IEA Iron & Steel sector data and analysis

It is the energy-agency answer to a question every steel participant eventually faces: how much energy does this industry burn, how much steel does it make, and how much CO2 does that release? IEA Iron & Steel sector data and analysis sits inside the agency's Energy system > Industry section and treats steel as roughly eight percent of global final energy demand and about seven percent of energy-system CO2 emissions. Three headline charts carry the tracking layer - energy demand in iron and crude steel production, iron and crude steel production itself, and CO2 emissions from iron and steel, each spanning 2015-2024 and stamped with its last update, the production series refreshed 18 June 2026.

The analytical layer beneath is what separates this from a bare statistics table: the Iron and Steel Technology Roadmap (October 2020, Energy Technology Perspectives series) models STEPS, the Sustainable Development Scenario and the Faster Innovation Case out to 2050 with charts covering end-use steel demand, in-use stocks, capacity age profiles, abatement by mitigation lever, production by process route, CO2 captured and hydrogen use. A September 2023 Net Zero Emissions guide extends the frame to aluminium; definitions for near-zero and low-emissions steel, measurement and data-collection guidance, India-specific standards work and a policy toolbox round it out. Supporting tools - the ETP Clean Energy Technology Guide (~640 technology designs) and the Clean Energy Demonstration Projects Database - sit alongside.

In Datadory's catalog of 1,744 datasets across 159 viable industries this record scores 6/10 for quality and anchors the steel vertical's energy-and-emissions corner - the layer you cite when a claim has to hold at planetary scale, not just across the countries that publish their own numbers.

Four rows that describe the whole record before you request anything:

# headline tracking series (annual, 2015-2024)
series=energy_demand_iron_crude_steel   unit=final energy   breakdown=region x fuel   weo_source=WEO 2025
series=production_iron_crude_steel     unit=tonnes         last_updated=2026-06-18
series=co2_emissions_iron_steel        unit=Mt CO2         scope=global with regional detail

# roadmap scenario rows (projected to 2050)
scenario=STEPS                    full_name=Stated Policies Scenario
scenario=SDS                      full_name=Sustainable Development Scenario
scenario=Faster Innovation Case   target=net zero by 2050

# shape of the slice
grain=annual   span=2015-2024 observed / to 2050 projected   layer=tracking charts + roadmap analysis

What those lines already tell you: the three tracking series share one clock - annual prints across 2015-2024 - so a demand row, a tonnage row and an emissions row for any year join on two keys without harmonization. Energy demand is the only series broken out by region and fuel, which makes it the one that answers 'which economies and which fuels carry steel's eight-percent share of world final energy'. And every chart-level figure carries its refresh stamp - 18 June 2026 on the production series - so a stale number announces itself instead of hiding inside a dashboard.

Seven verified core fields, written against the published charts and reports rather than guessed:

• Series - which quantity a row reports: energy demand in iron and crude steel production, iron and crude steel production, or CO2 emissions from iron and steel.
• Period - the reference year, 2015 through 2024 on the tracking layer.
• Value - the figure itself in the series' unit.
• Region - global aggregate or the regional split, deepest on the demand series.
• Fuel - the energy-carrier dimension of the demand series, separating the coking-coal-heavy blast-furnace route from scrap and DRI pathways.
• Scenario - roadmap pathway label on projected rows: STEPS, SDS or Faster Innovation Case.
• Last updated - the per-series refresh stamp, so staleness is a column not a mystery.

Everything analytical beyond that - process-route splits, abatement levers, capacity age profiles, in-use stocks, hydrogen and CCS figures, the ~640-design technology guide - folds under additional fields on request: named above, confirmed at sampling, delivered with your extract rather than promised blind.

Geography: global first. All three tracking series publish worldwide totals, with regional and fuel-level detail concentrated on the energy-demand side - enough to see which economies carry the sector's energy bill. Country-specific outputs exist but are thematic, like the India standards brief, not a national statistics service.

Temporal: two clocks, clearly separated. Observed values run annually across 2015-2024, anchored to World Energy Outlook 2025 vintage and refreshed in place - the production chart carries a 18 June 2026 stamp. Scenario work projects forward to 2050 across STEPS, SDS and the Faster Innovation Case, with the Sustainable Development Scenario reaching 2070.

Granularity: annual, sector-wide. One print per year per series at global or regional level - no monthly noise, no facility identifiers. That restraint is the feature: an annual benchmark layer joins cleanly underneath higher-frequency commercial data instead of fighting it for the primary key.

API, files, or your warehouse. Daily, weekly, or hourly.

Pick the channel your stack already speaks; the rows arrive identical either way - series, period, value, region, fuel, scenario, last-updated stamp. Name the years and regions when you request the sample: the full 2015-2024 tracking set, a single-country cut, or the roadmap scenario tables alone. The sample ships first either way, and the ongoing feed lands on whatever cadence your models need.

Every delivery includes the field dictionary above plus the sample rows, so your first join attempt happens against evidence rather than hope.

A global energy-and-emissions ledger earns its keep in four specific jobs:

• ESG and sustainability teams - sector-level decarbonisation disclosure needs a baseline nobody can accuse you of picking. The tracked CO2 series plus the roadmap's mitigation-lever split supply that baseline, and the scenario labels let you state explicitly which future each projection assumes.
• Market researchers and consultants - sizing steel demand starts with the authoritative account of how much energy and material the sector actually moves. End-use demand and in-use stock figures from the roadmap frame any addressable-market model built on top.
• Investors and quant researchers - asset-stranding theses live or die on capacity age and emissions intensity. The capacity-age-profile and process-route dimensions turn a narrative about blast-furnace retirement into a testable panel.
• Journalists, academics and students - the citation-grade original other outlets paraphrase. When an '8 percent of global energy' claim has to survive an editor, quoting the agency's own series beats quoting someone's chart of it.

Market Researchers & Consultants get the authoritative denominator behind every steel market model - the sector's energy share, production volumes and end-use demand straight from the source other analysts cite secondhand. Investors & Quant Researchers get a clean annual panel for stranding-risk screens: emissions, tonnage and capacity age on one clock, scenario-labelled so a policy-shock backtest knows which future it is pricing. Sustainability & ESG Teams get a baseline with institutional weight - the series regulators and NGOs already quote, delivered typed rather than re-typed from charts. Data Scientists & ML Engineers get a small, fully-typed annual cube whose seven fields join on period and region alone, ideal as macro features conditioned on scenario. Start from the steel data hub, then read best steel datasets for where this record ranks against commercial tonnage and price feeds.