Overview
Two utilities sit at a similar market-size level - roughly $97 billion each - yet they present sharply different ways to play the AI power story. Constellation Energy (CEG) is positioned as a merchant nuclear generator that can extract scarcity and premium pricing from hyperscale cloud providers. Duke Energy (DUK) is a regulated franchise that is adding data-center load to a rate-based model, trading higher predictability for more modest upside.
Quick comparative snapshot
| Metric | CEG | DUK |
|---|---|---|
| Price | $267.10 | $124.30 |
| Market Cap | $97.74B | $96.89B |
| Revenue Growth (LTM) | +23.4% | +7.2% |
| P/E (LTM) | 25.3x | 19.2x |
| P/E (Fwd) | 23.2x | 18.5x |
| V/EBITDA | 14.8x | 11.5x |
| Analyst Upside | +26.4% | +11.8% |
| Fair Value Upside (Pro) | -8.1% | -13.9% |
| Dividend Yield | 0.6% | 3.5% |
| FCF Yield | 1.2% | -3.4% |
| Debt/Equity | 65.4% | 179.1% |
| Beta | 1.12 | 0.37 |
| 1Y Return | -22.5% | +3.6% |
| EPS Proj (FY+1) | $11.78 | $6.72 |
CEG - Merchant Nuclear with a Pricing Advantage
Constellation’s core strength is structural: it operates the largest nuclear fleet in the United States, delivering continuous, carbon-free baseload power that matches the needs of hyperscale AI data centers. The company has established a template for premium long-term contracts through a 20-year power purchase agreement with Meta for the Clinton nuclear plant, covering 1,121 MW. That model supports the notion that nuclear can command pricing well above short-term spot markets when hyperscalers seek guaranteed 24/7 clean power.
Recent strategic moves are expanding Constellation’s scale and revenue base. The acquisition of Calpine increases capacity and provides greater financial flexibility, contributing to a projected FY+1 revenue figure of $33.79 billion, up from $25.53 billion. Additional structural supports include production tax credits for nuclear output, which act as a price floor and limit downside exposure for uncontracted generation.
Street sentiment toward CEG is bullish in aggregate. BMO Capital kept an Outperform rating and a $376 price target, down from $390 due to PJM regulatory considerations but still implying substantial upside versus the then-current share price of $267.10.
That said, risks remain. CEG returned -22.5% over the prior 12 months, reflecting valuation mean-reversion concerns. Free cash flow yield is a modest 1.2%, leaving limited buffer for investors who prioritize cash generation. The company also faces exposure to wholesale power price volatility in regions such as PJM for its uncontracted volume. Separately, regulatory reviews - notably the Texas Governor’s audit that covers 474 GW of data center interconnection requests - add friction to the broader momentum behind AI-related power demand.
DUK - Regulated Franchise with a Data-Center Tolling Angle
Duke’s exposure to AI comes through a different channel: growth in contracted electricity service to data centers that connect to its regulated grid. Duke recently reported a Q2 2026 earnings-per-share beat with $1.43 versus a $1.32 consensus, an outperformance of about 8.3%, and it reaffirmed guidance for 5-7% annual earnings growth through 2030. These results underscore the utility’s message that new AI-driven demand can be absorbed within its rate base while producing regulated returns.
DUK’s pipeline includes 7.8 GW of electric service agreements for data centers, with 5.2 GW currently under construction. The company has implemented a "Customer Protection Plus" framework requiring data-center customers to fund connection costs upfront; that mechanism protects existing residential and commercial ratepayers from bearing those one-time expenses and strengthens balance-sheet discipline during the buildout.
Some sell-side firms have moved to a more favorable stance. KeyBanc upgraded Duke to Overweight with a $139 target, while BTIG raised its target to $141 and labeled Duke as an attractive defensive utility name. Those assessments reflect the view that Duke offers steady regulated earnings and a 3.5% dividend yield that can cushion downside for income-focused investors.
However, Duke’s financial structure reflects heavy capital intensity. The company is deploying a $103 billion capital plan and carries a high debt-to-equity ratio of 179.1%. Free cash flow yield is negative at -3.4%, indicating cash outflows related to ongoing capex. Regulatory outcomes can also be constraining: a North Carolina rate settlement with Duke Energy Carolinas was set at $496 million, approximately half of what had been requested, signaling pressure on the company’s allowed returns in some jurisdictions.
Which profile fits which investor?
The comparison comes down to risk and return preferences. For investors prioritizing asymmetric upside tied directly to the AI electricity theme, Constellation’s combination of fast revenue growth (+23.4% LTM), substantial analyst upside (+26.4% consensus), and a premium PPA template with hyperscalers represents the more compelling pure-play option.
By contrast, Duke is tailored to investors seeking predictable income and lower market volatility. Its 3.5% dividend yield, low beta of 0.37 and regulated earnings construct make it a suitable vehicle for gaining exposure to incremental AI demand without taking merchant price risk. That predictability comes with tradeoffs: more modest analyst upside (+11.8% consensus), substantial ongoing capex, and a negative free cash flow yield during the expansion phase.
Insights
- Large-scale, continuous nuclear generation can command a scarcity premium for 24/7 clean power that intermittent renewables cannot replicate.
- Both companies trade above their InvestingPro fair value measures (CEG: -8.1%, DUK: -13.9%), suggesting some of the AI thesis has already been priced in.
- Duke’s negative FCF yield implies that its dividend is effectively being supported by capital markets while the utility executes its large capital program; this dynamic warrants monitoring through the 2027-2028 period.
- Regulatory processes such as the Texas ERCOT audit of 474 GW of data center interconnection requests are a sector-level headwind that could slow permitting and connections for data-center-driven demand.
Bottom line
Investors must choose between two distinct exposures to AI-driven electricity demand: Constellation for merchant, high-growth, premium-priced nuclear power and Duke for regulated, dividend-oriented, lower-volatility exposure attached to a large data-center pipeline. Both paths are credible, but they serve different strategic objectives within an investment portfolio.