South Korea's $3.1 Trillion Industrial Bet Outpaces Its Grid
Wood Mackenzie forecasts a 2.3 GW regional capacity shortfall by 2030 as South Korea's semiconductor and AI expansion pushes peak demand well above planned generation capacity.
National Grid on Monday (2026-08-17) committed GBP 1.2 billion ($1.63 billion) to reconductoring contracts under the second phase of its Electricity Transmission Partnership — part of a GBP 31 billion transmission investment programme running to 2031. The scale of that commitment illustrates what grid spending proportionate to a major industrial build-out can look like. South Korea is attempting a comparable transformation: a KRW470 trillion ($3.1 trillion) national programme anchored around semiconductor and AI data centres, without grid investment matched to that ambition.5
Wood Mackenzie put numbers to the gap in a July 2026 report. The consultancy said the Honam cluster — South Korea's flagship $528 billion semiconductor and AI data centre zone — could push regional peak electricity demand to 19.1 GW once fully built out, up from an estimated 11.8 GW as of the report's publication on July 13 (2026-07-13). Under current plans, effective regional generating capacity will reach only 16.8 GW by 2030, leaving a shortfall of roughly 2.3 GW.3
The solar arithmetic is stark. Closing a 2.3 GW gap through solar alone would require approximately 19.5 GW of new installations within five years. South Korea added only 679 MW of solar capacity region-wide in 2025 — less than 4% of what would be needed annually at that pace. Wood Mackenzie described the trajectory as unrealistic without a rapid battery energy storage build-out running alongside it.3
But the 2.3 GW headline figure already assumes full execution of existing generation plans. Data centres and semiconductor fabs require firm, uninterruptible supply that solar cannot provide on its own. Extending the Hanbit Nuclear Power Plant's operating life by 10 years could preserve about 2 GW of capacity otherwise set to retire; roughly 3 GW of Hanbit's 5.9 GW fleet is scheduled to reach end-of-life by 2035, according to Wood Mackenzie. Without that extension, the shortfall grows before it shrinks.3
South Korea's power market design adds another layer of difficulty. The Korea Chamber of Commerce and Industry said on Wednesday (2026-06-10) that the current market structure is not equipped to handle surging AI-driven demand or to integrate renewables at the pace required. The group called for reform without specifying the mechanisms it had in mind. That absence of detail matters: market reform in electricity systems typically takes years to implement, and the Honam cluster's demand is arriving on a faster schedule.4
The industry has a financing problem too. South Korean state-run power generators have been unable to provide long-term financial forecasts for their coal assets — including projected utilisation rates, early closure costs and carbon-related losses beyond 2030 — according to a May 2026 report. That opacity complicates the government's 2040 coal phaseout and makes it harder to build the investment case for replacement capacity.1
KEPCO, the state utility, has been tightening tariffs since 2021. EV charging discounts were removed in June 2022, household subsidies for low consumption were phased out the same year, and a 50% discount for customers with self-installed renewables was terminated by December 2023. Whether those adjustments translate into improved balance-sheet capacity for grid and generation investment is not apparent from KEPCO's available disclosures.2
The Wood Mackenzie forecasts carry real uncertainty. Demand projections for a single industrial cluster of this scale assume the semiconductor and AI investment runs to schedule and to capacity — a cyclical industry with a documented history of timing slippage. If the Honam build-out runs late, the power gap narrows. If it runs ahead of current assumptions, 2.3 GW understates the problem.
The Hanbit nuclear extension decision is the clearest near-term indicator of how seriously South Korea's government intends to close the firm-capacity side of the equation. Roughly 3 GW of that plant's fleet faces retirement by 2035. A decision to extend — or not — will arrive well before the Honam cluster reaches peak demand, and it will say more about South Korea's power trajectory than any renewable target announced so far.3