Chaos and Order

💡 왼쪽 원문을 읽으면서 오른쪽에 따라 써보세요. Tab 키로 힌트를 받을 수 있습니다.

원문 렌더가 준비되기 전까지 텍스트 가이드로 표시합니다.

Introduction — Reactors Switching Back On

A little over a decade ago, nuclear power was treated as a sunset industry. After the 2011 Fukushima accident, Germany committed to phasing out nuclear, and many advanced economies effectively froze new reactor construction. As the shale revolution drove natural gas prices sharply lower, aging U.S. reactors were retired early one after another on economic grounds. The dominant perception was that nuclear was an expensive, slow, and politically burdensome source of power.

Then, starting around 2023, the mood reversed sharply. Two variables drove the shift. First, a practical recognition that meeting climate targets requires carbon-free electricity that runs reliably around the clock, the so-called baseload. Second, the generative AI boom sent data-center power demand soaring, and Big Tech firms began partnering directly with nuclear operators to secure stable carbon-free power.

The most symbolic event was Constellation Energy's announcement, reported in September 2024. The company said it would restart Unit 1 at Three Mile Island, the very site whose 1979 partial core meltdown traumatized the U.S. nuclear industry, now renamed the Crane Clean Energy Center, and supply that power to Microsoft under a long-term power purchase agreement (PPA) reported to run roughly 20 years. The image of a plant once synonymous with disaster being reframed as a power solution for the AI era captures this narrative's turning point in a single frame.

This post lays out where the nuclear renaissance narrative comes from, how far the data actually supports it, and what the bull and bear cases each rest on, in a balanced way. Finally, it considers the implications for the Korean nuclear ecosystem, often called K-nuclear.

> This article is for information and educational purposes only and is not investment advice or a recommendation. Investment decisions and their consequences are your own responsibility. Consult a qualified professional when needed.

1. Why Nuclear, Why Now — Structural Change on the Demand Side

1-1. The Power Thirst of AI Data Centers

Traditionally, electricity demand grew gently alongside economic growth. In the United States, demand was nearly flat for roughly two decades from the mid-2000s, because efficiency gains offset rising consumption.

AI is now breaking that equilibrium. Training and running large language models requires enormous computation, which translates directly into electricity. The International Energy Agency (IEA) has projected that data-center electricity consumption will rise substantially over the coming years. A key point is that AI-focused data centers pack far higher power density per unit of floor space than conventional ones.

Big Tech's requirement is not simply for a lot of electricity. Because of ESG goals, they want carbon-free power, and to keep data centers running they want power that never cuts out, 24 hours a day. Solar and wind are carbon-free but intermittent. The central logic of the renaissance is that nuclear is nearly the only large-scale source that satisfies both conditions at once.

1-2. Big Tech Power Deals — The Substance of Demand

A reason the nuclear renaissance narrative carries weight beyond mere hope is that overwhelmingly well-funded Big Tech firms are reported to be entering into actual power deals. Below is a summary of representative cases reported in the press. (Terms and timelines are as reported, and variables such as regulatory approval may remain.)

| --- | --- | --- | --- |

Two things are central here. First, as power buyers signal their intent to purchase electricity in the form of long-term contracts, the demand uncertainty that historically undermined nuclear economics is partly resolved. Second, Big Tech has begun committing capital not only to restarting existing large reactors (Microsoft) but also to SMR development that is not yet commercialized (Amazon, Google). That said, most SMR-related deals are pre-purchase or investment arrangements premised on future deployment, so it should be clear that substantial time and uncertainty remain before actual power is delivered.

Spectrum of Big Tech power sourcing (conceptual)

higher certainty lower certainty

| |

[restart existing reactor] - [co-location] - [SMR pre-purchase/invest]

Microsoft-CE Amazon-Talen Google-Kairos

Amazon-X-energy

(immediacy of power up) (more of a future promise)

1-3. The Scarcity of Carbon-Free Baseload

A simplified view of generation sources looks like this.

| --- | --- | --- | --- | --- |

As the table shows, the only sources that are both carbon-free and reliable around the clock are roughly nuclear and hydro. Hydro has almost no suitable new sites left. As a result, nuclear is drawing renewed attention as the realistic option to fill the carbon-free baseload box.

To compare generation sources more quantitatively, it helps to look at three axes together: capacity factor, the approximate range of levelized cost of electricity (LCOE), and carbon emissions per unit of generation. The figures below vary widely by institution, region, and year, so treat them only as conceptual ranges for spotting trends; for real decisions you must consult primary sources such as the IEA, the Lazard LCOE reports, and the U.S. Energy Information Administration.

| --- | --- | --- | --- | --- |

The key takeaway from this table is that nuclear's weakness lies less in the generation cost itself than in upfront capital and construction risk. Once built, a reactor has a high capacity factor and a low fuel-cost share, so its operating-stage economics are relatively sound. Conversely, solar and wind have low generation costs but low capacity factors and intermittency, so they need storage or backup; once those system costs are added in, a simple LCOE comparison gives an incomplete picture.

Scarcity of carbon-free baseload (conceptual)

carbon-free

| [solar] [wind]

| (intermittent)

| [nuclear] <- scarce zone

| [hydro]

----------+---------------------> 24-hour reliability

| [natural gas] [coal]

| (emissions)

2. SMRs — A New Bet on Small Modular Reactors

2-1. What an SMR Is

A conventional large reactor is a massive facility with electrical output of roughly 1,000 to 1,400 megawatts (MW) per unit. Construction can cost billions of dollars and take around a decade from groundbreaking to operation. Cost overruns and schedule delays have been chronic problems.

The small modular reactor (SMR) is an attempt to solve those problems with a different approach. The core ideas are as follows.

- Design output small, on the order of roughly 50 to 300 MW per unit.

- Standardize and modularize major components in a factory, then assemble on site.

- Bundle multiple modules to scale output as needed.

A common analogy is mass-producing cars from standard parts in a factory instead of hand-building each one. In theory, learning effects could lower unit costs, shorten build times, and make siting more flexible.

2-2. Potential Advantages and Unproven Risks of SMRs

| Item | Potential advantage | Not yet proven |

| --- | --- | --- |

| Unit cost | Expected to fall via learning curve | First-of-a-kind cost may stay high |

| Build time | Shorter via modular assembly | Limited commercial operating record |

| Siting | Flexible due to small size | Regulatory and licensing paths immature |

| Safety | Emphasis on passive safety design | No long-run data for new designs |

| Waste | Varies by design | Spent-fuel handling remains common |

A point that must be stressed: as of 2026, the number of Western SMRs that have entered commercial operation is very limited. Many projects remain at the design-certification, site-acquisition, or financing stage. In other words, a large part of the SMR investment narrative still rests on a promise of the future, and that should be recognized clearly.

Notably, one NuScale Power project was reported to have been canceled in 2023 over rising costs and difficulty securing buyers. It is frequently cited as a warning that the SMR concept does not automatically translate into economic viability.

2-3. Comparing Major SMR Developers

SMR is not a single technology but a collection of designs with different cooling methods and fuel forms. Broadly, they split into light-water-reactor (PWR/BWR) families and non-light-water families (sodium-cooled, high-temperature gas, molten salt, and so on). The table below organizes frequently cited developers in simplified form; output, cooling, and licensing stage are conceptual summaries as of reporting and are not recommendations.

| --- | --- | --- | --- | --- |

Two things stand out. First, within the single word "SMR" sit very heterogeneous technologies, from 15 MW to 470 MW, and from proven light-water reactors to unproven fast and high-temperature gas reactors. So "investing in SMRs" is by itself ambiguous; one must distinguish which cooling method and which fuel (especially whether HALEU is required) a design uses. Second, light-water designs (GE Hitachi, Rolls-Royce) build on existing technology and may license relatively quickly, while non-light-water designs (TerraPower, X-energy, Oklo) offer larger potential advantages but face a trade-off of immature regulatory paths and fuel supply chains.

Two branches of SMR design (conceptual)

[light-water family] [non-light-water family]

GE Hitachi BWRX-300 TerraPower Natrium

NuScale VOYGR X-energy Xe-100

Rolls-Royce SMR Oklo Aurora

| |

built on existing tech new fuel (HALEU, etc.), high temp

faster licensing expected higher upside / higher unproven risk

3. The Nuclear Value Chain — What Sits Where

The nuclear investment narrative is not a single stock but a multi-layered value chain. Understanding it makes it easier to see exactly why a given name draws attention.

Nuclear value chain (simplified)

[uranium mining]

Cameco, Kazatomprom, etc.

[conversion / enrichment]

Centrus Energy, Orano, Urenco, etc.

[fuel fabrication]

[reactor design / construction]

large: Westinghouse, KHNP / Doosan, etc.

SMR: NuScale, Oklo, X-energy, etc.

[generation / operation]

Constellation Energy, Vistra, etc.

[power sales / PPA]

Big Tech data centers and other buyers

3-1. Upstream — Uranium and the Fuel Cycle

As reactors multiply, demand for uranium fuel grows too. In uranium mining, Canada's Cameco is a frequently cited name. Uranium prices were reported to be recovering from long-term lows during the 2020s, though one should note the high volatility characteristic of commodity prices.

In enrichment, the U.S.-listed Centrus Energy comes up often. In particular, some next-generation reactors and SMRs require a special fuel called high-assay low-enriched uranium (HALEU), and the fact that this supply chain is not yet mature is seen as both an opportunity and a risk. The enrichment supply chain has also historically been heavily dependent on Russia geopolitically, and analysts note that a reshaping of that supply chain is under way.

3-2. The Uranium Cycle in Depth

The most volatile part of the nuclear value chain, and the one with the hottest investment narrative, is the uranium fuel cycle. Natural uranium passes through several stages before becoming reactor fuel, and each stage forms a distinct industry with its own price.

The uranium fuel cycle (by stage)

[1. mining/milling] -> [2. conversion] -> [3. enrichment] -> [4. fabrication]

U3O8 (yellowcake) UF6 U-235 ratio up fuel rods/assemblies

Cameco Cameco, Orano Urenco, Orano national fabricators

Kazatomprom Centrus (HALEU)

Unpacking each stage:

1. **Mining and milling**: Uranium mined from the ground is refined into yellowcake (U3O8). The world's largest producer is Kazakhstan, where the state firm Kazatomprom holds an overwhelming share. Among Western-listed firms, Canada's Cameco is the representative name.

2. **Conversion**: Yellowcake is converted into uranium hexafluoride (UF6), a gaseous form suitable for enrichment. Conversion capacity is limited worldwide, and conversion prices were at one point reported to have surged.

3. **Enrichment**: This raises the U-235 fraction of natural uranium (about 0.7%) to power-grade levels (3 to 5%). Holders of centrifuge technology such as Urenco, Orano, and Russia's Rosatom have been the main players. Some next-generation reactors and SMRs require high-assay low-enriched uranium (HALEU), with U-235 raised to nearly 20%; the Western HALEU supply chain is still early-stage, and U.S.-listed Centrus Energy is frequently cited here.

4. **Fabrication**: Enriched uranium is made into fuel rods and assemblies and loaded into reactors.

Uranium Price Cycles and Supply-Demand

Uranium prices have historically shown extreme cycles. After a speculative spike around 2007 came a long slump, and following the 2011 Fukushima accident, prices stayed near long-term lows as new demand contracted. As a result, many mines cut output or were mothballed on poor economics, structurally reducing supply.

The mood shifted in the 2020s. Reactor life extensions, expectations of new construction starts, and a Western push to reduce dependence on Russian uranium and enrichment services combined to drive prices off their long-term lows, as reported. On top of that, physical uranium trusts such as the Sprott Physical Uranium Trust (SPUT) bought and stored physical uranium directly, and some analysts argued this absorbed loose spot material and acted as a new variable tightening supply.

Uranium price cycle (conceptual trend, not actual figures)

price

| /\

| / \ (2007 speculative spike)

| / \

| / \____

| / \____ (long slump after Fukushima)

| / \________

| / \____/‾‾‾ (2020s recovery)

+----+----+----+----+----+----+----+----> time

2005 2007 2011 2015 2018 2021 2024

The key supply-demand points:

| Factor | Direction | Explanation |

| --- | --- | --- |

| New reactors / life extensions | Demand up | More operating reactors mean more fuel demand |

| Output cuts from low prices | Supply down | Restarting mothballed mines takes time |

| Reduced Russia dependence | Supply reshaping | Western conversion and enrichment must expand |

| Physical trust (SPUT, etc.) buying | Spot absorption | Locks up loose material, tightens supply |

| New HALEU demand | New demand | High-assay supply for new designs immature |

That said, uranium is a classic commodity with very high price volatility, and a self-correcting mechanism applies: when prices rise, mothballed mines restart and supply grows. So it is dangerous to extrapolate a short-term spike into a long-term trend.

3-3. Mid- and Downstream — Design, Construction, Operation

In large-reactor operation, the most frequently cited U.S.-listed name is Constellation Energy. It is known to hold the largest U.S. nuclear generation portfolio and was the subject of the Microsoft-related reporting mentioned earlier.

In SMR design and development, names like NuScale Power, Oklo, and X-energy come up. That said, many of these have not yet begun meaningful commercial revenue or are loss-making, and their share prices tend to be highly volatile, which should be recognized.

3-4. Frequently Cited Names (Fact-Based, Not Recommendations)

The table below organizes names often mentioned in the narrative. It is not a buy or sell recommendation but a summary of why each is discussed and what its main risks are.

| --- | --- | --- | --- |

4. Construction Risk in Depth — Vogtle and Licensing

The most frequently underestimated variable in nuclear investing is construction risk. The bull narrative draws a simple causal chain: demand rises, reactors multiply, related firms benefit. In reality, reactor construction is an area where cost overruns and schedule delays have recurred structurally.

4-1. The Vogtle Case — A Textbook of Overruns and Delays

The Vogtle Units 3 and 4 in Georgia, U.S., are frequently cited as emblematic of the difficulty of building new large reactors in the West. According to reporting, the project was delayed by years against the original plan, and total cost swelled to well over double the initial estimate. Westinghouse, a major equipment supplier, was reported to have filed for bankruptcy protection in 2017 partly under construction-related burdens, which was bound up with the project's troubles.

The Vogtle case offers three lessons. First, first-of-a-kind units almost always overrun on cost and schedule because of the absence of learning. Second, when a complex supply chain combines with a shortage of skilled labor, delays snowball. Third, all of these costs are ultimately passed on to electricity rates or onto the operator's balance sheet, potentially eroding related firms' profitability.

Structure of reactor cost overruns (conceptual)

planned cost |##########

actual cost |###################### (1st unit, no learning)

-> later units may improve via

learning effects, but one-off

projects struggle to accumulate it

This is exactly the point the SMR bull case rests on: that by accumulating learning through standardization and mass production, the first-of-a-kind problem can be structurally reduced. But that logic holds only if a sufficient number of identical modules are actually built repeatedly, and as of 2026 it remains pre-validation, which must not be forgotten.

4-2. NRC Licensing — Time as a Cost

To build and operate a reactor in the U.S., one must pass licensing by the Nuclear Regulatory Commission (NRC). There are multiple stages, including design certification, the combined construction and operating license (COL), and safety reviews, each accompanied by strict safety standards and public procedures. These are essential for safety, but they are also an entry barrier that demands considerable time and cost.

In particular, new SMR and non-light-water designs often do not fit neatly into a regulatory framework built around conventional light-water reactors, so regulators and developers must jointly establish new review paths. This adds uncertainty to licensing timelines.

| Stage (conceptual) | Content | Investment implication |

| --- | --- | --- |

| Design certification | Safety review of the reactor design | Passing does not soon mean revenue |

| Site and construction permit | Site suitability and construction approval | Local acceptance and environmental variables |

| Operating license | Final pre-operation safety check | Cost accrues if startup is delayed |

| Operation and periodic inspection | Ongoing regulation during operation | Risk of shutdown on safety issues |

From an investment standpoint, the key is that news of a "design certification" does not mean revenue is being generated. Licensing progress is a positive signal, but from there to actual construction, operation, and power sales still requires long time and large capital.

5. The Reality in Data — How Far Have We Come

To check the gap between narrative and reality, consider some commonly cited quantitative indicators in simplified form. (The figures below are conceptual examples meant to show trends; for actual investment decisions you must always check primary sources such as the IEA, the World Nuclear Association, and the U.S. Energy Information Administration.)

Global new reactor construction starts (conceptual, units)

2010 |##########

2015 |######

2020 |########

2023 |###########

2025 |#############

(reporting suggests a recovery in starts,

but absolute numbers are limited vs. past peaks)

Data-center power demand outlook (conceptual)

now |######

2027 |##########

2030 |################

(AI demand growth is the key driver,

but estimates vary widely by forecaster)

Two points stand out. First, there are signals that reactor construction starts are recovering, but the absolute scale remains limited, and it takes a long time for new reactors to contribute to the grid. Second, data-center demand forecasts vary widely across institutions, so treating any single optimistic scenario as a foregone conclusion is dangerous.

6. Two Perspectives — The Bull and Bear Cases

6-1. The Bull Case

The bull case rests on the following arguments.

1. **Structural demand**: AI data centers, electric vehicles, and manufacturing reshoring drive a structural rise in power demand, and the scarcity of carbon-free baseload raises nuclear's value.

2. **Policy support**: Reports indicate that many countries classify nuclear as clean energy and are pursuing tax incentives and streamlined licensing.

3. **Big Tech capital**: Well-funded Big Tech firms underwrite demand with long-term PPAs, partly resolving the demand uncertainty that historically undermined nuclear economics.

4. **SMR potential**: If mass production succeeds, it could structurally improve the cost and schedule problems.

6-2. The Bear Case

The bear case warns of the following.

1. **Construction risk**: Nuclear has historically been prone to cost overruns and delays. Recent large Western reactor projects were also reported to have significantly exceeded their budgets.

2. **SMRs unproven**: There is little commercial operating record, and first-of-a-kind economics are unproven. The NuScale project cancellation is a representative example.

3. **Stretched valuations**: Some SMR and uranium-related stocks are seen as having run well ahead of fundamentals, so volatility could be large if expectations falter.

4. **Regulatory and public-opinion risk**: An accident or political shift could rapidly change policy direction, and spent-fuel handling remains an unsolved problem.

5. **Advancing alternatives**: If competing carbon-free technologies such as battery storage, geothermal, and next-generation geothermal advance quickly, nuclear's relative appeal could fade.

6-3. Comparing the Two Views

| Issue | Bull | Bear |

| --- | --- | --- |

| Demand | Structural and long-term | Wide forecast variance, may be overstated |

| Cost | Expected to improve via SMRs | Historically prone to overruns |

| Policy | Favorable shift | Vulnerable to political change |

| Valuation | Early stage, upside | Some overheating |

| Time frame | A 10-year-plus megatrend | Near-term realization is slow |

A balanced view is that the long-term direction may be favorable, but the near-term pace of realization and the valuations of individual names deserve caution. The megatrend being correct and buying a particular stock at a particular price being correct are two separate matters.

7. Comparing Investment Approaches — Where Are You Exposed

"Investing in the nuclear renaissance" actually spans very different approaches. Your risk-reward profile changes greatly depending on which layer of the value chain you are exposed to. Below are four representative approaches. (This is not a judgment that one is better, but a way to understand the differences in character.)

| --- | --- | --- | --- |

The key this table shows is that even within the same "nuclear theme," the bet is effectively different depending on the exposure.

- Operators are assets that already generate cash, so they are relatively defensive, but expectations may also be priced in.

- Uranium is closer to a bet on the commodity cycle, requiring attention to price volatility and the self-correcting supply mechanism.

- SMR developers are a high-risk, high-volatility bet of "big if it works, big if it fails."

- Equipment and construction benefit only if new builds actually happen, an approach sensitive to cycles and orders.

Risk-reward spectrum (conceptual)

lower risk/volatility <----------------------> higher risk/volatility

| |

[operators] --- [uranium/fuel] --- [equip/build] --- [SMR developers]

cash flow commodity cycle order cycle unproven growth

Restating the Bull and Bear Scenarios

To compress once more how each approach diverges across bull and bear scenarios:

| --- | --- | --- | --- | --- |

In the end, whichever approach you choose, the first order of business is checking whether it fits your risk tolerance and time horizon.

8. Risks and Checkpoints

Here are items worth checking before any investment decision. Again, this is a checklist, not a recommendation.

8-1. Macro and Policy Checkpoints

- Are major countries' nuclear policy directions (tax, licensing, classification standards) staying favorable?

- Are power-demand forecasts backed by actual data, or based only on expectations?

- How are geopolitical risks in the fuel supply chain, such as uranium, evolving?

8-2. Company and Project Checkpoints

- Does the company generate actual revenue and cash flow, or is it still at the promise stage?

- For SMR firms, are there binding orders or contracts?

- For large reactor projects, are cost and schedule being managed against plan?

- Is the valuation excessive relative to earnings expectations?

8-3. Risk Summary

| Risk type | Description | Impact |

| --- | --- | --- |

| Construction risk | Cost overruns, delays | Eroded profitability |

| Technology risk | SMRs unproven | Possible commercialization failure |

| Regulatory risk | Sudden policy or opinion shifts | Worse operating environment |

| Market risk | Stretched valuations | Sharp downside volatility |

| Supply-chain risk | Reliance on fuel enrichment | Cost and availability swings |

8-4. Additional Checkpoints — One Level Deeper

Here are further checks for seeing beyond surface-level news. The more clearly you can answer these, the easier it is to separate narrative from substance.

- **Binding nature of contracts**: Have you confirmed whether a Big Tech power deal is a binding PPA, or merely a memorandum of understanding or declarative agreement? Certainty of realization differs greatly by form.

- **Power delivery timing**: For SMR-related agreements, have you distinguished when power is actually delivered versus a pre-purchase premised only on future deployment?

- **Financing structure**: Is funding for a large project secured, and how dependent is it on government subsidies and tax credits? How vulnerable is it to policy change?

- **Fuel (HALEU) supply**: If a design requires HALEU, is that supply chain actually secured, or still immature?

- **Dilution risk**: Could a loss-making SMR or uranium firm increase its share count via secondary offerings and dilute existing shareholders?

- **Valuation basis**: Does the current price already reflect an optimistic scenario years out? How large is the downside if expectations falter?

- **Diversification and time horizon**: Are you overly concentrated in a single name or theme? Does your investment horizon match this megatrend's pace of realization?

Narrative -> substance check flow (conceptual)

[attractive news]

[binding? delivery timing? funding? fuel?] <- four core questions

[check valuation, dilution, diversification]

[fits your risk tolerance and horizon?]

9. K-Nuclear — Implications for the Korean Ecosystem

This global trend carries direct implications for the Korean nuclear ecosystem, often called K-nuclear.

9-1. Korea's Strengths — APR1400 and Build Capability

Korea has a large reactor design, the APR1400, and is credited with proving its ability to build on budget and on schedule through the Barakah project in the United Arab Emirates. Barakah, the first large-scale overseas export of a Korean-type reactor, was reported to have built four units relatively on schedule, raising confidence in Korean EPC (engineering, procurement, construction) capability. It contrasts with the Vogtle cost and schedule overruns examined earlier, and "building on budget and on schedule" is cited as a differentiator in the global market.

In addition, when a Korean-led consortium was reported in 2024 to have been selected as the preferred bidder for the Czech Dukovany new-reactor project, K-nuclear's export competitiveness drew renewed attention. On top of that, cooperation discussions in Eastern European markets such as Poland have also been reported, and some argue that the Korean-type reactor is emerging as an option amid the de-Russification and carbon-free transition trends.

9-2. The K-Nuclear Value Chain — Who Does What

On the value chain, a structure in which Korea Hydro and Nuclear Power (KHNP) is the operator and export lead, while Doosan Enerbility makes major components such as the reactor, is frequently cited. In particular, Doosan Enerbility is known to hold the facilities and capability to make large cast-and-forged components such as reactor pressure vessels and steam generators, an area with high entry barriers because only a handful of firms worldwide possess such large-scale casting and forging capacity.

| Player | Role | Talking points |

| --- | --- | --- |

| Korea Hydro and Nuclear Power (KHNP) | Operation and export lead | Barakah record, Czech preferred-bidder reporting |

| Doosan Enerbility | Major components and large castings/forgings | Pressure vessel and steam generator capability |

| Design and engineering | Reactor design and technology | Holds standard designs such as APR1400 |

| Construction partners | EPC execution | Track record on budget and schedule |

Doosan Enerbility has been reported to be seeking to supply major components not only for domestic large reactors and export projects but also for SMRs developed by overseas firms. In other words, the Korean ecosystem is pursuing opportunity along two branches: exporting its own reactor design and participating in the global SMR supply chain.

9-3. K-Nuclear Opportunities and Risks

| Category | Opportunity | Risk |

| --- | --- | --- |

| Exports | Reported entry into new markets such as the Czech Republic | Final contract and financing terms uncertain |

| Build capability | Track record on budget and schedule | Burden of localization and overseas regulation |

| SMR | Possible major-component supply role | Own design still moving toward commercialization |

| Policy | Domestic nuclear policy stance | Direction can shift with changes of government |

9-4. An Overall View

The core of the K-nuclear narrative is that the ability to manage construction cost and schedule reliably could be a differentiator in the global nuclear renaissance. That said, overseas orders involve many variables through final contract signing and financing, and the domestic policy stance has historically been subject to swings with changes of government, both of which should be weighed. Whether an order moves beyond the preferred-bidder stage to a binding contract and financing, and whether local regulatory and financing terms are favorable to the Korean side, is what determines whether it turns into actual results.

10. Closing Thoughts

The nuclear renaissance narrative is rooted less in a passing fad than in structural change: the scarcity of carbon-free baseload and the power demand of the AI era. In that sense, the long-term direction is worth taking seriously.

At the same time, nuclear is an industry with large construction and regulatory risks, and SMRs in particular are not yet sufficiently proven commercially. Some related stocks may have a great deal of expectation already priced in. The megatrend pointing in the right direction and buying a specific stock at a specific price being wise are entirely different matters.

In the end, what matters is the discipline to judge in a balanced way, based on data, the fundamentals of individual companies, and your own risk tolerance, rather than being swept along by the narrative.

> To stress it once more: this article is for information and educational purposes only and is not investment advice or a recommendation. Investment decisions and their consequences are your own responsibility. Consult a qualified professional when needed. The names mentioned in the text are examples only and are not buy or sell recommendations.

References

- [Reuters — Microsoft deal to revive Three Mile Island nuclear plant](https://www.reuters.com/markets/deals/constellation-energy-restart-three-mile-island-power-microsoft-2024-09-20/)

- [CNBC — Three Mile Island nuclear plant restart for Microsoft AI power](https://www.cnbc.com/2024/09/20/three-mile-island-nuclear-plant-will-reopen-to-power-microsoft-ai.html)

- [IEA — Electricity 2024 and data centre demand analysis](https://www.iea.org/reports/electricity-2024)

- [World Nuclear Association — Information Library](https://world-nuclear.org/information-library.aspx)

- [U.S. EIA — Nuclear explained](https://www.eia.gov/energyexplained/nuclear/)

- [Constellation Energy — Newsroom](https://www.constellationenergy.com/newsroom.html)

- [Reuters — NuScale Power and UAMPS project termination](https://www.reuters.com/business/energy/nuscale-uamps-agree-terminate-small-modular-reactor-project-2023-11-08/)

- [Bloomberg — Nuclear power and AI energy demand coverage](https://www.bloomberg.com/energy)

- [Yonhap News (yna.co.kr) — Korean consortium preferred bidder for Czech reactor reporting](https://en.yna.co.kr/)

- [The Korea Economic Daily (hankyung.com) — Doosan Enerbility and K-nuclear coverage](https://www.hankyung.com/)