Is Korea’s Construction Rerating Coming? The Answer Is Nuclear, Power and AI Infrastructure EPC

Sun, 14 Jun 2026 22:30:00 +0900

Connection This is a follow-up to Korea’s $350bn U.S. strategic investment law and Team Korea nuclear opportunity and Hyundai E&C and Woojin in U.S. nuclear expansion. Also read ETF flows are driving Korea’s volatility, AI data center capex and Korea’s bottlenecks, U.S. non-semiconductor rerating translated into Korea, the Korea Daily Market Hub and the Exclusive Analysis Hub.

TL;DR

The conclusion is not a broad construction-sector rerating. The names that can rerate are those being reclassified from housing contractors into nuclear, SMR, AI data center power, energy EPC and reconstruction EPC.
The old frame was construction = domestic housing + PF risk + low-margin EPC + low ROE. The new frame needs to become construction = nuclear/SMR + AI data centers + power infrastructure + energy EPC + reconstruction + development/operation income.
The clearest axes are Hyundai E&C = U.S. nuclear/SMR, Samsung C&T = data centers, SMR and renewable development, DL E&C = undervalued reconstruction and SMR optionality, and Samsung E&A = Middle East and energy EPC optionality.
This is not a sector-wide buy signal. Hyundai E&C is the cleanest direct exposure but has chase risk, DL E&C has the best asymmetry, Samsung C&T is the highest-quality platform but already reflects much of the story, and Samsung E&A needs flow confirmation.
Ranking: Hyundai E&C pullback buy, DL E&C conditional buy, Samsung C&T hold/pullback only, Samsung E&A wait, and GS E&C/Daewoo E&C as trading candidates.

Core View

A construction rerating can happen. But the multiple goes to power, nuclear, data center, energy and reconstruction EPC operators, not to generic housing contractors.

1. Why revisit construction now?

Korean construction stocks have been assigned low multiples for a long time. The reasons were clear: domestic housing cyclicality, PF risk, low-margin overseas EPC, and weak ROE.

That old market frame was:

construction = domestic housing + PF risk + low-margin EPC + low ROE

Under that frame, rerating is hard. Housing is cyclical, PF losses rarely end in one quarter, and overseas EPC can destroy margin when costs move against the contractor.

But the 2026 question is different:

construction = nuclear/SMR + data centers + power infrastructure + energy EPC + reconstruction EPC

That changes the debate. The new drivers are not domestic apartments. They are electricity for AI data centers, the U.S. nuclear renaissance, Korea-U.S. policy capital, Ukraine and Middle East reconstruction, and energy infrastructure such as renewables, ESS, gas and ammonia.

The important point is selectivity. This is not a blanket buy call on all builders. It is a reclassification trade. The market can price a housing-heavy contractor very differently from an EPC operator tied to nuclear, power, data centers and energy infrastructure.

2. What must happen for rerating: P × Q × C

The rerating needs price, quantity and cost to move together.

Factor	Rerating condition	What to verify
P, price	The market pays an infrastructure EPC multiple, not a housing multiple	Is the company being reclassified as nuclear, data center or energy infrastructure EPC?
Q, quantity	U.S. nuclear, SMR, data center, reconstruction and energy EPC orders increase	Do MOUs convert into FEED, EPC, O&M, equipment, development equity or recurring work?
C, cost	Domestic housing losses, PF provisions and overseas cost overruns stabilize	Does backlog convert into profit?

Construction stocks can move on order headlines. But true rerating happens after the order: backlog needs to expand, the mix needs to improve, margin needs to hold, and housing losses need to stop offsetting the new story.

So the key question is:

Which Korean construction names can be read as AI power, nuclear and infrastructure EPC rather than housing stocks?

3. Catalyst 1: Korea-U.S. strategic investment and U.S. nuclear

The Korea-U.S. strategic investment law is the first catalyst.

LawTimes describes the law as the legal framework for implementing the $350bn strategic investment package agreed under the Korea-U.S. MOU: $200bn of U.S.-bound strategic direct investment and $150bn of shipbuilding cooperation investment. (LawTimes)

Korea’s Ministry of Trade, Industry and Resources said the implementing decree establishes a 20-year Korea-U.S. Strategic Investment Corporation, a statutory capital base of KRW 2tn, and a structure that can delegate work to KEXIM, KDB, K-SURE, KIC, KOBC and KIND. The ministry said the law and decree are expected to take effect on June 18, 2026. (MOTIR)

Two caveats matter.

First, this is not a guaranteed order book for Korean companies. MOTIR also notes that actual projects will be determined after commercial-rationality reviews, operating committee deliberation, National Assembly reporting and U.S. negotiations.

Second, energy, AI and supply-chain projects are natural candidates. AI data centers require electricity, and the U.S. is pushing nuclear permitting and nuclear supply-chain rebuilding at the same time.

The U.S. Department of Energy has framed the policy goal as expanding U.S. nuclear capacity from around 100GW today to 400GW by 2050, while discussing TerraPower, Holtec, TVA, Palisades, uprates and AI-linked energy demand. (U.S. DOE)

Hyundai E&C has the most direct connection. The company announced a FEED contract with Fermi America for four large nuclear reactors inside an integrated energy and AI campus. The project combines AP1000 nuclear 4GW, SMR 2GW, combined-cycle gas 4GW, solar and BESS 1GW, for an 11GW independent power infrastructure tied to hyperscale AI data centers. Hyundai E&C said the parties are discussing an EPC contract targeted for the first half of next year. (Hyundai E&C)

That is the type of project that can change the market’s language. It is not housing. It is AI campus power, nuclear, SMR, gas, solar, batteries and data centers in one package.

4. Catalyst 2: AI data center power bottlenecks

The second catalyst is the data center power bottleneck.

IEA estimates global data center electricity consumption at about 415TWh in 2024 and projects it to reach about 945TWh by 2030 in its Base Case. The same report notes that data centers can be operational in two to three years, while the broader energy system requires longer planning, permitting and construction lead times. (IEA)

That is why EPC matters.

AI data centers are not built with GPUs alone. They need land, grid access, substations, transmission, cooling, backup power, gas generation, renewables, batteries, nuclear, mechanical and electrical systems, piping and project management.

But the market does not simply buy old-style construction. In the U.S., the rerating has concentrated in power grid, substation, electrical/mechanical, cooling, piping and modular construction names, not in generic EPC.

U.S. peer	Rerating driver	Read-through
Quanta Services	Grid, transmission, substations, renewables EPC	Direct AI power-grid bottleneck
EMCOR	Electrical/mechanical systems, data centers, industrial infrastructure	Data center systems and project execution
Comfort Systems	Mechanical systems, HVAC, cooling, piping	Data center cooling and power-density bottleneck
Fluor	Traditional EPC, energy/infrastructure	Can rerate, but premium is more limited

The message is simple.

The market is buying bottleneck execution, not construction as a label.

In Korea, that frame points to Samsung C&T, Hyundai E&C, Samsung E&A, DL E&C and HanmiGlobal. The multiple depends on whether each name solves a real data center, power, nuclear or energy infrastructure bottleneck.

5. Catalyst 3: Ukraine and Middle East reconstruction are options

Reconstruction matters, but it requires more caution.

The World Bank, the European Commission, the UN and Ukraine’s government estimate Ukraine’s reconstruction and recovery needs at almost $588bn over the next decade as of end-2025. Housing, transport and energy are among the most affected sectors, and energy-sector damaged or destroyed assets increased by around 21% versus the prior assessment. (World Bank)

That is a large TAM, but it is not yet revenue for listed companies.

Stage	Meaning	What investors need
Reconstruction need	TAM	Not revenue yet
Multilateral/government financing	Payment capacity	MDB, ECA, guarantees and insurance
Project package	Real opportunity	Energy, roads, ports, housing or industrial sites
Contract type	Margin and risk	EPC, EPCM, PM, equipment, O&M
Revenue recognition	Earnings	Cost inflation, FX, security and sanctions risk

The same applies to the Middle East and Iran. Samsung E&A, DL E&C, Hyundai E&C and Daewoo E&C can benefit from energy EPC or reconstruction headlines. But without sanctions clarity, financing, guarantees and order structure, this remains an option.

So the reconstruction axis should be framed this way:

Reconstruction can move stock prices. But the core is still AI power, nuclear and energy EPC.

6. Korean candidate matrix

The table below uses Thesis OS local DB as of the June 12, 2026 close. The flow column looks at recent 10-day foreign plus real-money flow, giving more weight to quality institutional money than to headline institutional totals.

Company	Price	20D	60D	Target upside	10D foreign + real money	View
Hyundai E&C	KRW 157,500	-2.0%	-1.3%	+40.6%	+KRW 160.5bn	Top priority, but no chase
Samsung C&T	KRW 432,000	+0.8%	+55.1%	+12.4%	-KRW 10.6bn	Best quality, less comfortable price
DL E&C	KRW 73,900	-11.2%	+57.2%	+69.8%	+KRW 42.9bn	Best asymmetry
Samsung E&A	KRW 47,600	-10.2%	+51.8%	+39.5%	-KRW 49.5bn	Good theme, wait for flow
GS E&C	KRW 28,800	-13.3%	+28.6%	+69.5%	+KRW 47.6bn	Housing beta plus reconstruction option
Daewoo E&C	KRW 21,850	-24.5%	+82.4%	+58.4%	+KRW 78.4bn	Event-driven, high volatility
HanmiGlobal	KRW 19,340	-21.2%	-1.4%	+89.6%	-KRW 1.5bn	Data center/reconstruction PM option, weak liquidity

Target upside alone is not enough. A large gap often exists because the market still discounts housing risk, PF risk, margin volatility or weak liquidity.

7. Hyundai E&C: the clearest nuclear and SMR leader

Hyundai E&C is the most direct name in this theme.

Its classification is Idiosyncratic Alpha. It has the Fermi large nuclear FEED contract, Holtec SMR-300 cooperation, and a U.S. nuclear execution story.

Strength	Why it matters
Direct exposure	The Fermi America FEED contract is official
AI power linkage	Fermi is an AI data center plus independent power infrastructure project
Nuclear EPC scarcity	Large nuclear EPC and PMO execution capability is scarce
Policy capital fit	The project aligns with U.S.-bound energy and AI investment themes
Flow	10D foreign + real money is +KRW 160.5bn

But a good company and a good entry point are different things. Hyundai E&C has already moved as the nuclear/SMR leader. After a sharp short-term rally, chasing is less efficient.

View: Wait / Pullback Buy.

The key is whether FEED converts to EPC. FEED is important, but EPC contract scope, risk sharing and margin determine earnings. If Hyundai E&C is recognized as the execution bottleneck, not merely a low-margin contractor, the multiple can move higher.

8. DL E&C: the cheapest reconstruction and SMR option

DL E&C has the best asymmetry among the candidates.

Its classification is Idiosyncratic Alpha. It is less direct than Hyundai E&C, but valuation is more comfortable. Based on the local dataset used here, DL E&C trades around 8x forward PER and 0.6x PBR, while target upside is +69.8%. The 20D return is -11.2%, while the 60D return is +57.2%, which means the larger trend is alive but the near-term heat has cooled.

DL E&C has two axes.

First, reconstruction optionality. If Ukraine or Middle East reconstruction turns into funded projects, discounted construction and plant EPC names can rerate quickly. But this remains an option until contracts, financing and margins are visible.

Second, SMR optionality. X-energy has highlighted Korean partners including DL E&C and Doosan Enerbility, and DL E&C is tied to SMR standardization design and collaboration. It is less direct than Hyundai E&C’s Fermi exposure, but lower valuation increases sensitivity to positive news.

View: Conditional Buy.

DL E&C is less direct, but the price is more forgiving. If the trade moves from “buy the leader” to “rerate undervalued second-line EPC,” DL E&C can have better expected return.

9. Samsung C&T: highest quality, less comfortable entry

Samsung C&T is a different type of stock.

Its classification is Quality Compounder. It is not just a builder. It is closer to a Samsung-group NAV, high-tech EPC, data center, renewable, ESS, IPP and SMR-option platform.

Axis	Meaning
Data centers	North American development and power infrastructure optionality
Renewables/ESS	Solar, storage, power sales and development
SMR/nuclear	Energy-transition projects and Samsung group optionality
High-tech EPC	Semiconductor, battery and bio manufacturing EPC record
NAV	Samsung group stake value and shareholder-return optionality

The issue is price. The local dataset shows Samsung C&T is up +55.1% over 60D, with target upside narrowed to +12.4%. Recent 10D foreign + real-money flow is -KRW 10.6bn, suggesting consolidation rather than fresh accumulation.

View: Hold / Pullback only.

The long-term quality is strong. But for new entry, Hyundai E&C and DL E&C are more compelling. Samsung C&T becomes more interesting after a pullback or after an official data-center development catalyst.

10. Samsung E&A: good energy EPC theme, but flow has not turned

Samsung E&A matters for the U.S. peer read-through.

If Quanta, EMCOR and Comfort Systems can rerate because they solve power, mechanical and data center bottlenecks, Samsung E&A needs to be read as more than chemical EPC. The rerating path requires energy transition, hydrogen/ammonia, gas, data-center power systems and Samsung-group high-tech infrastructure.

The research input is constructive. Samsung E&A’s 1Q26 revenue was KRW 2.2674tn, operating profit KRW 188.2bn and net profit KRW 163.3bn. New orders were KRW 4.6tn and backlog around KRW 20.6tn. That backlog base is not weak.

But stock selection also needs flow.

The local dataset shows Samsung E&A is up +51.8% over 60D and down -10.2% over 20D. Target upside is +39.5%, but 10D foreign + real-money flow is -KRW 49.5bn.

The theme is good, but the money is not yet back.

View: Wait.

It needs to reclaim KRW 48,000 and show foreign-selling exhaustion. Energy EPC headlines alone may not be enough without orders and flow.

11. GS E&C, Daewoo E&C and HanmiGlobal

These are secondary candidates, not core positions for this framework.

Company	Interpretation	View
GS E&C	Housing beta plus reconstruction option. Upside looks large, but housing and PF discounts remain.	Trading candidate
Daewoo E&C	Czech nuclear, Team Korea and reconstruction beta. Less direct to U.S. nuclear than Hyundai E&C.	Event trading
HanmiGlobal	Data center and reconstruction PM optionality, but liquidity and flow are weaker.	Watch

Daewoo E&C is often mentioned in nuclear baskets, but if the core question is U.S. nuclear, AI power and infrastructure EPC, it is less direct than Hyundai E&C. HanmiGlobal has an interesting PM angle, but liquidity and flow make it a watchlist name rather than a core position.

12. Entry strategy

This is a pullback and flow-confirmation market, not a blanket buy market.

Rank	Company	Action	Entry	Invalidation
1	Hyundai E&C	Pullback Buy	Support around KRW 150,000-155,000, or KRW 165,000 breakout with continued foreign/real-money flow	Fermi/Holtec EPC delay, or U.S. strategic investment first project not going toward nuclear
2	DL E&C	Conditional Buy	KRW 72,000-75,000 support plus sustained foreign/real-money flow	Housing cost risk returns, or overseas orders fail to convert into numbers
3	Samsung C&T	Hold / Pullback only	Recovery after a 20D moving-average break, or official data-center development catalyst	Samsung Electronics stake-value pass-through weakens, or data center/SMR options fail to become contracts
4	Samsung E&A	Wait	Reclaim KRW 48,000 plus foreign selling slows	Oil price shock, Middle East order delays, order-margin deterioration
5	GS E&C/Daewoo E&C	Trading	News flow plus flow confirmation	Housing risk returns, weak absorption of retail flow

13. What proves or kills the thesis?

Proof point	Meaning
FEED converts to EPC	Most important proof for direct names like Hyundai E&C
Backlog quality improves	Low-margin contracting versus high-value infrastructure EPC
Data-center development becomes official	Key for Samsung C&T, HanmiGlobal and power-infra chains
Foreign + real-money flow persists	Confirms institutional sponsorship beyond theme trading
Housing losses stabilize	Prevents old discounts from overpowering new rerating

Risk	Meaning
Only MOUs, no EPC	Policy headlines do not become revenue
Fixed-price contracts and cost inflation	Orders hurt margin instead of helping earnings
PF and housing losses return	Old construction discount overwhelms infrastructure story
Reconstruction financing is delayed	Ukraine/Middle East optionality stays as news flow
Foreign and real-money flow exits	Theme exists, but money does not follow

Final view

A construction rerating is possible. But the market is not buying the entire construction sector.

The question is how each company is reclassified.

Hyundai E&C can be reclassified as U.S. nuclear, SMR and AI power infrastructure EPC. It is the most direct name, but the entry should be disciplined.

DL E&C has the best asymmetry. It is less direct than Hyundai E&C, but its valuation and target gap are more comfortable.

Samsung C&T is the highest-quality platform, but much of the story has already been priced.

Samsung E&A has a good Middle East and energy EPC narrative, but flow needs to turn first.

Fund-manager summary:

The leader is Hyundai E&C, the asymmetry is DL E&C, the quality platform is Samsung C&T, and the event option is Samsung E&A. Do not buy construction broadly. Buy only names that can be explained as AI power, nuclear, energy or reconstruction EPC.

Evidence and Limits

Evidence

Korea-U.S. strategic investment law and decree: LawTimes, MOTIR
U.S. nuclear policy: U.S. DOE
Hyundai E&C Fermi America FEED: Hyundai E&C newsroom
Data center electricity demand: IEA Energy and AI
Ukraine reconstruction needs: World Bank RDNA5
Price, flow and target-price data: Thesis OS local DB, June 12, 2026 close

Limits

As of June 14, 2026, the first concrete project under Korea’s strategic investment law has not been finalized.
Fermi FEED is important, but final EPC contract and margin structure remain unconfirmed.
Ukraine and Middle East reconstruction are large TAMs, but financing, insurance, sanctions and contract structure remain uncertain.
Stock prices and consensus estimates can change after the June 12 close used in the dataset.

Ukraine Reconstruction on Korea Invest Insights