Public Service Enterprise Group (NYSE: PEG) — The Nuclear Option Got Cheaper

Institutional equity-research update | Information through September 2, 2026 | Market data through the September 1 close

⚡ Claude’s Take

The author’s subjective opinion; general information, not investment advice. The analytical body below carries no recommendation.

Rating: ACCUMULATE — upgraded from HOLD. Conviction: Medium. At $73.48, PEG has fallen 10.0% from the $81.62 reference price in our July 3 report even though management maintained 2026 operating-EPS guidance of $4.28–$4.40 and its 6%–8% growth outlook through 2030. The stock now trades at 16.9 times the guidance midpoint and yields 3.65% on the $2.68 annualized dividend. Its composite own-history valuation has fallen from the upper quartile to the 47th percentile. That is the thesis change: operating evidence did not improve enough to make PEG a more valuable company, but the price now requires much less faith in an unsigned nuclear contract.

Framing: quality compounder in a negative trend, not a momentum trade. PEG sits below its 21-, 50- and 200-day averages and has negative three-, six- and twelve-month returns. Yet utility-sector exposure explains most modeled variance, recent factor regimes are adverse but not extreme, and the Q2 earnings bridge did not break. Staging matters more than trying to call the exact low.

We would build positions from $70 to $76, view $77 to $89 as a broadly fair zone, and stop adding above $90 without a signed, creditworthy, long-duration contract for existing nuclear output or a demonstrably more constructive New Jersey rate settlement. The low end corresponds to roughly 16 times current-year operating EPS and compensates for the approaching rate case; the upper end starts to capitalize most of the 6%–8% plan and some nuclear optionality. This is an accumulation call, not a claim that regulatory or nuclear-contract risk has cleared.

The business is unusually attractive for a utility: about 86% of 2025 operating earnings came from PSE&G, New Jersey’s largest regulated electric-and-gas delivery franchise, while the retained 3,758 MW nuclear fleet adds scarcity upside. PSE&G’s recovered margin is still outrunning incremental depreciation, interest and operating costs on a per-share basis. Nuclear operations are strong, PJM remains short capacity, and the federal 45U credit supplies a downside floor when power prices are weak. Yet the two July triggers remain unresolved. PEG still has no long-dated existing-fleet contract comparable with deals signed by Talen, Vistra or Constellation, and New Jersey has not issued a hostile base-rate order. Instead, the risk has moved closer: a year-end 2026 rate filing will occur while the state explicitly reviews ROE, capital structure and recovery mechanisms.

What would make us more positive: a signed existing-nuclear agreement with firm volume, credit support, duration and economics; a PJM backstop award structured as contracted infrastructure rather than merchant development; or a rate settlement near the current 9.60% ROE and 55% equity layer with no material disallowance. What breaks the call: a material ROE/equity-ratio cut or disallowance that turns PSE&G’s incremental EPS bridge negative; equity issuance to fund the present plan; sustained nuclear underperformance; or guidance falling below the 6%–8% framework before any optional project contributes.

Trigger line: A signed, modelable existing-nuclear contract would flip the call more bullish; a rate outcome that turns the regulated incremental EPS bridge negative would flip it bearish.

📈 Stock Price Action — Five-Year Event Map

PEG’s adjusted close rose 34.2% from September 2, 2021 through September 1, 2026, but that endpoint conceals a complete utility de-rating, an AI-and-nuclear re-rating, and a second de-rating. The trailing-year adjusted-close range is $72.61–$85.51; the current quote is 14.1% below that high, 17.8% below the November 2024 five-year high and only 1.2% above the trailing-year closing low. The event attributions below distinguish measured moves from our interpretation; no single company announcement explains the latest decline.

Event 1 coincided with the rate shock and PEG’s completed fossil sale; event 2 was a gradual reassessment of the simpler asset mix and federal nuclear floor rather than one release. Event 3 followed PSEG’s Q1 2024 framing of 45U, data-center discussions, uprates and license life. Event 4 coincided with the record 2024 PJM capacity outcome. Event 5 includes the January 27 DeepSeek selloff in power equities, documented by Reuters. Event 6 followed PSEG’s larger inquiry disclosure and PJM’s capped 2025 auction. Event 7 lacks a single shock: the Q2 2026 release maintained guidance, while contracts remained absent and New Jersey proceedings advanced. The moves are facts; those causal assignments are interpretations.

At September 1, PEG was below its 21-, 50- and 200-day exponential moving averages of $74.93, $76.69 and $78.79. Raw adjusted-close returns were -5.4% over roughly three months, -11.0% over six months and -7.8% over twelve months. The stock fell only 0.5% on the August 4 results day, supporting the conclusion that this was a regime and expectations unwind rather than an earnings-miss event. Market-history data are from AZI Trading; the 2024 capacity context is from PJM.

Verdict: The tape is weak but not informationally catastrophic. Negative momentum makes the timing uncertain; the absence of a company-specific earnings break argues against reading the decline as fundamental impairment.

Changes Since July 3, 2026

Our prior report’s full-price assessment reflected a stock that capitalized a high-quality utility plus much of an uncontracted nuclear option. Five developments change the balance, although none fires the original bull or bear trigger.

First, valuation reset. The reference price fell from $81.62 to $73.48 while the 2026 range and long-term growth outlook stayed intact. At the new price, the forward operating P/E is 16.9 times rather than 18.8 times, the dividend yield is 3.65%, and the own-history composite is near the median rather than the upper quartile. The operating evidence has not changed as much as the embedded expectations.

Second, Q2 validated the earnings engine. Consolidated operating EPS rose to $0.86 from $0.77; first-half operating EPS rose 9.5% to $2.41. PSE&G earned $342 million in Q2 and $919 million in the first half, while Power & Other earned $83 million and $284 million. Distribution and transmission margin added about $0.13 per share in the first half; operating costs, depreciation and interest absorbed about $0.06. Power gross margin and cost control supplied the rest. The Q2 release maintained both the annual range and the 2030 framework.

Third, New Jersey made the regulatory bear case more explicit without crystallizing it. The state’s Phase 1 utility-model report recommends near-term scrutiny of capital spending, automatic recovery, ROE and capital structure, but it makes no final rate recommendation and recognizes the financing risk of poorly calibrated cuts. PSE&G expects to file a base-rate case by year-end rather than waiting until 2029. Separately, mandatory RTO membership could remove a 50-basis-point FERC adder from 2027, a roughly $40 million annual net-income headwind already contemplated in management’s growth framework. The NJBPU Phase 1 study also shows why policy cannot solve affordability by cutting PSE&G alone: distribution represents only about one-quarter of a typical residential bill.

Fourth, the state expanded nuclear-development optionality without contracting PEG’s existing fleet. The Power NJ Act launches a competitive process for at least 1,100 MW of new nuclear, but it requires federal financing and a net-ratepayer-benefit finding; final negotiations can extend into 2028, and management estimates roughly twelve years to build new nuclear. It is a supportive policy signal, not current earnings or an award. The Governor’s signing release is explicit that ratepayers bear no construction-period payments.

Fifth, the contracting gap widened. Management said it had submitted several PJM Reliability Backstop Procurement concepts, but no award exists and bilateral interest in existing output showed “no inflection.” Talen, Vistra and Constellation have already signed material long-duration nuclear arrangements. PEG still owns the option, but peers have converted more of theirs into evidence.

Verdict: The company changed less than the stock. PEG is cheaper, but investors are being paid to wait through a more immediate regulatory process—not being handed a resolved thesis.

1. Executive Summary

PSEG is best understood as two assets sharing a balance sheet. PSE&G is a government-protected, high-density New Jersey transmission-and-distribution franchise. PSEG Power & Other is principally a low-variable-cost, carbon-free nuclear fleet exposed to PJM energy and capacity economics, hedging, federal support and possible contracting. In 2025 PSE&G generated $1.745 billion, or about 86%, of $2.029 billion in consolidated operating earnings. That mix makes PEG a regulated utility first and a nuclear-scarcity security second. The segment and asset data come from PSEG’s 2025 Form 10-K.

The regulated case is straightforward but capital intensive. PSE&G ended 2025 with approximately $36 billion of rate base and plans $22.5–$25.5 billion of regulated investment in 2026–2030. Management expects 6%–7.5% annual rate-base growth and funds the broader $24–$28 billion capital program without planned common-equity issuance or asset sales. The latest full distribution settlement authorizes a 9.60% ROE, 55% equity and 7.07% after-tax weighted cost of capital. Transmission earns through FERC formula rates and annual true-ups, although the RTO adder is at risk. The financial test is not whether rate base grows; it is whether recovered margin exceeds depreciation, interest, operating cost and dilution. It did in 2025 and the first half of 2026, but by much less than the headline capital growth.

The nuclear case is more asymmetric. PSEG owns 3,758 MW across Hope Creek, Salem and half-owned interests at Peach Bottom. The fleet produced 15.8 TWh in the first half at a 93.7% capacity factor and $6.85/MWh fuel cost; more than 95% of expected 2026 output was hedged. Approximately 3,600 MW cleared the 2028/29 PJM auction at its $325/MW-day cap despite a 6.8 GW system shortfall. The federal 45U production tax credit protects downside through 2032 when gross receipts are low, but it phases out as prices rise. This creates a useful floor-plus-upside shape, not a free subsidy: PSEG recorded about $350 million of estimated benefit in 2024 and none in 2025 because receipts exceeded the threshold.

The most important debate is whether regulated compounding and nuclear scarcity deserve a premium after accounting for political affordability constraints. The Phase 1 New Jersey review confirms that capacity, not distribution, caused most of the 2021–2026 residential bill increase. Nevertheless, PSE&G is the visible regulated counterparty, and regulators can ration its capital through program scope, recovery timing, capital structure and allowed return. The current 9.60% ROE is close to the national distribution median of 9.50%, limiting the analytical case for a severe cut; politics does not require analytical consistency.

Our base case assumes no existing-fleet nuclear contract, no return from new nuclear or backstop projects, 6%–7% operating-EPS growth, stable share count, and gradual dividend growth. The bear case is not demand destruction: it is a negative regulatory-and-financing loop in which returns fall, recovery lags, debt costs rise, and equity becomes necessary. The bull case is a contract that monetizes scarcity without shifting development or commodity risk back onto PEG.

Verdict: PEG combines a defensible utility franchise with a real but uncontracted nuclear option. The regulated core is sufficient to underwrite the present earnings plan; optionality should improve outcomes, not be required to achieve them.

2. Business Overview and Economic Architecture

PSE&G: delivery infrastructure, not commodity retail

PSE&G serves approximately 2.4 million electric and 1.9 million gas customers across a 2,600-square-mile territory containing much of New Jersey’s population and commercial density. It owns the local wires, substations, gas mains and related delivery infrastructure. Customers may procure electricity or gas supply through state mechanisms or third parties, but PSE&G earns chiefly from delivery assets and authorized returns. Basic Generation Service and Basic Gas Supply Service commodity costs pass through without utility margin. The Conservation Incentive Program and successor mechanisms reduce sensitivity to weather and usage volumes. These mechanics are described in the 2025 Form 10-K.

That distinction matters whenever customer bills rise. A high bill can create political pressure without creating PSE&G profit. The NJBPU study estimates distribution at roughly 25% of a typical residential bill and attributes about 70% of the 2021–2026 increase to capacity, versus roughly 25% to distribution. For investors, the mismatch is a risk: PSE&G can be blamed for costs it largely passes through, while rate relief may still target the portion it controls.

The regulatory architecture contains several layers. New Jersey distribution is set through base cases and program riders. The October 2024 base-rate order allowed annual increases of $341 million electric and $164 million gas on $9.3 billion and $8.5 billion of respective rate base, materially below PSE&G’s updated requests but on constructive financing parameters. Electric transmission uses FERC formula rates with annual forecasts and true-ups. Efficiency, gas modernization and other programs can receive accelerated recovery but are negotiated separately. This reduces lag yet increases the number of points at which regulators can narrow scope or alter incentives.

Power & Other: nuclear economics plus smaller activities

PSEG Power & Other includes Hope Creek, PSEG’s 57% interests in Salem Units 1 and 2, 50% interests in Peach Bottom Units 2 and 3, gas supply, PSEG Long Island service fees, competitively awarded transmission and legacy obligations. Nuclear capacity is 1,172 MW at Hope Creek, 1,311 MW of owned share at Salem and 1,275 MW at Peach Bottom.

The fleet sells energy, capacity and ancillary services into PJM and normally hedges output up to three years forward. It is a price taker with low variable costs, not a retail franchise. Capacity payments reward accredited availability; energy margin reflects realized power prices, hedges and fuel. The 45U credit supplies a federal floor but phases down as qualifying receipts rise. Zero Emission Certificates for New Jersey units ended in May 2025 as the federal mechanism became available. Consequently, reported quarterly revenue and GAAP income can move in counterintuitive ways even when operating economics improve.

PSEG Long Island is an asset-light management-services arrangement rather than a regulated ownership franchise. Its declining fees and contract life make it a secondary valuation factor. Legacy environmental, decommissioning, pension and corporate items remain important for cash and GAAP presentation but are not the earnings-growth engine.

Earnings mix and segment quality

In 2025 PSE&G supplied about 86% of operating earnings. In the first half of 2026, it supplied $919 million of the consolidated $1.203 billion, or 76%, because Power benefited from stronger margins and cost control. That quarterly mix should not be extrapolated mechanically: nuclear outages, hedge settlements, capacity delivery years and tax effects make the Power contribution more volatile.

The earnings mix is strategically useful. The regulated segment lowers the cost of capital and provides a visible reinvestment runway; nuclear gives PEG exposure to a supply-constrained market without requiring uncontracted new-build spending in the base plan. The danger is cross-subsidization. Management’s stated hurdle—utility-like or contracted risk for incremental generation—is sensible. Any new nuclear or backstop project should be evaluated on which entity bears construction, fuel, operating, market and regulatory risk, not on a headline megawatt number.

Verdict: PEG is not a diversified merchant generator disguised as a utility. It is a regulated utility with a significant nuclear sidecar, and the two pieces require different return, risk and accounting lenses.

3. Industry Structure and Capital Cycle

The regulated utility capital cycle

A local distribution grid is a natural monopoly. Duplicating PSE&G’s wires and pipes would destroy industry economics, so government grants an exclusive service territory and regulates price, service and investment. Traditional competitive analysis therefore fails: there is no ordinary market-share contest, but there is continual bargaining over allowed return, capital structure, recovery speed and customer outcomes.

New Jersey needs investment. Electrification, aging infrastructure, storm resilience, cleaner gas networks, distributed resources and data-center demand require grid capacity. PSE&G’s planned regulated capex is therefore not speculative in aggregate. The capital-cycle constraint appears through regulation rather than new entry. When utility capital spending rises faster than customer affordability or perceived execution, the NJBPU can narrow programs, postpone recovery, require performance metrics or reduce the equity component. GSMP III illustrates this discipline: the Board approved up to $1.41 billion for roughly 600 miles, including $1.05 billion with accelerated recovery, but the result was materially narrower than the initial proposal and deferred some recovery to a later base case.

The July 2026 energy-efficiency extension provides another signal. The Board replaced electric CIP recovery with a lost-revenue adjustment mechanism, lowered the earnings-test threshold, and made performance returns vary from debt cost below 60% of the quality-performance index to full ROE only above 150%. Those terms are program-specific, not an across-the-board base-ROE reset. Still, they show how affordability policy can convert an apparently open-ended capex runway into conditional capital.

The PJM generation capital cycle

PJM faces the inverse problem: too little dispatchable supply relative to projected demand. The 2028/29 capacity auction cleared around its $325/MW-day cap and procured roughly 6.8 GW less than the reliability requirement. PJM’s proposed Reliability Backstop Procurement contemplates contracts as long as fifteen years and maximum willingness to pay near $555/MW-day for qualifying new resources. The cap therefore obscures, rather than disproves, scarcity.

In a textbook commodity cycle, high prices invite new supply and eventually erase excess returns. That response is slow in PJM. Interconnection, permitting, equipment, gas infrastructure, financing and construction constrain new plants; nuclear lead times are longer still. Price signals can therefore persist. Political mechanisms simultaneously truncate upside: auction collars, cost allocation, state procurements and affordability protections decide how scarcity rent is shared.

PSEG’s existing fleet benefits from this lag because it is already licensed, interconnected and operating. Its planned Salem uprates add about 112 MW of PSEG share with lower execution risk than greenfield generation. Yet its ability to capture scarcity is bounded by hedging and policy. More than 95% of expected 2026 output was hedged at midyear, so current spot strength arrives with delay. The 45U credit disappears at high receipts. Capacity collars suppress observed price. A bilateral contract could trade some merchant upside for duration and certainty; until signed, the option remains exposed to political bargaining.

Data centers: use the completion-adjusted number

Load demand is real but commonly overstated. PSEG previously cited an inquiry pipeline near 11.5–11.8 GW. PJM’s January 2026 completion-adjusted forecast begins with 39 data centers and only 394 MW of actual 2025 peak, then forecasts 785 MW in 2026, 2,910 MW in 2030 and 3,084 MW in 2031. Existing expansions and formal new-business requests receive full weight; feasibility-stage requests receive 50%; another 5.6 GW of capacity review and leads is excluded. The two numbers are not contradictory—they measure different funnels—but only the probability-weighted forecast belongs in a base case.

New Jersey’s Data Center Fair Share law improves economics for ordinary customers by requiring large facilities to bear their infrastructure and grid-upgrade costs in a separate class and accept curtailment before residences. It also adds conditions that can slow conversion. For PSE&G, the best evidence will be executed service agreements, deposits, credit support, substation construction and measured coincident load. Inquiry totals are marketing indicators, not rate-base authorization.

Industry profit pools

The regulated profit pool comes from equity invested in authorized rate base. It is predictable but deliberately capped. The nuclear profit pool comes from low operating cost relative to energy, capacity and contract prices; it is volatile but can earn above a regulated return when supply is scarce. Development profit is a third pool, but it exists only if contracts place construction and market risks with parties able and willing to bear them. PEG’s investment case is strongest when it refuses to chase the third pool on merchant terms.

Verdict: Both of PEG’s industries are supply constrained, but policy determines rent capture. PSE&G faces regulatory rationing of abundant proposed capital; nuclear benefits from scarce existing capital but faces political limits on scarcity pricing.

4. Competitive Position and Moat

PSE&G under the Greenwald framework

PSE&G has a durable advantage, but it is not brand strength. Its moat is government protection reinforced by local network scale. The exclusive franchise prevents an entrant from cherry-picking attractive customers, and the embedded grid makes physical duplication uneconomic. Customer retention is compulsory in delivery, not the result of loyalty. Market share is therefore structurally stable.

This advantage does not create unconstrained excess return. The regulator stands in for competition: if PSE&G appears to earn too much, the NJBPU can share gains through lower rates, lower equity, tighter recovery, disallowances or performance requirements. The correct proof of moat is not a 20% ROIC. It is the ability to deploy large amounts of capital at an authorized return, recover prudent costs and avoid equity dilution through repeated cycles.

Recent evidence is constructive. PSE&G net income increased $198 million in 2025 while depreciation rose $91 million and interest $62 million. In the first half of 2026, distribution and transmission margin added roughly $0.13 per share; operating expense, depreciation and interest removed about $0.06, leaving a positive regulated contribution. Diluted shares declined slightly. That is capital-cycle value creation, although the net benefit is much smaller than gross rate-base growth.

Service quality supports regulatory legitimacy. PSE&G has ranked highly in eastern large-utility customer satisfaction, and its July storm response restored nearly all of approximately 380,000 affected customers within 24 hours. Such outcomes do not create pricing power, but they make it harder to argue that investment delivers no customer value. The 2025 results release summarizes the rate-base and service record.

Nuclear under the Greenwald framework

PSEG Nuclear’s advantage is supply-side. Licensed sites, grid connections, specialized labor, fuel arrangements, operational history and extraordinarily long development timelines restrict new entry. Low fuel cost and high capacity factor strengthen its position in the dispatch stack. Those barriers are real, but they create no exclusive customer relationship. Hyperscalers, utilities and load-serving entities can negotiate with other generators, self-build, curtail or wait. PEG competes with other nuclear owners for the most valuable contracts.

That distinction is visible in evidence. Talen’s AWS arrangement can cover up to 1,920 MW through 2042; Vistra’s Meta contracts cover 2,609 MW including uprates; Constellation’s Microsoft contract supports an 835 MW restart. The terms are disclosed in Talen’s, Vistra’s and Constellation’s 2025 Forms 10-K. PSEG has no comparable existing-fleet contract as of this report. Its New Jersey location and operational fleet are scarce, but scarcity alone does not prove bargaining victory.

Operational performance is currently a strength. The fleet generated 30.9 TWh at a 91.2% capacity factor in 2025 and 15.8 TWh at 93.7% in the first half of 2026. Salem Unit 2 completed a second breaker-to-breaker run, and the fleet cleared about 3,600 MW in the 2028/29 auction. Subsequent-license-renewal applications for Salem and Hope Creek are planned for April–June 2027; approval would support operation to 2056, 2060 and 2066. Peach Bottom already has licenses into 2053/54. The NRC schedule verifies the planned application window, not approval.

Management and organizational advantage

PSEG has executed a difficult strategic simplification: exiting fossil generation, preserving nuclear, expanding the regulated base and holding shares roughly flat. Capital allocation has been more disciplined than the old conglomerate structure. Management’s refusal to put unsigned nuclear contracts into the 6%–8% outlook is good forecasting hygiene.

Incentives are less perfect. The 2025 annual program weighted operating EPS at 65%; long-term awards emphasize relative total shareholder return and operating-EPS growth. There is no explicit ROIC, free-cash-flow or affordability metric. In a debt-funded utility expansion, that can reward the size and timing of EPS more directly than capital efficiency. The board should add an efficiency or customer-affordability measure that cannot be met merely by investing more.

Verdict: PSE&G has a very durable regulated moat; nuclear has durable supply barriers but must still win contracts. Management has improved the asset mix, while compensation remains more growth-oriented than capital-efficiency-oriented.

5. Growth History and Forward Opportunities

The base plan

Management guides to 2026 operating EPS of $4.28–$4.40 and 6%–8% compound growth through 2030 from the 2026 midpoint. At 6%, 7% and 8%, that arithmetic implies approximately $5.48, $5.69 and $5.90 of 2030 operating EPS. The plan is anchored in $22.5–$25.5 billion of regulated capex, 6%–7.5% rate-base growth, cost control and expected nuclear output at market prices above the 45U threshold. It excludes a long-duration existing-output contract, new nuclear and incremental regulated projects.

The plan’s credibility has improved through execution. Management rebased the outlook higher in consecutive years, 2025 operating earnings reached $4.04 per share, first-half 2026 operating EPS grew roughly 9.5%, and the full-year range was maintained despite the $40 million prospective RTO-adder issue. The H1 bridge was balanced across both segments rather than dependent on one accounting item.

Still, rate-base growth is not EPS growth. Depreciation, interest and operating costs rise as assets enter service; tax-flowback benefits can fade; regulatory lag may widen. The current bridge suggests that about half of gross distribution/transmission margin growth can be absorbed before reaching EPS in a six-month period. The base case should therefore use the company’s 6%–8% earnings framework, not simply apply the 7.5% upper rate-base rate to consolidated earnings.

Regulated upside

Potential regulated upside includes resilience, advanced conductors, dynamic line ratings, electrification, gas-system modernization, energy efficiency, storage interconnection and data-center infrastructure. PSE&G’s density makes many projects economical, and formula transmission rates reduce lag. The July heatwave produced a 10,446 MW system peak, the highest in fourteen years, showing that existing demand can stress the grid before speculative loads arrive.

Data centers could enlarge the investment runway, but the 2.91 GW completion-adjusted 2030 forecast is the proper anchor. The separate-rate-class law improves cost causation and could make regulators more willing to approve dedicated infrastructure. It can also require projects to self-supply, accept curtailment or bear higher deposits. The best form of upside is customer-funded infrastructure on binding commitments; the worst is anticipatory capex against cancelable inquiries.

The PSE&G base case expected by year-end 2026 is a catalyst because it will define how much historic and near-term capital enters rate base, the requested return and equity ratio, and bill impact. An order near 9.60%/55% with limited disallowance would validate the compounding runway. A material reduction would not merely cut one year; it would lower the return on every incremental dollar and potentially raise financing cost.

Existing-nuclear upside

Existing nuclear offers four growth channels. First, higher capacity prices flow into future delivery years. Second, energy hedges roll into prevailing prices. Third, the Salem uprates add roughly 112 MW of PSEG share. Fourth, a bilateral contract could exchange merchant exposure for a durable premium or floor. The first three are partly represented in management’s plan; a material long-term contract is not.

Contract economics matter more than headlines. Investors should require volume, node, duration, escalation, capacity treatment, outage provisions, collateral, credit quality, transmission responsibility and treatment of 45U. A fixed contract that transfers all upside at a weak price could reduce value even as it increases visibility. A well-structured agreement should compensate PEG for scarcity, preserve operational incentives and allocate transmission and credit risks to the beneficiary.

New generation and Power NJ

PSEG says it submitted multiple concepts to PJM’s proposed backstop process and is evaluating Power NJ. These could be material because fifteen-year commitments and state support can turn new generation into infrastructure-like investment. They are absent from the base case for three reasons: no award, no disclosed capital or return, and unresolved cost allocation. New nuclear’s roughly twelve-year timeline makes it a strategic option rather than a 2030 earnings driver.

The critical discipline is ring-fencing. PEG should not place regulated credit or the existing nuclear franchise behind first-of-a-kind construction risk without a contract that covers prudent cost, delay, financing, tax and operating risks. A ratepayer-benefit requirement is socially sensible but can create asymmetric negotiation if developers bear overruns while customers retain price upside.

Dividend growth

The $2.68 annualized dividend is the fifteenth consecutive increase and represents about 62% of the 2026 EPS midpoint. That payout leaves retention for capital while providing a meaningful cash return. If EPS grows 6%–8% and the payout stays broadly stable, mid-single-digit dividend growth is sustainable. Because simple free cash flow is frequently negative after the full investment burden, dividends are economically funded by consolidated operating cash, retained earnings and access to debt rather than by surplus free cash alone.

Verdict: The 6%–8% plan is achievable without heroic data-center or contract assumptions. Existing-nuclear contracting and milestone-backed regulated investment are upside; unawarded new generation is not yet an investable forecast.

6. Financial Quality and Quality of Earnings

Reported earnings versus operating economics

PEG’s GAAP and non-GAAP results answer different questions. GAAP captures the full change in shareholder accounting value, including nuclear-decommissioning-trust investments and mark-to-market power positions. Operating earnings remove those volatile items to show management’s view of recurring period economics. Neither measure is “the truth” by itself.

In Q2 2026, revenue declined 9.0% to $2.554 billion and GAAP net income fell to $334 million, or $0.67 per share, from $585 million and $1.17. Operating earnings rose to $425 million, or $0.86, from $384 million and $0.77. For the first half, revenue rose 6.2% to $6.402 billion, GAAP EPS fell to $2.15 from $2.35, and operating EPS increased to $2.41 from $2.20. The H1 $128 million adjustment from GAAP to operating earnings reflected a $147 million pre-tax NDT gain, a $299 million pre-tax mark-to-market loss and $24 million of associated tax benefit. Those are genuine accounting changes, but they do not measure delivered power margin in the same period. The filed Q2 presentation provides the segment bridge and the 10-Q supplies the full statements.

The adjustment is economically defensible because an NDT is dedicated to decommissioning obligations and forward positions can reverse before physical delivery. It should not become a license to ignore cash. Hedge settlements eventually enter realized margin, trust assets support a real liability, and tax treatment changes the economics. We therefore anchor forecasting in operating earnings, reconcile every adjustment to GAAP, and test the thesis against cash flow and balance-sheet change.

Five-year financial shape

The table, compiled from PSEG’s 2023 and 2025 Forms 10-K and its Q2 2026 Form 10-Q, exposes three recurring traits. Revenue and GAAP profit are poor trend measures because commodity pass-through and mark-to-market items dominate comparisons. Operating cash is volatile with collateral, tax, working-capital and storm timing. Capital spending persistently consumes most or more than all operating cash before dividends. This is the normal shape of an expanding utility, but “normal” is not the same as free: shareholders rely on regulators and capital markets to convert today’s financing gap into future per-share earnings.

H1 2026 produced $1.821 billion of operating cash against $1.459 billion of PP&E additions, apparently leaving $362 million. After $668 million of dividends, the simple residual was negative $306 million. Moreover, PP&E omits $263 million of energy-efficiency investment and $74 million of removal costs classified in operating cash. Conversion benefited from a $300 million derivative-loss add-back and a $272 million tax-receivable movement, while prepayments and cash collateral consumed $350 million and $192 million. It was an improved half, not evidence that the funding gap vanished.

Returns on capital

GAAP consolidated ROE was approximately 7.3% in 2022, 17.6% in 2023, 11.2% in 2024 and 12.8% in 2025; annualized H1 2026 was about 12.5%. The 2023 spike reflects mark-to-market effects. Annualized H1 operating earnings imply roughly 14% consolidated operating ROE, while PSE&G net income on average utility equity was about 9.1% in 2025 and H1 2026. That utility result is close to authorized returns and fits the regulated model.

A conservative consolidated ROIC—operating income after a 21% statutory-tax proxy divided by average debt plus equity less cash—was approximately 3.3% in 2022, 8.4% in 2023, 5.0% in 2024, 5.9% in 2025 and 5.9% annualized in H1 2026. Accounting noise makes the early values imprecise, but the message is robust: PEG does not earn Greenwald-style 15%–25% commercial ROIC. Its advantage is protection and recovery on a vast capital base, not extreme return on each dollar.

This creates a useful discipline for assessing growth. If incremental investment earns near 9% on equity and is financed partly with debt, per-share earnings can compound at mid-to-high single digits even though consolidated ROIC stays around 6%. Value is created only if the allowed return exceeds financing and operating cost after lag and disallowance. The positive EPS bridge indicates that condition currently holds.

Balance sheet and refinancing

At June 30, securities debt was $24.541 billion: $950 million of commercial paper and loans, $850 million of current maturities and $22.741 billion of long-term debt. Equity was $17.329 billion, placing debt at 58.6% of debt plus equity. Cash plus restricted cash was $216 million. Committed facilities totaled $3.825 billion, with $3.210 billion unused. These values reconcile to the Q2 2026 Form 10-Q.

Liquidity is adequate, but leverage is a structural fact rather than a temporary choice. First-half funding included $500 million of parent 4.80% notes due 2031, $500 million of PSE&G 4.20% notes due 2031 and $500 million of PSE&G 5.63% notes due 2056. The parent retired $450 million of 0.95% notes and prepaid a $500 million term loan. Refinancing very cheap legacy debt raises the interest burden even if market rates do not rise further; H1 interest expense increased 10.6%. The next twelve-month PSE&G maturities include $425 million in September 2026 and $425 million in May 2027.

The “no equity” funding promise matters because shares have stayed around 500 million since 2021. It also removes a pressure valve. If regulatory cash recovery weakens or projects grow beyond plan, PEG must slow capital, accept higher leverage, sell assets or issue equity. Credit capacity is therefore part of the regulatory bargain.

Taxes and EDIT

PSE&G’s effective tax rate increased to 12.4% in H1 2026 from 8.1% as historical mixed-service deduction flowback declined. At 2025 year-end, protected excess deferred income taxes were about $1.2 billion and the tax-related regulatory liability was $1.7 billion. PSE&G returned $436 million of EDIT and repair-related deferred taxes to customers in 2025, reducing tax expense by $313 million. These benefits support customer affordability and timing; they are not an unregulated tax moat and should not be capitalized as permanent margin.

The 45U credit is similarly conditional. PSEG estimated a $350 million benefit for 2024 but none for 2025 because qualifying gross receipts exceeded the phase-out threshold. At weak power prices, the credit supports nuclear economics; at strong prices, market revenue replaces it. A forecast that adds both full high-price margin and full PTC value double counts the floor.

Verdict: Underlying earnings quality is good, but cash conversion is utility-like and debt dependent. Operating earnings are the best forecast base only when reconciled to GAAP marks, collateral, taxes and the full investment burden.

7. Capital Allocation and Management

Portfolio decisions

Management’s most valuable capital-allocation decision was subtraction. PSEG sold approximately 6,750 MW of fossil generation in 2022, exited offshore-wind development and concentrated on regulated networks plus nuclear. That reduced commodity and construction exposure, simplified the balance sheet and raised the proportion of earnings eligible for a utility multiple. The retained nuclear fleet preserved asymmetric exposure to PJM scarcity and federal support.

The current allocation hierarchy is sensible: maintain nuclear safety and reliability; invest in authorized utility infrastructure; fund the dividend; pursue incremental projects only with regulated or contractual risk protection. Management describes new generation opportunities as “utility-like,” an appropriate hurdle. The phrase is not enough. Each award must demonstrate return, counterparty, duration, cost recovery and downside allocation.

Reinvestment and funding

The $24–$28 billion five-year program is more than 90% regulated, with $22.5–$25.5 billion at PSE&G. That is the highest-confidence use of capital because the assets enter a legal recovery framework. Recent evidence shows positive per-share accretion, but approvals are not automatic. The 2024 base case settled below the requested revenue increase; GSMP III was narrowed; the energy-efficiency extension changed return mechanics. Management deserves credit for funding current approvals without new equity, not for treating the entire upper capex range as earned.

Flat shares are a real advantage relative to utility peers that issue stock annually. Diluted shares were 504 million in 2021 and 499 million in H1 2026. Recent repurchases are primarily intended to offset compensation and employee-plan issuance, not exploit valuation. A November 2025 Rule 10b5-1 repurchase plan bought 850,000 shares at $79.50 in December and 865,000 at $83.13 in March. At those prices, buying shares while borrowing for capex was not obviously superior to retaining balance-sheet capacity. Repurchases should remain secondary.

Dividend

The annualized $2.68 dividend is up 6.3% from 2025 and marks fifteen consecutive increases. At the $4.34 guidance midpoint, the payout is 61.8%. That is a sound balance: enough yield to compensate utility shareholders, enough retention to support investment. Dividend growth broadly in line with EPS is sustainable if financing access and recovery remain intact.

The economic funding source is important. Because cash after the full investment burden and dividends is normally negative, the dividend is not covered by conventional free cash flow in many years. It is covered by the combination of regulated earnings, depreciation cash, debt capacity and future recovery. Cutting the dividend is not a base-case risk, but neither should investors mistake the payout for excess cash.

Compensation and insider evidence

CEO Ralph LaRossa’s 2025 total compensation was $13.87 million. Annual incentives weighted operating EPS at 65%, a scorecard at 25% and strategic goals at 10%; the payout factor was 1.37 times. Long-term awards were 70% performance units and 30% restricted shares, with performance units weighted 50% to relative TSR, 25% to operating-EPS growth and 25% to ESG. The 2023–2025 cycle paid at 163% of target. Share-based compensation was $43 million, less than 0.4% of revenue. The incentive design and ownership data are in PSEG’s 2026 proxy.

EPS and TSR alignment is better than vague strategic goals, but neither penalizes inefficient balance-sheet expansion as directly as ROIC, financing-adjusted EPS or customer-affordability performance. During a $24–$28 billion build, the board should reward net per-share value after interest, depreciation, dilution and customer outcomes.

The sixty-month SEC corpus contains no transaction-code “P” open-market insider purchase. Most recent executive sales were scheduled. LaRossa’s August 3 and September 1 sales of 2,083 shares each were under a plan adopted in November 2025 and are low-signal. Director Richard Thigpen’s August 11 sale of 8,000 shares at $74.56 was not under a 10b5-1 plan; he retained 20,970 shares and has sold 23,520 shares in four non-plan transactions over two years. That is mildly negative evidence, not a thesis break. Aggregate officer/director ownership remains below 1%.

M&A and future projects

There is no need for large-scale M&A. PSEG already has a regulated runway and scarce nuclear assets. Buying another utility could import unfamiliar regulation and require equity; buying merchant generation would reverse simplification. Small grid capabilities or contracted projects can make sense, but management should prefer organic investment and contract monetization.

New nuclear is the capital-allocation stress test. A twelve-year project can consume multiple management cycles and expose shareholders to permitting, labor, supply-chain, financing and policy change. Federal loans and state procurement reduce but do not eliminate those risks. PEG should proceed only if contracts allocate construction overruns and delays away from the common shareholder or compensate them explicitly.

Verdict: Historical allocation is a strength: simplification, regulated reinvestment, a covered dividend and flat shares. The next test is whether management maintains that discipline when scarce-generation opportunities invite much larger commitments.

8. Changes and Headwinds — Two-Year View

Changes that strengthened the thesis

The federal nuclear floor became operational. The 45U credit now protects eligible existing generation through 2032 when qualifying receipts are weak, materially reducing the risk that low wholesale prices force uneconomic closure. PSEG’s 2024 benefit and 2025 phase-out demonstrate that the mechanism works as intended rather than as a permanent subsidy.

PJM scarcity became observable. Capacity auctions reached imposed caps, forecast peak demand continued to outrun supply additions, and the 2028/29 auction failed to procure the reliability requirement. PJM’s backstop proposal acknowledges that ordinary market signals have not produced enough timely capacity. Existing nuclear is consequently more strategically valuable than it appeared before 2024.

The regulated plan expanded without current equity issuance. Rate base reached approximately $36 billion, the five-year regulated plan rose to $22.5–$25.5 billion, and shares remained flat. PSE&G’s earnings bridge stayed positive after financing and depreciation. This is better per-share execution than gross-capex headlines alone would imply.

New Jersey created cleaner cost-causation rules for data centers and a process for new nuclear. Requiring large loads to pay dedicated infrastructure costs reduces the risk that existing customers subsidize speculative development. Power NJ gives long-duration nuclear development a state pathway. Both can improve investability if final contracts preserve returns.

Changes that pressure the thesis

Affordability is now the organizing regulatory issue. New Jersey electric bills rose about 20% in 2025, largely because of capacity, according to the NJBPU’s August 2026 report. The Phase 1 report concludes no single utility-business-model reform can solve the bill problem, but its near-term agenda explicitly includes ROE, capital structure, automatic recovery and capital-accountability review. The year-end base filing puts those concepts into a live proceeding.

The RTO-adder law is a concrete headwind. Mandatory participation could remove PSE&G’s 50-basis-point incentive from January 2027 and reduce annual net income and cash by about $40 million. It is included in management’s long-term outlook, but it shows that previously durable regulatory economics can change through legislation.

The data-center narrative became more sober. Management’s inquiry pipeline reached double-digit gigawatts, but it also said historical conversion may be only 10%–20% and that New Jersey interest was leveling. PJM’s 2.91 GW completion-adjusted 2030 forecast is material, yet far below the headline funnel. Investors now require milestones rather than megawatt inquiries.

Nuclear peers moved from narrative to contracts while PEG did not. This does not make PEG’s fleet worse, but it weakens any assumption that a premium agreement is inevitable or imminent. Customer location, transmission constraints, tax incentives, ownership structures and counterparty preferences differ. The opportunity cost of waiting is visible.

Financing costs rose. PSEG replaced sub-1% debt with new paper near 4%–6%, while interest expense already absorbs a meaningful part of the regulated margin bridge. A stable policy rate still implies higher average interest as legacy maturities roll.

Mixed developments

The proposed PJM backstop is both opportunity and risk. Fifteen-year revenue can support useful investment, but cost allocation may create customer backlash and project terms may leave construction risk with developers. PSEG’s submissions are evidence of participation, not value.

Advanced-grid legislation similarly cuts both ways. It encourages dynamic ratings, advanced conductors and power-flow controls, which can improve utilization. It also requires NJBPU certificates for supplemental transmission projects of 100 kV or more, adding state review to projects that previously relied more heavily on PJM and FERC processes.

The Maryland/northern Virginia 500 kV competitive transmission project shows execution risk outside the core territory: the $424 million award no longer expects its directed 2027 in-service date because of Maryland regulatory timing. Formula rates do not eliminate permitting risk.

Verdict: The quality of PEG’s assets improved relative to system need, while the political claim on their economics intensified. Scarcity is real; capture remains negotiated.

9. Risk Analysis

Regulatory downside is the dominant risk

The most plausible route to permanent value impairment is not a one-quarter miss. It is a multi-year regulatory-financing loop. Suppose the NJBPU lowers authorized equity and ROE, disallows capital or lengthens recovery. Cash generation weakens just as interest expense rises. PEG then slows projects, increases leverage or issues equity; each response reduces per-share growth. A lower credit profile raises debt cost and can make a nominal customer-rate cut partly self-defeating. The Phase 1 study recognizes this feedback, which should limit extreme action but does not eliminate political risk.

The quantitative sensitivity is meaningful. Every 50 basis points of return on roughly $18 billion of distribution rate base is about $90 million pre-tax before capital-structure and timing details. The enacted FERC-adder issue alone represents approximately $40 million after tax. A combined return and equity-layer change could absorb a material share of annual EPS growth.

Nuclear tail risk

Nuclear economics have low variable cost but high consequence. An extended outage loses energy and capacity revenue while many costs continue. A safety event can affect multiple units, licensing and public support. Decommissioning obligations and trust performance create long-duration exposure. Insurance pools and regulatory frameworks distribute some risk, but catastrophic scenarios cannot be valued as ordinary volatility.

Ownership reduces diversification. PSEG operates Hope Creek and Salem and holds non-operating interests at Peach Bottom, so one site issue can affect a large share of output. Recent 92%–94% capacity factors are reassuring, not a guarantee. Subsequent license renewal requires continued capital and regulatory approval; depreciation assumptions should not be mistaken for licenses.

Merchant and contract risk

The 45U floor is legislated only through 2032 and depends on compliance and receipts. Strong prices eliminate the credit; weak prices activate it. Policy can change, and the credit does not protect against outages. Hedges stabilize near-term cash but create collateral and counterparty exposure and can defer participation in rising markets.

A long-duration contract could introduce basis, credit and opportunity-cost risk. A data-center counterparty may be creditworthy today yet negotiate technology or location flexibility. Node and transmission rules can separate contract price from plant realization. Contract excitement should never substitute for reading the terms.

“Catastrophic loss is impossible” is the wrong utility framing

Regulated monopolies can recover prudent costs, but equity is not guaranteed. Severe operational failure, criminal cyberattack, nuclear incident, sustained adverse regulation or financing closure can impair capital. The right statement is that ordinary demand competition is unlikely to destroy PSE&G’s franchise. The tail distribution remains real.

Verdict: Regulatory return and financing are the highest-probability threats; nuclear operations are the highest-severity tail. The present valuation compensates for ordinary execution risk, not a catastrophic event.

10. Valuation — What the Current Price Embeds

Current snapshot

At the September 1 close, PEG traded at 17.2, 16.9 and 16.7 times the low, midpoint and high of 2026 operating-EPS guidance. The dividend yield was 3.65%. Using 498–499 million shares, equity value was approximately $36.6 billion. Adding $24.678 billion of total debt and subtracting $192 million of cash produces current-price enterprise value near $61.1 billion. Against $4.390 billion of trailing EBITDA, EV/EBITDA is 13.9 times and net debt/EBITDA is 5.58 times.

Those calculations intentionally mix a live share price with the latest available quarter-end debt, cash and trailing operating data. They should not be confused with a data vendor’s June-quarter valuation, which used a different stock price. EBITDA is also not cash available to equity: it precedes interest, tax, nuclear obligations and enormous capital needs.

Own-history valuation

The most important movement is the composite reset. July’s “full price” premise depended on upper-quartile price-to-book and price-to-sales readings. They are now near the middle of PEG’s own record. TTM P/E did not fall proportionately because TTM GAAP EPS declined to $4.034 when mark-sensitive H1 2026 replaced H1 2025. Forward operating EPS, book and sales are cleaner update metrics.

Own history has limits. PEG exited fossil generation and now deserves a different mix multiple than it did five years ago. Rates, nuclear policy and PJM scarcity also changed. A median historical valuation is not automatically fair value; it simply shows that the earlier statistical premium has dissipated.

Peer cross-check

Using September 1 prices and each company’s current 2026 adjusted-EPS guidance midpoint produces forward P/Es of roughly 24.6 times for Entergy, 19.8 for Ameren, 19.2 for Evergy, 18.5 for Xcel and 17.6 for CMS. The five-company median is 19.2 times, about 12% above PEG’s 16.9 times. These companies were selected from PEG’s empirical utility-factor neighbors, not because their assets are identical.

The discount is directionally useful, not a mechanical arbitrage. Entergy has different nuclear and jurisdiction exposure; Ameren and Evergy differ in growth, financing and regulation; Xcel has a distinct capital plan; CMS is more purely regulated. PEG carries merchant nuclear volatility and a concentrated New Jersey relationship, but it also has no planned equity issuance and owns scarce generation. The peer test says PEG is no longer expensive among comparable utility risk profiles. It does not prove that peers themselves are cheap.

Merchant-nuclear owners are poor direct comps. Constellation, Vistra and Talen have different leverage, contract portfolios, generation mixes and earnings volatility. Applying their multiples to PEG’s 3,758 MW while also valuing PSE&G at a utility multiple can double-count central costs, debt and tax benefits. A sum-of-the-parts can be informative only with normalized nuclear cash flow, explicit net debt allocation and decommissioning obligations; present disclosures do not make a single precise SOTP superior to earnings scenarios.

Embedded growth

Compounding the $4.34 2026 midpoint at management’s 6%–8% range produces 2030 operating EPS of approximately $5.48–$5.90. The current quote is 12.4–13.4 times that prospective earnings range before dividends. That comparison is not an exit valuation; it shows how much growth has to arrive before today’s multiple falls into the low teens.

Assume four annual dividends from 2027 through 2030, growing 6%, with no reinvestment. To earn a 10% annualized total return, required EPS growth is about 8.2% at a 16-times terminal multiple, 6.6% at 17 times and 5.1% at 18 times. Thus an unchanged 17-times multiple requires growth near the center of management’s outlook. One turn of multiple compression requires slightly more than the high end; one turn of expansion permits under-delivery.

These are return sensitivities, not forecasts of future market prices. The bear assumes regulation, financing and weaker nuclear economics reduce growth below guidance. The base assumes delivery at the midpoint, stable shares and no existing-fleet contract. The bull requires evidence outside the plan: contracted nuclear monetization, successful uprates or backstop awards, or load conversion without adverse cost allocation. The return calculation excludes dividend reinvestment and uses a clean FY2026-to-FY2030 four-year convention.

The asymmetry is now reasonable rather than extraordinary. The bear still produces a small positive annual return because dividends and some growth offset multiple compression; it does not model a severe nuclear or regulatory event. The base reaches a double-digit annualized return without a heroic catalyst. The bull requires both better earnings and re-rating, so it should carry a materially lower probability until a contract appears.

Factor and regime lens

FactorsToday estimates PEG as a defensive, dividend-sensitive utility. In its base style model, Dividend Yield exposure is +0.787 and Momentum +0.356, while empirical Quality is -0.423 and Value -0.374. The “Quality” coefficient is a statistical portfolio loading, not a conclusion that PSE&G is low quality. In the broader model, Utilities is the dominant exposure at +1.09 and explained fit rises from 34% to 68%; stock-specific annualized volatility is 11.5%. Empirical neighbors are utility ETFs and regulated companies, supporting the peer set above. Methodology is available from FactorsToday.

Recent regime data are mildly adverse but not extreme. The pure Utilities factor was negative over 21, 63, 126 and 252 trading days, but no dominant factor reached an absolute two-standard-deviation signal. PEG’s one-year risk-adjusted record is poor, while the five-year annualized price return is only about 6.2% with 20.4% volatility. Factor data therefore support staged timing and humility, not a claim that negative momentum predicts deteriorating earnings.

Verdict: The current price embeds roughly mid-guidance growth at a stable utility multiple and assigns limited value to an unsigned nuclear catalyst. Valuation is no longer the objection; regulatory execution and contract evidence are.

11. Variant Perception

The superficial bullish view is that PEG owns nuclear capacity in a power-short region and will eventually sign a hyperscaler contract. The superficial bearish view is that New Jersey will cut utility returns to solve affordability. Both statements contain truth and both are too coarse for an edge.

Our first variant is that the regulated core, not a data-center contract, determines whether the current return framework works. At today’s forward multiple, delivery around the 6%–8% framework can produce a satisfactory modeled return with dividends and a stable multiple. That reduces catalyst dependence. A contract would be incremental rather than necessary to the base arithmetic.

Our second variant is that New Jersey regulatory risk is more concrete than a generic “constructive jurisdiction” label suggests, but less economically simple than a blanket ROE-cut thesis. The state has already narrowed programs, altered efficiency-return mechanics, added transmission review and placed ROE/capital structure on the agenda. Yet capacity drove most of the bill increase, distribution is only one-quarter of the bill, and PSE&G’s 9.60% ROE is near the national median. A punitive cut could raise financing cost and undermine investment. The probable result is tougher accountability and selective rationing, not confiscatory regulation; the probability distribution is wider than consensus shorthand.

Our third variant is that the 11.5 GW inquiry pipeline is not the relevant data-center forecast. PJM’s completion-adjusted 2.91 GW for 2030 is still economically meaningful, but it is much closer to decision-grade evidence. Using the larger headline produces illusory upside and encourages anticipatory capital. Milestones, deposits and measured load should move estimates; inquiries should not.

Our fourth variant is that 45U is an insurance policy, not a second revenue stream. In low-price years it supports nuclear; in high-price years it phases out. PEG’s 2024 and 2025 outcomes demonstrate the offset. Analysts who add a full credit to strong merchant prices overstate normalized earnings, while analysts who ignore the floor overstate closure risk.

Finally, peer contracting is both encouraging and cautionary. It validates the demand for long-duration nuclear output but also shows that counterparties have alternatives. PSEG’s fleet is scarce, not unique. A delayed contract can reflect transmission, ownership, location or pricing discipline. We prefer management to wait rather than sign a weak agreement, but delay has an opportunity cost.

What the market may be missing is not a secret project. It is the combination of a still-positive regulated per-share bridge, flat shares, a nuclear downside floor and a valuation reset sufficient to make base execution investable. What bulls may be missing is how readily policy can capture scarcity rent before shareholders do.

Verdict: The differentiated case is modest: base execution matters more and optionality less at the new valuation, while regulatory risk deserves more granular monitoring than either bulls or bears typically apply.

12. Fact Versus Interpretation

The table matters because PEG’s narrative contains many category errors. A pipeline can be a management fact and still be unsuitable for valuation. A tax credit can be a factual floor and still not be additive to high prices. An authorized return can be a fact while future earning of it remains a forecast. Separating those layers prevents a compelling system-need story from turning into unsupported earnings.

Verdict: The hard evidence supports the base earnings engine and asset scarcity. It does not support assuming a contract, a new-build award or unchanged regulatory terms.

13. Open Questions and Decision-Relevant Diligence

The year-end base-rate filing

What revenue increase, rate base, ROE and equity ratio will PSE&G request? Which investments move from riders into base rates, and what is the proposed customer-bill effect? How much historic capital faces prudence review? The filing is only the opening position; the decision-useful comparison is request versus staff testimony, settlement and final order.

Track four thresholds. First, is requested and authorized ROE close to the current 9.60%? Second, does the equity layer remain near 55%? Third, are material assets disallowed or deferred? Fourth, does recovery timing keep the incremental margin bridge positive after interest and depreciation? A small ROE adjustment with intact equity and recovery may be manageable. A coordinated cut across all four is not.

Phase 2 utility-model review

Which performance metrics will the NJBPU quantify, and are rewards symmetric with penalties? Will capital-accountability tools apply prospectively or revisit approved programs? How will credit and financing effects be modeled? Can shared savings reward measurable customer outcomes without reducing the incentive to invest?

The key distinction is base distribution versus program-specific mechanics. The energy-efficiency extension already reduced conditional returns; extrapolating it to all assets would be premature. Conversely, treating Phase 1 as an academic exercise ignores its explicit near-term agenda.

Existing-nuclear contract economics

Is PSEG prioritizing Hope Creek/Salem in New Jersey, Peach Bottom in Pennsylvania, or a portfolio structure? What volume is available after hedges, and when? Would a contract include capacity, energy, clean attributes and uprates? Who pays transmission and interconnection costs? How does 45U value flow between generator and buyer? What credit support survives a data-center relocation or delay?

No announcement should change estimates without enough terms to model cash. Duration alone is insufficient. A premium contract can transfer outage, basis or opportunity-cost risk back to PSEG. The high-quality outcome is a creditworthy counterparty, minimum payment, appropriate escalation and preservation of upside for operational performance.

PJM backstop submissions

Which technologies and sites did PSEG submit, what capital is required, and does PSEG propose ownership or service? Are fifteen-year payments indexed? Who bears construction delay, interconnection, fuel and capacity-performance penalties? Does cost allocation create exposure for PSE&G customers and therefore political opposition?

Until PJM and FERC approve a mechanism and select projects, the submissions are strategic positioning. A contracted gas or uprate project may fit PEG’s hurdle; a merchant plant with long construction and fuel risk may not.

Data-center conversion

How much of the 2.91 GW 2030 completion-adjusted path has binding service agreements, deposits and construction milestones? Which loads are behind the meter, curtailable or required to bring generation? What rate-base spending is customer funded versus socialized? Does actual coincident peak track the forecast from 394 MW in 2025 to 785 MW in 2026?

Future disclosures should reconcile the inquiry funnel to PJM stages. A rising headline pipeline with flat milestone-backed load is negative evidence. A smaller funnel with executed agreements is positive.

Nuclear longevity and operating cost

What capital and outage schedule accompanies the Salem uprates? When will Hope Creek and Salem subsequent-license applications be accepted, and what incremental safety or environmental spending is required? How does fuel procurement respond to geopolitical supply restrictions? What level of sustaining capex is embedded in the 6%–8% outlook?

The current capacity factor and fuel cost are strong. License-extension value should be discounted until applications, NRC milestones and required capital are visible.

Funding and credit

What FFO-to-debt and parent-credit thresholds underpin the no-equity plan? How much of the upper $28 billion program is discretionary? What average coupon does management assume as legacy debt matures? Would a larger regulated or generation award be incremental to the no-equity commitment or replace existing projects?

The plan remains credible at current scope. The question becomes more important if management celebrates major awards without simultaneously updating financing.

Capital efficiency and governance

Will the compensation committee add ROIC, financing-adjusted growth or affordability metrics? Why has there been no open-market insider purchase in five years, and what explains repeated non-plan director sales? Neither issue determines value alone, but better efficiency alignment would strengthen confidence through the regulatory cycle.

Verdict: The next twelve months are unusually rich in falsifiable evidence. The rate filing, data-center milestones, contract terms, NRC process and funding disclosures can move the thesis from narrative to measurable outcomes.

14. What Must Be True

Base case: regulated compounding carries the investment

The following must hold together:

1. PSE&G earns close enough to its authorized return that rate-base investment produces 6%–8% consolidated operating-EPS growth after depreciation, interest and operating cost.
2. The year-end base case and Phase 2 review preserve a workable equity layer and avoid material disallowance, even if they add reporting and performance conditions.
3. Shares remain approximately flat and the balance sheet retains investment-grade funding access.
4. Nuclear capacity factors stay around the low-90% range, hedge realization and capacity payments support the plan, and 45U behaves as a downside floor.
5. No long-duration nuclear contract, new nuclear or backstop award is required to reach the plan.

Base falsification: two consecutive reporting periods in which incremental distribution/transmission margin fails to cover incremental O&M, depreciation and interest; a financing update that introduces material common equity for the existing plan; or formal guidance below 6% growth without a clearly temporary cause. A modest annual miss does not by itself falsify a four-year compound thesis; a broken per-share bridge does.

Bull case: scarcity becomes contracted value

The bull case requires evidence beyond high PJM prices:

1. PEG signs a long-duration agreement for existing output or uprates with creditworthy minimum payments, attractive escalation and sensible outage, transmission and 45U treatment.
2. The Salem uprates proceed on time and on budget, with license renewal extending the economic life of the sites.
3. PSEG wins backstop or Power NJ opportunities under contracts that cover construction, financing and operating risk and produce utility-like returns.
4. Data-center load converts near or above PJM’s completion-adjusted path, and dedicated rates prevent cross-subsidy backlash.
5. The NJ settlement remains near current return/equity parameters, allowing upside to reach shareholders rather than offsetting regulatory erosion.

Bull falsification: another year with no signed existing-fleet contract while peers continue to contract; disclosed terms that merely cap merchant upside without adequate floor or premium; repeated uprate/license delays; backstop or Power NJ awards requiring uncapped shareholder construction risk; or 2030 load milestones falling materially behind PJM’s 2.91 GW path. A press release naming negotiations is not evidence.

Bear case: regulation and financing overwhelm gross growth

The bear case requires a chain, not one headline:

1. NJBPU cuts allowed ROE materially below 9.60%, lowers the 55% equity layer, disallows capital or lengthens recovery.
2. The RTO-adder loss and refinancing at higher coupons compound the regulatory reduction.
3. PSE&G’s recovered margin no longer covers incremental depreciation, O&M and interest on a per-share basis.
4. PEG must issue equity, sell a valuable asset or reduce high-return investment to protect credit.
5. Nuclear performance or realized pricing weakens while policy support or contract opportunities disappoint.

Bear falsification: a base settlement near present parameters with limited disallowance; continued positive net regulated EPS contribution; stable or improving credit metrics; and funding of the stated program without equity. Those outcomes would show that affordability policy is modifying, not breaking, the compact.

Monitoring dashboard

Both July triggers are still open. The bull contract threshold has not been met; the bear regulatory threshold has not been met. The appropriate response is to update probabilities as evidence arrives, not to declare victory from a cheaper price.

Verdict: PEG works if the regulated core compounds without dilution and nuclear merely performs. It becomes exceptional if scarcity is contracted on good terms; it fails if regulation turns gross capital growth into negative per-share economics.

15. Public Source Appendix

Company and SEC filings

• PSEG, “PSEG Announces Second Quarter 2026 Results”, August 4, 2026, company earnings release.
• PSEG/SEC, Q2 2026 Form 10-Q, filed August 4, 2026, regulatory filing.
• PSEG/SEC, Q2 2026 results presentation, filed August 4, 2026, earnings exhibit.
• PSEG/SEC, 2025 Form 10-K, filed February 26, 2026, annual regulatory filing.
• PSEG/SEC, 2026 proxy statement, filed March 12, 2026, proxy filing.
• PSEG, “PSEG Announces 2025 Results”, February 26, 2026, company earnings release.

Regulation, power markets and policy

• NJBPU/E3, “New Jersey Electric Utility Business Model Study: Phase 1”, July 2026, commissioned regulatory study.
• NJBPU, PSE&G 2024 base-rate order, October 9, 2024, final regulatory order.
• NJBPU GSMP III order, November 21, 2025; and Energy Efficiency Triennium 2.5 order, July 15, 2026.
• New Jersey energy and data-center legislation release, July 7, 2026.
• New Jersey Power NJ Act release, July 13, 2026.
• NJBPU New Jersey capacity-market report release, August 13, 2026.
• PJM, PSE&G 2026 load-forecast documentation, January 2026, system-planning forecast.
• PJM, Reliability Backstop Procurement overview, July 31, 2026, market-operator release.
• IRS zero-emission nuclear power production credit, accessed September 2, 2026.
• NRC subsequent-license-renewal schedule, updated August 27, 2026.

Market and methodology

• AZI Trading PEG historical market data, data through September 1, 2026.
• FactorsToday methodology and PEG factor endpoints, accessed September 2, 2026.
• PJM 2024 market review, January 7, 2025.
• Reuters, “US power stocks plummet as DeepSeek raises data center demand doubts”, January 27, 2025, authoritative secondary market report.
• Nuclear-contract comparisons: Talen, Vistra and Constellation 2025 Forms 10-K.
• Utility valuation cross-check: 2026 guidance from Entergy, Ameren, Evergy, Xcel and CMS, each published with Q2 2026 results.

This report is research, not individualized investment advice. Forward scenarios are sensitivities rather than assurances.