Nonprofit research institute · Seoul, Koreacontact@planit.institute

Prepare for the Electricity Appetite of Steel and Petrochemicals

Korea's steel and petrochemical industries emit over 150 million tonnes of CO₂ per year — 25% of the country's total. Decarbonising them requires electrification, and the scale of the electricity demand that follows is rarely discussed honestly.

The Electrification Routes

Steel: Replacing the blast furnace/basic oxygen furnace (BF-BOF) route with hydrogen direct-reduction and electric arc furnace (H₂-DRI-EAF) cuts CO₂ by 95%. The catch: electric arc furnaces run on electricity.

Petrochemicals: Naphtha crackers currently operate above 800 °C using methane as a heat source. Switching to electric heating (e-cracking) eliminates 94–100% of process CO₂. Again: electricity is the input.

The Numbers

Korea currently produces 67 million tonnes of crude steel annually and 9 million tonnes of ethylene. Maintaining that output while electrifying 100% of both processes would require 345–405 TWh of electricity per year — roughly 60% of Korea's entire 2023 generation (590 TWh).

The steel and petrochemical clusters in Pohang and Gwangyang alone would need 247–283 TWh, six to seven times more power than those regions currently consume. This implies explosive electricity demand growth outside the Seoul metropolitan area — something Korea's grid planning has not adequately addressed.

What the 11th Basic Plan Gets Wrong

The government's 11th Basic Plan for Electricity Supply and Demand (전기본) projects industrial electrification demand of just 63 TWh by 2038. That figure covers only the very early stages of hydrogen-reduction steelmaking and electric cracking introduction. Full electrification for carbon neutrality by 2050 requires more than five times that amount — for steel and petrochemicals alone.

Six more rounds of Basic Plan revisions remain before 2050. Each cycle must grapple seriously with how to project and accommodate industrial electrification at this scale. The conversation cannot wait.

Mode: