Korea’s Power Equipment Boom: How AI Data Centers Are Reshaping the Grid
Quick Take
The artificial intelligence investment cycle is no longer only about GPUs and semiconductor capacity. It is increasingly becoming a power infrastructure story.
Data centers require new generation, substations, transformers, switchgear and distribution equipment — and many of those assets take longer to manufacture and connect than the servers they ultimately power. The International Energy Agency expects global data-center electricity consumption to more than double to around 945 TWh by 2030, while U.S. data centers alone could consume roughly 9.5%–15.3% of national electricity by the end of the decade under Lawrence Berkeley National Laboratory scenarios.
That mismatch between rapidly deployable computing infrastructure and slower-moving power infrastructure is benefiting Korean electrical-equipment manufacturers.
HD Hyundai Electric, Hyosung Heavy Industries and LS Electric are exposed to different parts of the opportunity. The important investment question is therefore not simply whether AI data centers will use more electricity, but where the next grid bottleneck appears, how long equipment pricing remains favorable, and which Korean suppliers can convert record backlogs into cash flow without destroying returns through excessive capacity expansion.
AI Data Centers Are Turning Computing Growth Into a Grid Problem
AI infrastructure has an unusual characteristic: enormous computing capacity can be installed relatively quickly, but the physical power network surrounding it cannot.
The IEA estimates that global data-center electricity demand could rise at roughly 15% annually between 2024 and 2030 — more than four times the growth rate of electricity consumption from other uses. Electricity consumed by accelerated computing servers, the hardware most closely associated with AI workloads, is expected to grow even faster.
The United States is at the center of this shift. The IEA expects data centers to account for nearly half of the increase in U.S. electricity demand through 2030. Lawrence Berkeley National Laboratory's June 2026 update estimates that data centers could represent about 11.8% of total U.S. electricity consumption in its central 2030 projection, with a scenario range of 9.5% to 15.3%.
This creates a problem that semiconductor investors sometimes overlook.
A hyperscale data center needs more than electricity generation. Power must move through the transmission network, enter substations at the correct voltage, pass through transformers and switchgear, and finally reach server infrastructure with extremely high reliability.
In simplified form:
Generation → Transmission Grid → Substation → High-Voltage Transformer → Switchgear → Medium/Low-Voltage Distribution → Data Center Electrical System → Servers
A shortage at almost any point can delay the entire project.
Analysis: This changes the economic importance of power equipment. A transformer or switchgear unit may represent only a fraction of a data center's total construction cost, but an unavailable unit can prevent billions of dollars of computing equipment from becoming productive.
That makes delivery time, manufacturing slots and proven reliability unusually valuable.
Why Korea Is Well Positioned in the Power Equipment Cycle
Korea did not enter the global power-equipment market because of generative AI. Korean manufacturers had already spent decades developing high-voltage transformers, switchgear and related grid equipment for domestic utilities and export markets.
AI is accelerating demand into an industry where Korean suppliers already had factories, engineering expertise, certification histories and customer references.
Korea Customs Service data illustrates the shift. Korean exports of major power equipment increased from approximately USD 4.54 billion in 2021 to USD 7.10 billion in 2024. During January–November 2025, exports reached about USD 7.13 billion, already exceeding the full-year 2024 level and rising 11.3% year over year. Transformers alone accounted for approximately USD 2.33 billion, or 32.6% of the total. The United States was the largest export market.
The important point is not simply that exports increased.
Two demand cycles are overlapping:
First, new demand. AI data centers, semiconductor fabs, electrification and renewable-energy infrastructure are creating additional loads.
Second, replacement demand. The United States and other developed markets need to modernize aging transmission and distribution infrastructure even without AI.
This distinction matters because it makes the current cycle less dependent on a single technology theme than the phrase “AI power demand” might imply.
If AI spending slows, grid replacement does not disappear. If grid modernization slows, hyperscalers may still need dedicated electrical infrastructure to obtain power.
That does not eliminate cyclicality, but it broadens the demand base.
The Three Korean Leaders Are Not the Same Investment
The three largest Korean names exposed to this theme participate in different sections of the electrical system.
| Company | Q2 2026 Revenue | Q2 Operating Profit / Margin | Reported Backlog | Core Exposure |
|---|---|---|---|---|
| HD Hyundai Electric | KRW 1.142 trillion | KRW 287 billion / 25.1% | USD 8.49 billion | Large power transformers, distribution equipment, rotating machinery; strong U.S. exposure |
| Hyosung Heavy Industries | KRW 1.687 trillion consolidated | KRW 264.3 billion / 15.7% calculated consolidated margin | Approximately KRW 17.5 trillion | Extra-high-voltage transformers and grid equipment; expanding U.S. localization |
| LS Electric | KRW 1.577 trillion | KRW 179 billion / 11.3% | KRW 7.0 trillion | Switchgear, transformers, automation and increasingly data-center/DC distribution solutions |
The margin comparison is particularly revealing.
HD Hyundai Electric's Q2 operating profit increased 37.3% year over year against 26.0% revenue growth. LS Electric's operating profit increased 64.4% against 32.2% revenue growth. Hyosung Heavy Industries' consolidated operating profit increased 60.9% while revenue rose only 10.6%.
Analysis: The industry is not merely shipping more equipment. Recent profitability suggests that product mix, factory utilization and pricing are also working in suppliers' favor.
For investors, that distinction is crucial. Volume growth creates revenue. Scarcity plus favorable mix creates operating leverage.
HD Hyundai Electric: Scarce Transformer Capacity Meets High Margins
HD Hyundai Electric currently provides one of the clearest examples of the transformer shortage translating into earnings.
In Q2 2026, the company reported:
- Revenue of KRW 1.142 trillion, up 26.0% year over year
- Operating profit of KRW 287 billion, up 37.3%
- Operating margin of 25.1%
- Quarterly orders of USD 1.44 billion
- First-half orders of USD 3.24 billion
- Order backlog of USD 8.49 billion, up 29.6% year over year
The company specifically connected data-center expansion with stronger demand for distribution equipment and rotating machinery.
The 25% operating margin is notable for an industrial equipment company. It suggests HD Hyundai Electric is benefiting from more than simple end-market growth.
Its current economics reflect a combination of constrained global transformer supply, selective order intake, higher-value products and strong U.S. demand.
The company is also expanding its American manufacturing base. Its Alabama transformer operation has increased annual capacity over time to approximately 105 units, and the second Alabama plant under construction is expected to raise transformer output capacity by roughly another 50% when completed. The expanded facility is also intended to handle equipment up to 765 kV.
HD Hyundai Electric also announced a KRW 173 billion U.S. order in 2026 covering 765 kV extra-high-voltage transformers and reactors.
Investment implication: HD Hyundai Electric currently offers perhaps the cleanest combination of high-voltage equipment scarcity, strong backlog visibility and elevated profitability among the three Korean leaders.
The risk follows directly from that strength.
If global transformer capacity catches up with demand, the company's unusually high margins could normalize even if revenue continues growing. Investors therefore need to distinguish between structural demand growth and temporarily exceptional pricing power.
Hyosung Heavy Industries: U.S. Localization Becomes the Competitive Strategy
Hyosung Heavy Industries occupies a particularly strong position in extra-high-voltage equipment.
The company's Q2 2026 consolidated revenue reached approximately KRW 1.69 trillion, while operating profit increased 60.9% year over year to KRW 264.3 billion. Its heavy-industry segment generated roughly KRW 1.14 trillion of revenue and KRW 229.8 billion of operating profit, equivalent to a segment operating margin of about 20.2%.
Order momentum has been even more striking.
Reported first-half 2026 new orders reached roughly KRW 7.5 trillion, while the backlog at the end of Q2 was approximately KRW 17.5 trillion. The company raised its full-year order target as U.S. and other international power-grid demand accelerated.
The more important strategic development is localization.
In June 2026, Hyosung HICO agreed to establish a U.S. joint venture with a subsidiary of Quanta Services. The venture, HYOSUNG HICO BREAKER, is designed to manufacture high-voltage circuit breakers ranging from 72.5 kV to 800 kV in Pennsylvania, with production planned to begin in October 2026.
Hyosung Group explicitly tied the investment to growing power demand from AI, data centers and U.S. grid modernization.
The partnership is strategically interesting because Quanta is not simply an equipment buyer. It is one of North America's largest electric-infrastructure engineering and construction networks. That can potentially give Hyosung greater proximity to utility and data-center projects.
Hyosung has also invested approximately USD 300 million in its Memphis transformer manufacturing operation from acquisition through its current expansion program.
Analysis: The competitive advantage is gradually shifting from “Can you export a transformer to America?” toward “Can you manufacture a broader set of critical grid equipment inside America and integrate yourself into the local project ecosystem?”
Hyosung appears to be pursuing exactly that transition.
One caveat for equity investors is that Hyosung Heavy Industries is not as pure an electrical-equipment exposure as HD Hyundai Electric. Its consolidated results also include the construction business. Strong power-equipment economics can therefore be partially diluted by unrelated segment volatility.
LS Electric: The Data Center Opportunity Is Moving Downstream
If HD Hyundai Electric and Hyosung Heavy Industries are primarily associated with the large-transformer bottleneck, LS Electric provides exposure to another increasingly important layer: switchgear, distribution systems and the electrical architecture inside or near the data center.
Q2 2026 revenue rose 32.2% year over year to approximately KRW 1.58 trillion, while operating profit jumped 64.4% to KRW 179 billion.
More strikingly, new orders reached KRW 2.08 trillion, up 243% year over year, while backlog expanded 81.5% to KRW 7.0 trillion.
Within that backlog:
- Switchgear reached approximately KRW 2.00 trillion, up 109.4% year over year.
- High-voltage transformer backlog reached approximately KRW 3.35 trillion, up 88.1%.
- Medium- and low-voltage transformer backlog rose 144.5%.
The company specifically identified U.S. and Korean data centers and on-site generation as contributors to its broader customer base.
There is also direct customer evidence.
In May 2026, LS Electric announced an approximately USD 70 million power-equipment contract for a U.S. big-tech data center, including high-end distribution equipment such as vacuum circuit breakers.
LS Electric is additionally developing equipment for higher-voltage DC distribution architectures as data centers search for ways to reduce conversion losses between the grid and increasingly power-dense AI computing systems.
Analysis: This could become important as the bottleneck migrates.
The first phase of the power-equipment boom was dominated by large transformers. But once new transformer factories come online, the constraint may shift toward switchgear, breakers, substations, distribution systems and grid interconnection.
That potentially gives LS Electric a different earnings pathway from the two transformer-heavy competitors.
The balance-sheet trade-off should not be ignored. LS Electric reported higher net debt as its backlog and working-capital requirements expanded. Rapid order growth is positive only if the company converts those orders into profitable cash flow without requiring disproportionate financing.
How Korean Suppliers Compare With Global Power Equipment Leaders
Korean suppliers do not operate in an empty market.
Eaton is expanding U.S. medium-voltage switchgear capacity, including a new Nebraska factory, specifically citing rising demand related to data centers and electrification. GE Vernova is adding transformer capacity and offers a broad portfolio spanning high-voltage substations, switchgear, transformers, breakers and generation. Hitachi Energy has highlighted the extreme urgency surrounding transformer delivery for hyperscale data-center projects.
Siemens Energy has also reported a rapid rise in data-center-related orders within its grid technologies business, demonstrating that this is a global capital-equipment boom rather than a uniquely Korean phenomenon.
That creates a useful distinction.
Global leaders such as Eaton, GE Vernova, Hitachi Energy and Siemens Energy often possess broader system-integration, software, service and installed-base advantages.
Korean companies, however, have become particularly relevant where physical equipment capacity itself is scarce.
Their opportunity is strongest when customers care about:
- Available manufacturing slots
- High-voltage engineering capability
- Proven utility references
- Reliable delivery
- Competitive manufacturing cost
- Local U.S. production
- The ability to expand capacity before the shortage disappears
The competitive question is therefore not simply whether Korean products are cheaper.
It is whether Korean manufacturers can become permanent suppliers in global grid infrastructure after today's shortage normalizes.
What Could Drive the Next Earnings Leg?
The current boom has already moved through three stages.
First came stronger orders.
Then came larger backlogs.
Now those backlogs are converting into revenue and higher margins.
The next stage will depend on whether capacity expansion can create another earnings leg without eroding pricing.
For the Korean companies, several mechanisms matter.
Capacity expansion
New U.S. factories can increase addressable revenue and reduce trade, logistics and local-content risks. But they also introduce fixed costs and execution risk.
Product mix
Extra-high-voltage transformers, high-end switchgear and specialized data-center equipment can generate better economics than commodity electrical products.
Customer diversification
Winning business directly from hyperscalers, utilities, engineering contractors and data-center developers reduces reliance on a narrow customer group.
Localization
American manufacturing is becoming strategically more important as utilities and large infrastructure customers prioritize supply-chain security and delivery certainty.
Expansion beyond transformers
The long-term winner may not necessarily be the company with the most transformer capacity today. It may be the supplier able to capture more of the electrical chain surrounding each data center.
The Strongest Counterargument: This Is Still a Capital-Goods Cycle
The bullish argument has considerable evidence behind it, but power equipment has not stopped being cyclical.
Every shortage eventually attracts capital.
HD Hyundai Electric is expanding transformer production. Hyosung is expanding U.S. transformers and breakers. LS Electric is adding U.S. electrical-equipment capacity. Global competitors including Eaton and GE Vernova are doing the same.
That creates the most important counterargument to the current investment thesis:
Today's shortage could become tomorrow's capacity normalization.
Several things could weaken the cycle.
A slowdown in hyperscaler capital expenditure could delay planned data centers. Grid interconnection constraints could postpone projects even when computing demand remains strong. New global factories could shorten delivery times and reduce pricing power. Raw-material costs, labor shortages and working-capital requirements could also absorb some of the earnings benefit.
There is another subtle risk.
The power infrastructure bottleneck itself can delay AI data-center construction. In the short term, that makes scarce equipment more valuable. But if grid access becomes the binding constraint for entire regions, equipment demand can also be deferred because projects cannot proceed.
So the same bottleneck that creates pricing power can eventually become a constraint on order timing.
Valuation: A Boom Does Not Automatically Mean the Stocks Are Cheap
A strong industry does not automatically create undervalued equities.
Investors should avoid concluding that Korean power-equipment stocks deserve higher valuations simply because AI electricity demand is increasing.
A sustained rerating would require evidence that today's earnings power can survive beyond the immediate shortage.
The most important valuation questions are:
Can margins remain structurally above the previous cycle?
If current profitability is mainly shortage pricing, normalized earnings may be lower than reported earnings.
Will new factories earn attractive returns on capital?
Capacity growth only creates value when incremental profit exceeds the cost of the additional capital required.
Can backlog convert into cash?
Large order books are valuable, but working capital, inventories and customer payment schedules determine how much accounting profit becomes free cash flow.
Will Korean companies retain customers after global capacity expands?
The strongest long-term outcome would be for today's emergency suppliers to become permanent preferred vendors.
Can they move higher in the value chain?
Selling a transformer is valuable. Supplying transformers, switchgear, breakers, DC distribution and integrated electrical solutions can produce a broader and more durable opportunity.
Analysis: The sector's next rerating is less likely to come from another headline about AI electricity demand. It would more likely come from evidence that the current shortage has produced a structurally stronger global franchise.
What to Watch
For HD Hyundai Electric, watch whether the Alabama expansion adds volume while the company preserves its unusually high operating margin. A sharp decline in pricing or margin despite higher capacity would indicate that scarcity economics are normalizing.
For Hyosung Heavy Industries, watch U.S. order growth, the ramp-up of its Pennsylvania breaker joint venture, Memphis capacity utilization and whether heavy-industry profitability continues to outweigh volatility elsewhere in the group.
For LS Electric, the most important indicators are switchgear and transformer backlog, additional hyperscaler or data-center customer wins, U.S. factory progress, DC-distribution commercialization and free-cash-flow conversion.
Across the industry, investors should pay particular attention to transformer delivery times, global capacity announcements, U.S. data-center power demand, utility transmission spending and order-to-backlog conversion.
If orders remain strong after the current wave of factory expansion becomes operational, the evidence for a structural cycle becomes much stronger.
If lead times collapse and order growth slows while capacity rises, the market may discover that part of the boom was conventional capital-goods scarcity after all.
Sources & Data
- International Energy Agency — Energy and AI, data-center electricity-demand outlook through 2030.
- Lawrence Berkeley National Laboratory — United States Data Center Energy Usage Report: 2025 Update, June 2026.
- Korea Customs Service — Korean power-equipment export statistics and AI-related electricity-demand analysis, January–November 2025 data.
- HD Hyundai Electric — Q2 2026 earnings results, July 28, 2026.
- HD Hyundai Electric — U.S. manufacturing expansion and 2026 major-order disclosures.
- Hyosung Heavy Industries — Q2 2026 earnings release, July 31, 2026, supplemented by reported segment and order data.
- Hyosung Group — Hyosung HICO and Quanta Services U.S. circuit-breaker joint venture announcement, June 14, 2026.
- LS Electric — Q2 2026 Earnings Release.
- LS Electric — U.S. big-tech data-center power-equipment contract announcement, May 18, 2026.
- Eaton, GE Vernova, Hitachi Energy and Siemens Energy — global grid-equipment capacity and data-center infrastructure disclosures.
Data as of August 21, 2026
Investment Disclaimer
This article is for informational and educational purposes only and does not constitute investment advice. Investors should conduct their own research before making investment decisions.
Comments
Post a Comment