Nuclear Electric Power Generation (United States)

NAICS 2022 code 221113 — an investor primer

For public-market and private investors alike. Core industry statistics are from U.S. federal sources (Census Bureau, Energy Information Administration, Nuclear Regulatory Commission). Every figure is cited to the numbered Sources list at the end. Reported facts and forward-looking judgments are worded to keep them distinct.

---

1. Overview

Nuclear power is the largest single source of carbon-free electricity in the United States, running around the clock and supplying roughly 18–19% of all U.S. electricity from about 94 reactors at some 54 plants [5][6]. For most of the past decade this was a "managed decline" business — flat demand, cheap natural gas, and a dozen plant closures. In the space of about three years that story has flipped. Electricity demand is rising again, data centers want 24/7 clean power, and no one can build a new reactor quickly, so the existing fleet has become a scarce, strategic asset. The industry itself is narrow: it means operating the reactor, not mining the fuel, building the plant, or making the equipment. There are two ways to get exposure. Public-market investors buy the utilities and power producers that own the reactors — almost all of them large, diversified holding companies where nuclear is one segment. Private and infrastructure investors take direct or joint ownership stakes in plants, finance restarts and uprates against long-term contracts, buy transferable tax credits, or back next-generation reactor developers. This primer explains what the industry is, how big it is, how the money works, and how both kinds of investor get in.

---

2. What it is and how it is structured

Scope. NAICS (North American Industry Classification System) code 221113 covers establishments primarily engaged in operating nuclear-powered electric generation facilities — plants that use fission to boil water, spin steam turbines, and deliver electricity into the grid [1]. The classification follows the plant, not the corporate owner. A utility that owns coal, gas, and nuclear plants has only its nuclear stations counted in 221113 [1]. So the industry is a fleet of physical reactors, while the companies an investor can buy are mostly multi-technology firms.

What it excludes. These adjacent activities are separate NAICS industries — several are themselves investable and are where much of the "nuclear renaissance" venture and equipment money is flowing [1]:

So reactor vendors, uranium miners, enrichers, engineering contractors, and pre-commercial small modular reactor (SMR) developers are adjacent nuclear exposure — not operating exposure to 221113 [1].

Ownership mix. U.S. reactors are often owned in fractions, with a licensed operator holding part of a plant and municipal utilities, cooperatives, or other utilities holding minority interests [7]. Weighting by capacity, roughly four-fifths of U.S. nuclear capacity is owned by shareholder-owned utilities and competitive (merchant) generators, and about one-fifth by federal, state, municipal, public-power, or cooperative entities (both reports' capacity-weighted estimate; there is no single official national tally) [7]. The largest non-shareholder owner is the Tennessee Valley Authority (TVA), a federal corporation operating seven reactors (Browns Ferry, Sequoyah, Watts Bar) and wholesaling to 150-plus municipal and cooperative distributors [7][15]. Other public and cooperative owners include Energy Northwest, Nebraska Public Power District, Santee Cooper, the Municipal Electric Authority of Georgia, Oglethorpe Power, and the municipal owners at Palo Verde and the South Texas Project [7].

Within the shareholder-owned block, owners split into two very different business models — rate-regulated and merchant — which is the single most important distinction for an investor (see Section 5).

---

3. How big it is

Two different measuring sticks matter here, and they do not match — for a good reason.

Federal business statistics (Census). For NAICS 221113 the 2022 Economic Census reported $34.6 billion of revenue across 34 firms [2]. The County Business Patterns program (2023) reported 166 establishments, 38,822 employees, and $5.28 billion of annual payroll [3]. The industry is extraordinarily concentrated: the top 4 firms account for 61.9% of revenue, the top 8 for 78.7%, and the top 20 for 99.8% [2]. (The Herfindahl-Hirschman Index, a standard concentration measure, is suppressed in our source [2].) The Small Business Administration's size standard for the industry is 1,150 employees [4] — every real operator vastly exceeds it, so this is a large-firm industry with essentially no small-business cohort.

Important undercount. The Economic Census largely excludes government-owned utility activity, so its $34.6 billion omits TVA and the municipal, public-power, and cooperative owners that hold roughly a fifth of the fleet [2][7]. It is therefore not total revenue for the whole U.S. nuclear fleet — the true figure is higher. A commonly cited broader estimate is IBISWorld's modeled U.S. nuclear-generation market of about $39–41 billion (2025–2026); treat that as an order-of-magnitude private estimate, not a Census tabulation, and lead with the $34.6 billion Census figure for the counted (private-sector) industry [Opus, IBISWorld].

Physical size (EIA). The fleet is physically enormous even though it is corporately tiny. The Nuclear Regulatory Commission counts 94 operating commercial reactors [6], at about 54 plant sites, with roughly 98 GW (gigawatts) of net summer capacity [5]. It generated about 782 TWh (terawatt-hours) in 2024, near 18–19% of U.S. electricity — the world's largest fleet, ahead of France and China [5][6]. Its capacity factor was about 90% in 2024 (down from ~93% in 2023 on outage timing), the highest utilization rate of any electricity source [5]; refueling outages are the main downtime. Vogtle (Georgia), at four reactors and ~4.5 GW, is the largest plant; Vogtle Units 3 & 4 (2023–2024) were the first newly built U.S. reactors in decades, while about a dozen reactors have closed permanently since 2013 [Opus][5].

The takeaway: a ~$35–41 billion-a-year industry, ~98 GW of capacity, ~782 TWh of output, ~18–19% of U.S. power — concentrated in ~54 plants owned by a few dozen entities.

---

4. The investable universe

Almost no pure public "nuclear play" exists. The listed owners are diversified utilities and power producers where nuclear is one part of a much larger business, so the reported market caps and dividend yields below are company-wide, not nuclear-only. (Clean per-company nuclear revenue is not broken out in public data; nuclear capacity is the more meaningful size marker and is shown instead.)

Publicly traded owners (market data at the July 17, 2026 close; will change) [32]:

Other listed owners with nuclear in the mix include Entergy (ETR), NextEra Energy (NEE), Public Service Enterprise Group (PEG), Xcel Energy (XEL), DTE Energy (DTE), Ameren (AEE), American Electric Power (AEP), Pinnacle West/Arizona Public Service (PNW), Evergy (EVRG), and PG&E (PCG) [7][8]. Note that Exelon (EXC) spun its generation fleet off as Constellation in 2022 and now owns transmission and distribution only [Opus].

Major private and non-shareholder owners (not directly buyable on a stock exchange): TVA (federal, 7 reactors); Energy Northwest and Nebraska Public Power District (public power); Santee Cooper, Oglethorpe Power, and the Municipal Electric Authority of Georgia (cooperative/municipal joint owners of Vogtle and others); and the municipal and public owners at Palo Verde and the South Texas Project [7][15]. Private and infrastructure capital typically reaches the industry through these joint interests, through restart and uprate financing, or through tax-credit transfers rather than through a listed share (see Section 10).

Adjacent tickers that move with the theme but are outside 221113: uranium miners (Cameco, CCJ), enrichers (Centrus, LEU), equipment and services (BWX Technologies, GE Vernova), SMR developers (Oklo, NuScale/SMR), and thematic exchange-traded funds (ETFs) such as NLR, NUKZ, URA, and URNM [Codex][Opus].

---

5. How the money works

Nuclear economics split sharply between the two ownership models. This is the most important thing to internalize.

Rate-regulated model (Duke, Southern, Dominion, Entergy, Xcel, Arizona Public Service). A state public utility commission (PUC) lets the utility recover a revenue requirement = operating costs + depreciation + taxes + (allowed return × rate base) [14]. "Rate base" is the depreciated capital the utility has prudently invested; the allowed return on equity (ROE) — typically about 9.5–10.5% — is applied to the equity portion. Earnings grow by growing the rate base (more approved capital), not by selling more electricity [14]. For nuclear, rate-base additions include license-renewal work, uprates, steam generators, spent-fuel storage, and security upgrades. The catch: a PUC can disallow imprudent or overrun spending. Plant Vogtle is the cautionary tale — total project cost ballooned from a ~$14 billion budget to about $35–37 billion; Georgia Power's net investment in Units 3 & 4 was roughly $10.7 billion, of which regulators let it recover only part from customers, so shareholders absorbed the rest [24][Opus]. For an investor, regulated nuclear is bond-like: low-beta, dividend-anchored earnings whose upside is capex-driven rate-base growth (often guided at ~5–8% a year) and whose risk is disallowance and regulatory lag [14].

Merchant / competitive model (Constellation, Vistra, Talen). These owners sell into wholesale markets run by regional transmission organizations (RTOs) / independent system operators (ISOs) such as PJM (the Mid-Atlantic grid) and ERCOT (Texas), and earn on the spread between market prices and their very low running cost. EIA's 2024 data put nuclear production expense at about $23.08/MWh (operations $9.87 + maintenance $6.84 + fuel $6.37), versus $41.32/MWh for fossil-steam plants [16]. That figure excludes depreciation, financing, and capital projects, but it shows why a well-run, already-built reactor throws off cash when market revenue clears well above it. Merchant revenue comes in layers:

• Energy — wholesale $/MWh; historically volatile (cheap gas nearly bankrupted the merchant fleet in the mid-2010s).
• Capacity payments — the RTO pays generators to be available. PJM's July 2026 auction (2028/29 delivery) cleared at the $325/MW-day administrative cap — about $118,600 per MW-year — and still came up ~6,831 MW short of its reliability target, hard evidence of scarcity [17].
• Long-term contracts / power purchase agreements (PPAs) — increasingly 20-year deals with data-center operators (Section 6).
• Tax credits. The Inflation Reduction Act's Section 45U zero-emission nuclear production tax credit (PTC) pays existing reactors up to $15/MWh (1.5¢/kWh) — a base 0.3¢ times a 5× multiplier for meeting prevailing-wage rules — for electricity sold 2024 through 2032 [18]. It phases out as a plant's gross receipts rise above an inflation-adjusted threshold (2.6¢/kWh in 2025), so it works as a price floor, not a bonus on top of high prices [18][19]. That floor ended the era of premature retirements. New reactors and uprates can instead use the technology-neutral Section 45Y clean-electricity PTC or the Section 48E investment tax credit (ITC) [19].
• State zero-emission credits (ZECs). Illinois, New York, New Jersey, and Connecticut created ZEC subsidies (2016–2019) that kept plants open; New York has extended its program through 2049, while New Jersey approved about $300 million over three years [20]. Several older ZECs are now being replaced by data-center PPAs.

---

6. What drives demand

1. Data centers and artificial intelligence — the defining new tailwind. After ~15 years of flat U.S. electricity demand, load is inflecting: EIA measured demand growth of about 1.7% a year in 2020–2025 versus 0.1% in 2005–2019, and forecasts continued growth (~1.9% in 2026, 2.5% in 2027), heaviest in the PJM and ERCOT regions where many reactors sit [23]. Lawrence Berkeley National Laboratory (LBNL) estimated U.S. data centers used 176 TWh (4.4% of U.S. electricity) in 2023, rising to 325–580 TWh (6.7–12.0%) by 2028 [22]. Nuclear is the preferred match because it is 24/7, carbon-free, weather-independent, and already grid-connected at large single sites. Landmark deals: - Constellation–Microsoft (2024): 20-year PPA to restart the ~835 MW former Three Mile Island Unit 1 (renamed Crane Clean Energy Center), targeted ~2028 [21]. - Constellation–Meta (2025): 20-year deal from June 2027 supporting Clinton's 1,121 MW plus a 30 MW uprate, replacing the plant's expiring state credit [21]. - Talen–Amazon: contract ramping to as much as 1,920 MW from Susquehanna by 2032, running through 2042 [Opus][Codex]. - Vistra–Amazon and Vistra–Meta: up to 1,200 MW from Comanche Peak, plus Meta agreements (announced January 2026) covering 2,609 MW including 433 MW of planned uprates [Codex]. A caution flagged by both reports: a PPA usually allocates price and clean-energy attributes contractually — it does not necessarily mean electrons flow through a private wire from one reactor to one data center, and FERC (the Federal Energy Regulatory Commission) scrutiny of on-site "co-location" can change project economics [Codex].

2. Electrification of transport, buildings, and industry, plus manufacturing reshoring.

3. Decarbonization and clean-firm mandates — nuclear is the largest U.S. carbon-free source, valued in state clean-energy standards and corporate net-zero goals [7].
4. Grid reliability and capacity scarcity — reflected in record-high capacity prices as reserve margins tighten [17].
5. Energy security and bipartisan policy — the 2024 ADVANCE Act (faster licensing), DOE loans, and uranium reshoring [21][28].

---

7. Regulation

Nuclear is the most heavily regulated part of the power sector. Regulation is both the moat (new competitors are nearly impossible to permit quickly) and the main cost and risk center.

• Nuclear Regulatory Commission (NRC) — the federal safety regulator; controls operating licenses, license renewals, uprates, restarts, and decommissioning [6]. Reactors are first licensed for 40 years, renewable in 20-year steps to 60 years, then to 80 years via subsequent license renewal (SLR) if the plant's aging-management case is accepted [28]. Each SLR extends a plant's earning life by 20 years at modest incremental cost, making it one of the highest-return moves available in the industry [28]. The 2024 ADVANCE Act directs the NRC to modernize and speed licensing.
• Federal Energy Regulatory Commission (FERC) — regulates interstate wholesale power sales, transmission, and RTO/ISO market rules (the PJM capacity and energy auctions, co-location questions, merger approvals). FERC generally does not set retail rates or decide whether a state utility may build a plant; ERCOT's Texas market sits largely outside FERC's wholesale-rate jurisdiction [16].
• State public utility commissions (PUCs) — for regulated owners, set the rate base, allowed ROE, cost recovery, and prudence of nuclear spending. A plant can run flawlessly and still deliver poor equity returns if its commission disallows costs [14].
• Environmental Protection Agency (EPA) — governs radiation limits (40 CFR Part 190), Clean Water Act cooling-water intake rules (Section 316(b)), and steam-electric effluent guidelines [29]. Because reactors emit no carbon dioxide, sulfur, or mercury, EPA rules that penalize fossil plants improve nuclear's relative position — but cooling-water availability and temperature can constrain output during heat or drought.
• Department of Energy (DOE) — provides loan guarantees (a $1.52 billion conditional loan backs the Palisades restart) and funds domestic fuel-supply reshoring [21].
• Liability — the federal Price-Anderson framework pools industry accident coverage above $16 billion, with $500 million of primary insurance per site and roughly $158 million per reactor in potential retrospective assessments [30].

---

8. Competitive dynamics and consolidation

• The merchant side is consolidating fast. Exelon spun its generation off as Constellation (2022); Constellation then bought Calpine (announced January 2025, closed January 2026) to become the largest U.S. power generator, with roughly 22 GW of nuclear [9]. Vistra bought Energy Harbor (2024) to add ~4 GW of PJM nuclear [10]. Scale buys centralized outage expertise, purchasing leverage, hedging platforms, and the balance sheet to fund restarts and uprates — but it does not remove correlated risk (a common reactor-design flaw or regional price drop hits many units at once) [10].
• Scarcity is the defining dynamic. No new merchant reactors are being built at scale, the fleet is finite, and demand is rising — so existing reactors have repriced as strategic, hard-to-replicate assets, which is why the merchant equities have re-rated well above traditional utility multiples [17][32].
• Barriers to entry are extreme: NRC licensing, roughly $10 billion or more per GW for new build, decade-plus timelines, scarce trained operators, specialized fuel supply chains, and siting hurdles. This protects incumbents' pricing power.
• Competition for the demand comes from natural gas (cheap, flexible, but carbon-emitting), renewables plus storage (cheap but intermittent), and eventually SMRs — which could be either new competition or a growth avenue for incumbents. More renewables can simultaneously depress daytime energy prices and raise the value of firm, always-on capacity [Codex].

---

9. Risks

1. New-build cost and schedule. Vogtle's ~$35–37 billion (about 2.5× budget, years late) is the warning; overruns can be disallowed by PUCs and land on shareholders [24].
2. Merchant power-price risk. The 45U credit provides a floor, but the upside rides on volatile wholesale and capacity prices; high contract prices can also shrink the 45U benefit through its phase-out [18].
3. Operational and outage risk. A reactor earns heavily when running and bleeds fixed cost during an outage; an NRC-ordered shutdown, a safety finding, or an aging-component failure removes a large earnings block. Aging plants (many 40–50 years old) carry equipment risk even after SLR [28].
4. Catastrophic tail risk. A serious incident anywhere can shift policy quickly (as Fukushima did) and destroy asset value beyond insurance [30].
5. Spent fuel and waste. The U.S. has more than 90,000 metric tons of spent fuel stranded on-site with no permanent repository; the federal government's disposal-failure liability has climbed toward the tens of billions of dollars, and plant-level storage, security, and litigation exposure remains [25][Codex].
6. Fuel-supply concentration. EIA reported about 99% of the uranium concentrate used by U.S. reactors in 2023 was foreign-origin; the 2024 ban on Russian uranium imports (waivers to 2028) is a near-term supply-security risk until domestic conversion and enrichment scale up [26].
7. Decommissioning cost — multi-hundred-million-to-billion-dollar per-unit liabilities; owners hold trust funds, but shortfalls are a risk [Codex].
8. Data-center demand risk. Forecasts could overshoot if AI computing gets more efficient or projects fail to secure financing and grid connections; underwrite signed contracts separately from speculative interconnection queues [Codex].
9. Policy-expiration and valuation risk. The 45U credit covers power sold only through 2032; a slowdown in data-center capex or a failure to replace 45U could compress the merchant equities' rich multiples [18][32].

---

10. How to invest and the outlook

Two ways in.

Public-market investors. The choice is essentially income versus growth.

• Merchant / growth: Constellation (CEG), Vistra (VST), and Talen (TLN) offer the most leveraged exposure to rising demand, capacity scarcity, and data-center PPAs — with commensurately higher power-price and valuation risk and low or no dividends [32]. At the July 17, 2026 close these traded at roughly 21× (CEG), 26× trailing / 17× forward (VST), and ~15× forward (TLN) earnings, with dividend yields under 0.7% (Talen pays none) [32].
• Regulated / income: Duke (DUK), Southern (SO), and Dominion (D) offer rate-base-driven earnings and dividend yields of about 3.2–3.8%, trading at classic utility multiples of roughly 19–24× earnings; nuclear is one segment of a regulated network [32].
• Basket / thematic: ETFs (NLR, NUKZ) and adjacent fuel-cycle and equipment names (CCJ, LEU, BWXT, GE Vernova) — outside 221113 but correlated [Codex].

Private and infrastructure investors. Direct ownership of an operating reactor is rare (most sit inside public companies, regulated utilities, or public-power entities) and any transfer requires NRC approval and proof of technical, financial, insurance, and decommissioning capability [6]. The realistic entry points are:

• Direct or minority (joint) plant ownership — highest operating exposure, and full responsibility for outages, capex, fuel, and decommissioning.
• PPA-backed restart or uprate financing — a 20-year offtake can support debt and equity for incremental capacity (the template used at Palisades, backed by a $1.52 billion DOE loan plus a long-term PPA) [21].
• Project or corporate debt against contracted cash flows.
• Tax-credit transfer financing — buying transferable 45U/45Y credits, subject to eligibility and recapture diligence [18].
• Public-private partnerships — DOE loans, state support, and municipal ownership paired with private operating or development capital.
• Advanced-reactor / SMR development — the highest-risk, highest-return avenue; pre-revenue and outside 221113 until a project reaches commercial generation. Private underwriting should focus on the investor's net beneficial megawatts, the NRC license expiration, capacity factor and outage record, PPA price and counterparty, fuel and enrichment coverage, decommissioning-trust sufficiency, and 45U phase-out — and on exit liquidity, which is far lower than for gas or renewable assets [Codex].

Outlook — the most positive nuclear environment in a generation, with real execution risk. The base case (an analytical judgment, not a fact): continued operation of most economic reactors, more 80-year license renewals, selective uprates, restarts where a long-term buyer absorbs the risk, more hyperscaler PPAs — and only gradual net capacity growth, because Vogtle showed how painful large new build remains [24][28]. The bull case adds data-center demand near the top of LBNL's range, sustained capacity scarcity, and commercial SMRs; the bear case is cheaper gas, a data-center slowdown, an outage or safety event, PUC disallowances, or no replacement for 45U after 2032 [22][23]. The industry's defining feature — a finite, hard-to-replicate fleet meeting suddenly rising demand — is bullish for the value of existing assets and is the central reason nuclear has moved from managed decline to strategic growth. For income, the regulated owners; for growth, the merchants; for private capital, contracted restarts, uprates, and SMR project equity, usually alongside federal credit support.

---

Sources

1. U.S. Census Bureau — 2022 NAICS Definitions and Manual, "221113 Nuclear Electric Power Generation" (2022). https://www.census.gov/naics/
2. U.S. Census Bureau — 2022 Economic Census, Utilities basic statistics / concentration for NAICS 221113 (receipts $34.6B; 34 firms; CR4 61.9%, CR8 78.7%, CR20 99.8%; HHI suppressed) (2024). https://data.census.gov/
3. U.S. Census Bureau — County Business Patterns 2023, NAICS 221113 (166 establishments; 38,822 employees; $5.28B annual payroll). https://www.census.gov/programs-surveys/cbp.html
4. U.S. Small Business Administration — Table of Size Standards, NAICS 221113 (1,150 employees), effective 2023. https://www.sba.gov/document/support-table-size-standards
5. U.S. Energy Information Administration — Electric Power Annual 2024 and Today in Energy (world's largest fleet; ~98 GW net summer capacity; ~782 TWh generation 2024; ~18–19% of U.S. electricity; ~90% capacity factor) (2025). https://www.eia.gov/electricity/annual/; https://www.eia.gov/todayinenergy/detail.php?id=65104
6. U.S. Nuclear Regulatory Commission — Operating Reactors (94 operating commercial reactors) (2026). https://www.nrc.gov/reactors/operating
7. U.S. Energy Information Administration — U.S. Nuclear Plant Ownership and Reactor Capacity (fractional ownership; public/cooperative owners) (2025). https://www.eia.gov/nuclear/reactors/ownership.php
8. Company disclosures (SEC filings) for listed nuclear owners (Entergy, NextEra, PSEG, Xcel, DTE, Ameren, AEP, Pinnacle West, Evergy, PG&E).
9. Constellation Energy — Form 10-K and Calpine acquisition (~22 GW nuclear; 14 stations/25 units; deal closed Jan 2026). https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=CEG
10. Vistra Corp. — Energy Harbor acquisition (2024) and Form 10-K (~6.4 GW nuclear). https://investor.vistracorp.com/
11. Duke Energy — Form 10-K (11 reactors; ~10.8 GW regulated). https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=DUK
12. Dominion Energy — Nuclear Facilities and Form 10-K (7 reactors; Millstone merchant). https://www.dominionenergy.com/
13. Southern Company / Southern Nuclear — Annual Report (~8.2 GW; Vogtle, Hatch, Farley). https://www.southernnuclear.com/
14. National Association of Regulatory Utility Commissioners — Ratemaking Fundamentals (revenue requirement; rate base; allowed ROE). https://www.naruc.org/
15. U.S. Energy Information Administration — Today in Energy, "TVA is the largest government-owned electricity provider in the United States." https://www.eia.gov/todayinenergy/detail.php?id=49136
16. U.S. Energy Information Administration — Electric Power Annual, Table 8.4 average operating expenses (nuclear $23.08/MWh; fossil-steam $41.32/MWh, 2024). https://www.eia.gov/electricity/annual/
17. PJM Interconnection — Reliability Pricing Model / capacity-auction results (July 2026 auction: 138,318 MW procured, $325/MW-day cap, ~6,831 MW short). https://www.pjm.com/markets-and-operations/rpm.aspx
18. Internal Revenue Service — Zero-Emission Nuclear Power Production Credit (Section 45U; up to $15/MWh; 2024–2032; gross-receipts phase-out). https://www.irs.gov/credits-deductions/zero-emission-nuclear-power-production-credit
19. Internal Revenue Service — Clean Electricity Production/Investment Credits (Sections 45Y, 48E; incremental nuclear). https://www.irs.gov/credits-deductions/clean-electricity-production-credit
20. New York Department of Public Service and New Jersey Board of Public Utilities — state zero-emission-credit (ZEC) programs. https://dps.ny.gov/; https://www.nj.gov/bpu/
21. U.S. Department of Energy — Palisades restart ($1.52B loan) and Constellation announcements (Crane/TMI restart; Meta–Clinton). https://www.energy.gov/edf/palisades
22. Lawrence Berkeley National Laboratory / DOE — 2024 U.S. Data Center Energy Usage Report (176 TWh / 4.4% in 2023; 325–580 TWh / 6.7–12.0% by 2028). https://energyanalysis.lbl.gov/publications/2024-lbnl-data-center-energy-usage-report
23. U.S. Energy Information Administration — Electricity Demand Growth and AEO 2026 (1.7%/yr 2020–2025 vs 0.1% 2005–2019; ~1.9% 2026, 2.5% 2027). https://www.eia.gov/todayinenergy/; https://www.eia.gov/outlooks/aeo/
24. Southern Company / Georgia Power and Georgia Public Service Commission — Vogtle 3 & 4 cost (~$35–37B total; ~$10.7B Georgia Power net investment; partial disallowance). https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=SO
25. U.S. Government Accountability Office / Nuclear Energy Institute — spent-fuel inventory (>90,000 metric tons) and federal disposal-liability estimates. https://www.gao.gov/
26. U.S. Energy Information Administration — U.S. uranium imports and Russian-supply ban (~99% foreign-origin uranium 2023; import ban effective 2024, waivers to 2028). https://www.eia.gov/todayinenergy/detail.php?id=64444
27. Congress.gov — ADVANCE Act (2024) and Prohibiting Russian Uranium Imports Act (H.R.1042, 2024). https://www.congress.gov/
28. U.S. Nuclear Regulatory Commission — License Renewal and Subsequent License Renewal (40-year initial license; +20 to 60; +20 to 80). https://www.nrc.gov/reactors/operating/licensing/renewal/
29. U.S. Environmental Protection Agency — 40 CFR Part 190 radiation limits; Clean Water Act Section 316(b) cooling-water rule; steam-electric effluent guidelines. https://www.epa.gov/radiation/
30. U.S. Nuclear Regulatory Commission — Nuclear Insurance / Price-Anderson Act (>$16B coverage; $500M primary per site; ~$158M per reactor). https://www.nrc.gov/reading-rm/doc-collections/fact-sheets/nuclear-insurance
31. IBISWorld — Nuclear Power Generation in the US (NAICS 221113) modeled market size ~$39–41B (2025–2026); broader/other estimate. https://www.ibisworld.com/united-states/industry/nuclear-power/1911/
32. StockAnalysis.com — market and valuation data for CEG, VST, TLN, DUK, SO, D (July 17, 2026 close). https://stockanalysis.com/