding access in lower-income regions is necessary to meet basic needs worldwide within constraints.

This study quantifies the feasible global supply of steel and cement under Paris-compliant carbon budgets while explicitly incorporating uncertainties in CCUS and non-emitting electricity deployment. Results show likely shortfalls relative to expected demand by 2050, despite growth in hydrogen-based and recycling-based production routes. The work provides concrete benchmarks for the level of demand-side resource efficiency required—on the order of 60% material reduction in construction and 40% in manufacturing under a 1.5 °C pathway—and motivates a shift to forecasting feasible supply before backcasting demand-side actions. Policy and industry should act on two fronts: rapidly scale essential infrastructure and simultaneously implement resource-efficiency strategies to hedge against deployment risks. Future research should refine regionalized infrastructure feasibility assessments, deepen analysis of supplementary cementitious materials and alternative fuels availability, integrate broader system interactions across sectors, and explore governance mechanisms for equitable material allocation within constrained feasible supply.

- Infrastructure uncertainty representation: CCUS and non-emitting electricity are sampled uniformly within bounds derived from IEA scenarios and current plans; true probability distributions and regional constraints may differ. - No large-scale carbon dioxide removal is assumed; this requires faster near-term industrial decarbonization than some scenarios (e.g., IEA NZE). - Global aggregation: Results are global and do not capture region-specific infrastructure, resource, or policy constraints (e.g., natural gas access for fossil DRI, location-specific CO₂ storage). - Technology progress assumptions: Linear improvements to 2050 and adherence to current best practices may under- or overestimate actual trajectories. - Hydrogen and electricity proxies: Electricity availability is used as a proxy for green hydrogen scale-up; electrolyzer, storage, and transport constraints are not modeled in detail. - Material inputs for cement: Availability of alternative clinker substitutes and waste-derived fuels is assumed based on literature but not fully coupled to competing sectoral demands. - Sectoral budget allocation: Sector budgets are proportional to current emissions and mitigation rates; alternative allocation schemes could change feasible supply. - Monte Carlo size and distributions: 1,000 iterations with uniform distributions capture range but not tails or correlated uncertainties among variables.