Duke Energy Corp DUK

Thesis

Duke Energy operates regulated electric and gas utilities in the Carolinas, Florida, the Midwest, and Indiana, serving roughly 8.4 million electric customers and 1.7 million gas customers. The business is the canonical Buffett utility: an essential service offered under cost-of-service regulation, where state commissions allow Duke to earn an authorized return (roughly 9.6%-10.5% ROE depending on jurisdiction) on a rate base that grows with each capital cycle. Earnings compound to the extent regulators let Duke spend capex and recover it in rates. The growth runway is real: Carolinas load is accelerating after two decades of flat demand, driven by data centers, onshoring, and Sun Belt population growth. Management guides to ~$83B of capex over five years, which on a current ~$120B rate base translates into rate-base growth of roughly 7-9% and EPS growth in the 5-7% band, plus a ~3.6% dividend yield.

The scorecard tells the rest of the story. ROIC is structurally low (10y avg 3.96%) because the business is capital-intensive; that is a feature, not a bug, for a regulated utility. FCF conversion is negative (-37.4% over 5y) because capex exceeds operating cash flow every year — the deal is to fund the gap with debt and equity and earn a regulated return on the new asset base. Net debt/EBITDA of 6.06x and interest coverage of 2.34x are at the upper end of investment-grade utility norms, which means the model needs cooperative regulators and stable rates to keep working.

IV math: the scorer's range is $139-$264 with base $208 against a $129 price (px/IV 0.62). But the scorer flags maintenance capex uncertainty twice and explicitly says 'widen the range.' I lean to the low end of IV. Buy with conviction below ~$110; trim above ~$165.

Moat

Duke's moat is the textbook regulated-utility monopoly: a state-granted exclusive franchise to deliver electricity within defined service territories, paired with a regulated cost-of-service rate structure that, in a constructive jurisdiction, allows recovery of prudent capex plus an authorized ROE. Buffett has written about this model repeatedly. He praises BHE's 'recession-resistant earnings, which result from these companies offering an essential service for which demand is remarkably steady' [3], and notes that the core compact is a 'state-by-state promise of a fixed return on equity' in exchange for the willingness 'to finance their growth' through huge multi-year capital programs [4]. Duke fits this template exactly.

Moat type 1 — Cost advantage / scale. Duke Energy Carolinas plus Duke Energy Progress together run one of the largest regulated generation fleets in the United States, with a balanced mix of nuclear (the Oconee, McGuire, Catawba, Robinson, Brunswick, Harris, and Crystal River units), combined-cycle gas, coal in run-off, hydro, and a growing solar/storage book. Once an integrated nuclear-and-transmission fleet is built and rate-based, no greenfield competitor could replicate it economically — the all-in levelized cost of an existing baseload nuclear plant is far below the cost of building new generation, and the transmission/distribution network represents a sunk-cost barrier with replacement value of hundreds of billions. Stress test: a $10B competitor with five years cannot get a CPCN to overbuild in Duke's territory, cannot buy land for transmission, and cannot poach customers because retail choice does not exist in the Carolinas, Florida, or Indiana.

Moat type 2 — Intangibles (regulatory franchise). Each operating subsidiary holds an exclusive certificate to serve a defined territory granted by the state public utility commission. This is the deepest part of the moat. NCUC, PSCSC, FPSC, IURC, and PUCO each set rates and authorize ROE. Buffett's framing is that the utility holds up its end (reliability, customer satisfaction, prudent investment) and 'in turn we look to our utilities' regulators … to allow us an appropriate return on the huge amounts of capital we must deploy' [6]. Duke's Carolinas regulators have historically been constructive: multi-year rate plans (MYRP) under HB 951 in North Carolina formalize forward-looking rate setting with annual adjustments, performance incentives, and storm-cost securitization mechanisms. The 2024 South Carolina rate case and the 2025 NC rate case both produced workable outcomes.

Moat type 3 — Switching costs. Customers cannot switch — there is no retail competition in Duke's largest jurisdictions. This is not a switching-cost moat in the SaaS sense; it is a legal monopoly. Customer satisfaction nonetheless matters because dissatisfied customers complain to commissions and weaken Duke's regulatory standing. JD Power scores for the Carolinas have been mid-pack — adequate but not BHE's '95.3% very satisfied' [4].

Moat type 4 — Pricing power. Bounded. Duke can raise rates only to the extent commissions agree they reflect prudent recovery. In a benign environment this is enough to keep ROE near authorized levels. In a hostile environment — California PG&E wildfire bankruptcy, Hawaiian Electric — pricing power evaporates. Buffett warned in 2023: 'the regulatory climate in a few states has raised the specter of zero profitability or even bankruptcy … the fixed-but-satisfactory-return pact has been broken' [Munger excerpt #4]. Duke is not in California or Hawaii; it is in jurisdictions that have so far honored the compact. But this risk is not zero, especially as climate change drives storm and wildfire exposure.

Moat type 5 — Network effects. Not applicable.

Erosion risk: the moat erodes if (a) state commissions lag-recover capex (regulatory lag widens), (b) hurricanes/wildfires create casualty losses regulators refuse to securitize, (c) FERC opens transmission to merchant build, or (d) distributed generation plus storage reaches grid parity for commercial customers and bypasses the retail rate.

Moat verdict: NARROW. The franchise is durable but the returns are capped, the moat is a function of a legal arrangement that can change, and ROIC of 3.96% over a decade is not the signature of a wide-moat business — it is the signature of a regulated one.

Management & Capital Allocation

Duke's capital-allocation playbook is structurally constrained: the dominant use of capital is reinvestment in rate base, because that is the only way a regulated utility creates shareholder value. Buffett and Munger have endorsed this model when run well — 'Since we purchased MidAmerican ten years ago, it has never paid a dividend. We have instead used earnings to improve and expand our properties' [6]. Duke does pay a dividend, which is the convention for publicly listed utilities, but the principle is the same.

Five choices:

1. Reinvestment. The headline. Duke is mid-cycle on a five-year ~$83B capital plan, the largest in its history, weighted to grid hardening, transmission for data-center load growth in the Carolinas, Lake Norman gas conversions, solar/storage in the Carolinas and Florida, and the Florida coal-to-gas transition. ROIIC of 10.98% over five years (per scorecard) is well above WACC of roughly 6%, indicating that incremental capital is earning the authorized regulated return. This is the single most important number in the analysis: the engine works as long as ROIIC stays above cost of capital. Grade for reinvestment: B+. Capex is large, mostly mandatory, and earning a fair return.

2. Acquisitions. After divesting the Commercial Renewables business in 2023 (sold to Brookfield for ~$2.8B) and exiting the Latin American power business years prior, Duke has shrunk back to a pure-play regulated utility footprint. The Commercial Renewables sale was sensible — sub-scale, capital-intensive, contract-dependent, and not in rate base. The earlier merger with Progress Energy (2012) gave Duke its current Carolinas dominance and is the foundational scale advantage. Recent acquisition discipline: A.

3. Debt. Net debt / EBITDA of 6.06x and interest coverage of 2.34x are high in absolute terms but standard for the industry. The maturity ladder is well-spread, with most debt issued at the holding-company and operating-subsidiary level at investment-grade ratings (BBB+ at parent, A-/A at operating utilities). The risk: rising rates compress equity returns because incremental debt costs more, and the regulatory lag in updating authorized ROE means a few years of underearning can compound. Grade for debt management: B-. Acceptable, but the balance sheet is not a fortress.

4. Buybacks. Effectively none. Share count change over 10 years is +1.4% (per scorecard), reflecting modest issuance net of small repurchases. This is correct policy: a utility trading near IV (or even above IV-low) should not be repurchasing. Issuance funds rate-base growth at par, which is value-neutral if shares are issued near IV. The scorecard does not provide an avg-buyback-P/IV figure because there is no meaningful buyback program. Grade for buyback discipline: B (correct restraint, not a positive contribution).

5. Dividends. Duke is a Dividend Aristocrat-adjacent name with ~99 consecutive years of dividend payment and consistent annual increases. Current dividend ~$4.18/sh, yield ~3.25% on $129. The payout ratio (~70% of EPS, but 100%+ of FCF given negative FCF conversion) is funded by issuing new debt and equity to grow the asset base — this is the regulated-utility model, not a flaw, but it does mean the dividend is structurally dependent on continued market access and constructive regulation.

Communication quality: investor relations is professional, with clear five-year capex roadmaps, jurisdiction-by-jurisdiction rate-case calendars, and consistent 5-7% EPS growth targets. Management has not promised what it cannot deliver. The CEO transition to Harry Sideris (effective April 2025) is recent — execution risk during a record capital cycle is non-trivial but not alarming.

Capital allocator: B. The framework is correct, the discipline on M&A and buybacks is good, and the reinvestment engine works. The grade is not higher because the balance sheet has no slack, the dividend is structurally fragile (negative FCF), and the model leaves no room for error if a single major jurisdiction turns hostile.

Industry Structure

Porter's Five Forces for the U.S. regulated electric utility industry, as applied to Duke:

1. Threat of new entrants — VERY LOW. Entry requires a state-granted Certificate of Public Convenience and Necessity, eminent-domain authority for transmission, decades of capex, and a willingness to operate under a regulated ROE cap. No rational private actor enters this business. The only meaningful competitive threat is behind-the-meter solar plus storage, and even there, regulators have allowed Duke to recover lost throughput through fixed charges and demand-based tariffs.

2. Bargaining power of suppliers — LOW TO MODERATE. Fuel (natural gas, uranium, coal) is a commodity pass-through under fuel adjustment clauses in every Duke jurisdiction. Equipment (transformers, turbines) has been in tight supply post-2022 and lead times have stretched, but cost increases are recoverable in rates. Labor is unionized in some plants but cooperatively managed. Capital is the binding 'supplier' — utilities depend on the bond and equity markets, and rising rates have raised the all-in cost of capital, which works through into customer bills with a one- to three-year lag.

3. Bargaining power of buyers — STRUCTURALLY LOW, BUT POLITICALLY MEDIUM. Residential and small commercial customers cannot switch suppliers in the Carolinas, Florida, or Indiana. Their power expresses itself through the political process — commissions are appointed by elected officials, and rate cases are public. Large industrial customers (manufacturers, paper mills) have more leverage and can negotiate bilateral special contracts. The new and rising buyer is the hyperscale data center: AWS, Google, Meta, and Microsoft are signing multi-hundred-megawatt deals in the Carolinas and demanding clean-energy supply, but they want firm power and are willing to share infrastructure cost — net positive for Duke.

4. Threat of substitutes — RISING. Rooftop solar plus battery storage is the substitute. Today it is uneconomic for most residential customers in the Southeast under net-metering rules Duke has helped shape. Five to ten years out, if storage costs fall another 40-50% and federal tax credits persist, behind-the-meter generation could erode kWh sales. The mitigant: Duke earns return on rate base, not on kWh sold, and as long as commissions allow grid-connection charges to recover fixed costs, distributed generation is more annoyance than threat. Microgrids for industrial customers are a more credible long-term threat, particularly for data centers.

5. Rivalry among existing competitors — NONE within service territory. The whole point of the regulated monopoly is that rivalry is replaced with a regulator. Competition exists only in capital markets (for investor dollars relative to other utilities) and in regulatory beauty contests (Duke vs. Dominion vs. NextEra reputations).

Value pool: the value pool sits at the meter — Duke captures the spread between authorized ROE and its true cost of capital, scaled by the size of its rate base. The pool is growing because (a) Carolinas load is finally accelerating after twenty flat years, driven by data centers and Sun Belt migration, and (b) the energy transition requires massive new capex (gas, nuclear extension, solar, storage, transmission) that will all flow into rate base. Trajectory: positive for the next decade, assuming no regulatory rupture.

Industry Verdict: GOOD. Not Excellent because returns are capped and capital intensity is brutal; not Average because the franchise is genuinely durable and the value pool is expanding for the first time in a generation.

Inversion (Bear Case)

I am now the short. Here is the strongest credible bear case.

1. The single event that kills this. A Category 4 hurricane makes landfall in the Florida Panhandle or eastern North Carolina, takes out 1.5-2 million Duke customers, and triggers $5-8B of restoration spend. South Carolina or North Carolina commissioners — under populist political pressure from rate-shocked customers and a state legislature with shifted composition — refuse to fully securitize the storm costs and impose a $2-3B disallowance. Simultaneously, the wildfire risk that Buffett flagged at BHE in 2023 [Munger excerpt #4] manifests in the Carolinas as ignition liability from Duke transmission lines during a drought year, producing a multi-billion-dollar civil settlement that the commission likewise refuses to fully recover. The 'fixed-but-satisfactory-return pact' breaks at Duke. Equity falls 40-50% on a multiple compression and dividend cut. This is not hypothetical — it has happened to PG&E (Chapter 11 in 2019, equity wiped out) and is actively threatening Hawaiian Electric. Duke has more storm exposure than either. Hurricane Helene in September 2024 was a preview: it caused billions in damage in western NC and is referenced in the 10-K. The question is not whether the next storm comes; it is whether the regulator absorbs it.

2. Why the moat is narrower than bulls think. Bulls treat the regulated franchise as a permanent asset. It is a political contract that can be amended. North Carolina HB 951 was passed under one political coalition; it can be repealed or weakened under another. The shift to performance-based regulation (PBR), increasingly popular in commissioner conferences, would narrow the gap between authorized ROE and actual ROE by creating clawbacks for service quality, reliability misses, and bill-impact thresholds. Distributed generation is approaching grid parity for commercial customers in Florida and the Carolinas — a 5MW corporate solar-plus-storage system is now competitive with retail rates for many industrial loads, and Duke's wedge is shrinking. Behind the meter is the only place a competitor can attack, and the attack is starting. Critically, ROIC over the past decade has averaged just 3.96% (per scorecard) — that is what 'wide moat' looks like in this industry. It looks the same as 'narrow moat' or 'no moat' from an unlevered-return standpoint. The only thing standing between Duke and a commodity-style return is the regulator.

3. Why management is worse than it appears. The capex plan has grown from $63B (2023 plan) to $83B (current plan) in two years. Each upward revision implies prior plans were wrong. Capex inflation in transformers, switchgear, and labor has been 15-25% per year — the company's per-MW build cost is not what it was. The Edwardsport IGCC plant in Indiana is a $3.5B+ embedded loss that was completed years late and over budget; it earns less than its capital cost and is a precedent for how big-bet generation projects go wrong. Coal ash remediation in the Carolinas was a multi-billion-dollar disallowance battle. The new CEO Harry Sideris took over in April 2025 in the middle of the largest capex cycle in company history; execution risk is real. Communication is professional but the EPS growth target (5-7%) requires perfect ratemaking, perfect storm response, and perfect execution simultaneously.

4. What bulls are extrapolating that won't hold. Bulls extrapolate Carolinas load growth at 2-4% indefinitely. But the data-center demand surge is concentrated in a handful of hyperscale customers, and they are simultaneously announcing on-site generation deals (Microsoft + Constellation at Three Mile Island, Amazon + Talen at Susquehanna, Google + Kairos for SMRs). If hyperscalers structurally bypass the utility for their incremental load — taking nuclear behind-the-meter, signing PPAs with merchant generators — Duke captures only the residential and small-commercial residual, which grows much slower. Bulls also extrapolate 5-7% rate-base CAGR translating linearly to EPS growth; in a higher-cost-of-capital world this requires equity issuance that dilutes per-share growth. The last share count expansion (+1.4% over a decade per scorecard) understates the issuance the next decade will require to fund $83B+ capex while maintaining BBB+ credit metrics. Realistic per-share EPS growth is closer to 3-5%, not 5-7%.

5. Valuation trap. P/E TTM is 22.5x vs. 10y avg of 33x — bulls call this cheap. It is not cheap; the 10y average was inflated by a zero-rates regime that is gone. Normalized utility multiples in a 4-5% Treasury world are 16-19x earnings. EV/FCF is meaningless (3,880x per scorecard) because FCF is structurally negative — the entire business model is 'spend more than you earn, fund with debt and equity, earn return on the new asset.' That model breaks if (a) cost of capital exceeds authorized ROE (it nearly does today: 30-yr utility bonds at 5.5-6%, authorized ROE ~10%, levered equity return ~9%), (b) regulators tighten ROE to reflect the lower long-rate world they assume is coming, or (c) equity markets refuse to fund issuance at NAV. The reverse-DCF implied growth of 1.9% (per scorecard) is roughly fair: at $129 the market is paying for low-single-digit terminal growth, which is what utilities deliver across cycles. The IV-base of $208 assumes 5-6% growth in perpetuity; that is bullish.

If I am right, the stock could be worth $80-95 within 3-5 years.

Lollapalooza Bias Check

Biases active in me as the analyst right now:

Authority bias. Buffett owns BHE and writes lovingly about regulated utilities. The canon excerpts in this brief are heavily Buffett, and his praise of MidAmerican/BHE [1][3][4][6] is doing real work in my head. I notice myself reaching for his framings ('essential service,' 'recession-resistant,' 'fixed return on equity') as if they automatically apply to Duke. They mostly do — but BHE is privately held inside Berkshire, has earnings diversification across many states and pipelines, and crucially has 'the inherent advantage of the company being owned by a strong parent' [3] that lowers its cost of debt. Duke does not have that. I should weight Buffett less and the Duke-specific facts more.

Anchoring. The scorecard reports IV-base of $207.77 against a $129 price, implying 60% upside. That number is anchoring me toward a Buy. But the same scorecard flags TWICE that maintenance capex is uncertain with >50% spread and explicitly says to widen the IV range. The honest read is that the IV is closer to the low end ($139) than the base, and possibly below it. The reverse-DCF implied growth of 1.9% is the more disciplined number — it tells me the market is paying for low growth, not pricing a deep discount.

Recency. I am writing this in 2026 against a backdrop of (a) AI-driven data-center load surge that is genuinely transforming Carolinas demand, (b) higher long rates that have pressured utility multiples, and (c) Hurricane Helene's recent damage. All three are top of mind and may be over-weighted in my forecast. The data-center thesis in particular is the most consensus 'utility renaissance' narrative on the Street; that should make me skeptical, not enthusiastic.

Social proof. Utilities are widely owned by income funds, retirees, and ESG mandates. The shareholder base creates a structural bid. I should not confuse that bid with intrinsic value — bond-proxy buyers exit fast when rates move.

Commitment / consistency. I have been positive on regulated utilities as a category for years. I notice myself wanting Duke to fit the framework. The honest counter: Duke's ROIC of 3.96% over a decade is a low number even for a utility, the balance sheet is not fortress-grade (interest coverage 2.34x), and FCF has been negative for years. A genuinely Buffett-quality utility looks more like a long-duration bond with optionality on rate-base growth — and the entry yield matters enormously.

Deprival super-reaction. There is no good way to express the 'storm/wildfire goes wrong' downside in a base case, so I keep tucking it into a low-probability tail. PG&E investors did the same thing for years before the tail bit. I should price the tail more honestly: 5-10% probability of a multi-billion-dollar non-recovery event over the next decade.

Net effect of the bias check: I am inclined to a Hold, not a Buy, at $129. A Buy at $105-110 would clear most of these biases by paying for the downside cases.

10-Year Outlook

Same fundamental business model in 2036? Yes, with very high probability. Cost-of-service regulation will persist; the energy transition only deepens the political need for regulated entities that can finance and build. Distributed generation will erode the edges, but the wires-and-poles plus baseload generation business is structurally entrenched.

Customer base larger? Modestly larger. Carolinas population growth (1-1.5% per year), continued Sun Belt migration, and electrification of heating and transport push residential customer counts up perhaps 10-15% over the decade. The bigger driver is per-customer load: data centers and industrial reshoring may add 30-50% to total Carolinas load by 2036.

Profit per customer higher? Yes nominally, flat to slightly down in real terms. Authorized ROEs have drifted lower over thirty years (from ~12% in the 1990s to ~9.6-10.5% today) and that drift continues. Inflation in capex erodes real returns unless authorized ROE keeps pace. Net: nominal earnings per customer grow with rate base; real earnings per customer roughly tread water.

Moat wider? About the same. The regulatory franchise is binary — either it holds or it breaks — and there is no path to widening it. Risk of narrowing is greater than of widening.

Single biggest threat? A multi-billion-dollar storm or wildfire liability that the regulator refuses to fully securitize, triggering a Hawaiian Electric / PG&E-style reratings event. Climate change makes this more likely than it was. The North Carolina and South Carolina commissions have been constructive historically, but 'historically' is exactly the wrong word for climate tail risk.

Second biggest threat: cost-of-capital squeeze. If long rates stay at 4.5-5%+ and authorized ROE is set by commissions assuming a lower rate environment, Duke's spread between authorized return and cost of capital narrows toward zero, and per-share value compounds slowly or not at all.

Third threat: hyperscaler bypass. If AWS, Google, Microsoft, and Meta increasingly procure power behind-the-meter from on-site nuclear (SMRs), gas, or dedicated renewables-plus-storage, Duke captures only transmission tariffs on the largest growth wedge and loses the fat margin on incremental load.

The business will exist and resemble itself. Whether it has compounded shareholder value at 8-10% per year (the bull case) or 3-5% per year (the bear case) depends on regulatory durability and entry valuation.

CONFIDENCE: medium