Constellation Energy at $240: Nuclear Power Meets the AI Boom

For most of the last decade, owning a nuclear reactor was a liability. Aging plants, flat power demand, subsidies that came and went, and a market that treated a merchant power producer as a leveraged bet on the spot price of electricity. Then something changed. Artificial intelligence started eating electricity faster than the grid could build it, and the one thing an AI data center needs above all else, carbon-free power that runs 24 hours a day, 365 days a year, turned out to be exactly what a nuclear fleet produces. Constellation Energy owns the largest one in the United States.

That is the whole story of why this stock went from a sleepy utility spin-off to a market darling, and why it now trades at about $239.71 (as of July 8, 2026), well off its 52-week high of $412.70 but still carrying a growth multiple no independent power producer wore five years ago. Twenty-one analysts sit at a mean target of $357.81, roughly 33% above the current price, a spread that tells you the sell-side is pricing in a lot of the AI-power thesis that hasn’t shown up in the reported numbers yet.

So here’s the question I want to answer. CEG is not Visa or Microsoft, businesses whose cash flows you can plug into a clean discounted-cash-flow model. It’s a capital-heavy, commodity-exposed independent power producer whose free cash flow was negative three of the last four years and whose return on invested capital swings with the price of power. You cannot value it on free cash flow, and I’ll show you why the 83x EV/FCF multiple is a mirage. What you can ask is this: how much of the AI-power premium is already baked into the ~18x forward earnings multiple, and is the uplift from data-center power deals durable, or just a commodity re-rate that fades? Every figure below is computed from CEG’s own statements. This is educational analysis, not advice.

#First, classify the business — because it changes everything

Before I touch a multiple, I have to be honest about what CEG is, because getting that wrong is how you value it wrong. This is an independent power producer (IPP): it owns generation, mostly nuclear, and sells the electricity into wholesale markets and increasingly under long-term contracts. That classification carries three consequences that reshape the entire analysis.

One: free cash flow is the wrong lens. IPPs carry heavy, lumpy generation capex, and they post billions in collateral against their hedging books, which swings working capital violently from year to year. CEG’s free cash flow ran -$4.0B, -$7.7B, -$5.0B, then +$1.3B across 2022 to 2025. Feed that into a reverse-DCF and it spits out nonsense: the 83x EV/FCF multiple on the payload is not a signal that the stock is expensive, it’s a signal that the denominator is broken. So I’ll value this on earnings and EBITDA, not cash flow.

Two: the returns are cyclical, not compounding. A wide-moat compounder earns a steady, high return on capital. An IPP’s return rides the power-price cycle. CEG’s ROIC went 1.2% → 5.2% → 16.7% → 8.7%. That is not moat erosion and it is not a broken business; it’s what a commodity producer’s returns look like. I’ll narrate it that way and refuse to dress it up as something it isn’t.

Three: the moat, such as it is, lives in the asset, not the economics. You cannot build a new nuclear fleet. Licensing, siting, and capital make new reactors close to impossible in a reasonable timeframe. That scarcity, colliding with surging AI power demand, is the real edge. It shows up as pricing power on long-term contracts, not as a stable ROIC line.

Hold those three facts. They govern every chart that follows.

#The returns swing with the power cycle — read the ROIC line honestly

Return on invested capital is the first number I check on any business, because it answers whether a company creates or destroys value per dollar it puts to work. For a wide-moat compounder, you want to see a high, flat line sitting comfortably above the cost of capital, year after year. That’s the fingerprint of a durable advantage.

CEG’s line does not look like that, and it shouldn’t. It swings. In 2022, a brutal year for power margins and a big working-capital drag, ROIC was 1.2%, well below the ~7.7% WACC, meaning the business destroyed value that year. Then 5.2% in 2023, still short of the hurdle. Then a spike to 16.7% in 2024 as power prices and generation margins ran hot. Then back down to 8.7% in 2025, just barely above the cost of capital.

I want to be precise about what that pattern means. It is not evidence of a deteriorating moat, and anyone who reads the dip that way is misapplying a compounder framework to a commodity producer. The swings track power prices, mark-to-market movements on the hedge book, and a steadily growing invested-capital base (the denominator keeps rising as CEG invests in its plants). The honest read: this is a cyclical business whose returns crossed above its cost of capital only recently, and only in the good part of the cycle.

i

×

A cyclical line, not a compounding one

This is return on invested capital — after-tax operating profit divided by the capital in the business — against CEG's ~7.7% cost of capital (the dashed line). Anything above the line is value created; anything below is value destroyed. Notice the shape: 1.2% in 2022 (well below the line), climbing to a 16.7% peak in 2024, then falling back to 8.7% in 2025. It only clears the hurdle in the good years.

For a moat business you'd want this line high and flat. For an independent power producer, this jagged shape is exactly what you should expect and exactly what you must not misread. The swings come from power prices, mark-to-market movements on CEG's hedging book, and a rising invested-capital base. This is commodity cyclicality, not competitive decline.

The practical takeaway for valuation: because ROIC is cyclical, you cannot lean on a single year's number. The 16.7% of 2024 flatters the business; the 1.2% of 2022 understates it. The right question is where returns sit through a full cycle — and whether the AI-power contracts can lift and stabilize that average above the cost of capital, which is the whole bull case.

#The margins swing too, and the cash finally turned positive

Here’s the chart that explains why the free-cash-flow lens fails on this company. On the left axis is operating margin, which swung from a razor-thin 2.0% in 2022 to a fat 18.5% in 2024 and back to 12.1% in 2025. On the right axis is free cash flow, and this is the striking part: it was deeply negative for three consecutive years, -$4.0B, -$7.7B, -$5.0B, before finally turning positive at +$1.3B in 2025.

Why was cash flow so negative while the company was reporting profits? Two reasons, both structural to an IPP. First, generation capex is heavy and constant, running $1.7B to $2.9B a year to maintain and upgrade the fleet. Second, and larger, an IPP posts collateral against its hedging positions, and when power prices move, those margin calls swing working capital by billions. In 2023, the change in working capital and operating outflows drove operating cash flow to negative $5.3B even as the company earned $1.6B in net income. That gap is not a business failing to generate cash; it’s the mechanics of a commodity hedger.

i

×

Why free cash flow is the wrong lens here

The green line (left axis) is operating margin; the blue line (right axis) is free cash flow. Both lines are volatile because both are driven by the price of power. The margin line more than tripled from 2022 to 2024, then gave some back in 2025 — that's the commodity cycle, not an operational story. The free-cash-flow line spent three years deeply negative before crossing into positive territory at +$1.3B in 2025.

The negative FCF years look alarming until you understand the mechanics. Independent power producers run large capex and, crucially, post billions in collateral against their hedging books. When power prices move, those margin calls whipsaw working capital, which is why operating cash flow can go sharply negative in a year the company reports a healthy accounting profit. It's a cash-timing artifact of hedging, not a solvency problem.

This is the concrete reason the 83x EV/FCF multiple on the data feed is meaningless and why I refuse to run a free-cash-flow DCF on CEG. You'd be dividing enterprise value by a denominator that is essentially random from year to year. The company has to be valued on earnings and EBITDA, which smooth over the hedge-collateral noise, plus a judgment on the durability of the power-demand catalyst.

#The earnings are real, but they’re volatile

If free cash flow is unusable, earnings are the better lens, but you have to respect how much they move. Revenue is actually the stable part of the story: $24.4B, $24.9B, $23.6B, $25.5B, essentially flat across four years. It’s the bottom line that whipsaws. Net income ran -$0.16B → $1.62B → $3.75B → $2.32B. That 2024 peak of $3.75B was a banner year for power margins; the 2025 figure of $2.32B is more of a mid-cycle read.

The 2022 loss deserves a word. A negative $160M net income on $24.4B of revenue looks scary in isolation, but it was largely driven by unrealized mark-to-market losses on the hedging book during an extraordinarily volatile energy year (the year of the European energy crisis). Those are accounting swings that reverse; they’re not the operating business bleeding cash. Still, the honest summary is that CEG’s earnings are genuinely volatile, and that volatility is exactly why the market historically slapped a low multiple on IPPs. The AI-power thesis is, at its core, a bet that long-term contracts will smooth and lift that jagged earnings line.

i

×

Flat revenue, volatile profit — the commodity signature

The tall bars are revenue, which barely moves: roughly $24-25.5B every year. The short bars are net income, which is all over the place — a small loss in 2022, then $1.6B, a $3.7B peak in 2024, and $2.3B in 2025. When revenue is stable but profit swings this hard, you're looking at a business whose costs and realized prices move with a commodity.

The 2022 loss was driven largely by unrealized mark-to-market losses on the hedging book during the energy-crisis year — an accounting swing that reverses, not the operating fleet losing money. The 2024 peak, conversely, flatters the business with unusually strong power margins. Neither extreme is the true run-rate.

For valuation, this means you should mentally normalize toward a mid-cycle earnings number rather than anchor on the best or worst year. The bull case is precisely that long-term data-center contracts convert this jagged bar chart into something smoother and higher — replacing merchant-price volatility with contracted revenue. Until that shows up in the numbers, the volatility is a reason for caution, not a reason the thesis is wrong.

#The balance sheet: leverage is normal for what this is

A power producer that owns tens of gigawatts of long-lived generation is going to carry real debt, and CEG does. The question is whether it’s a problem, and the answer here is a clear no. Net debt of roughly $21.6B against EBITDA gives a net-debt/EBITDA of 2.7x, and operating income covers interest about 6 times over. The red-flag screen does register leverage above the 2.5x line in one of five years, but for a utility-adjacent generator this sits squarely in the normal band. There’s no refinancing cliff visible in the data and no solvency concern.

i

×

Debt that funds the fleet, not debt that threatens it

These bars track balance-sheet debt against total cash & equivalents. Debt rose from ~$5.8B in 2022 to ~$9.0B in 2025 as CEG invested in its generation fleet, while cash built to $3.6B. Two definitional notes, because the leverage screen uses different figures than these bars. On the debt side, the net-debt metric (~$21.6B, 2.7x EBITDA) uses a broader definition that captures leases and other obligations (~$22.5B gross), so it runs well above the ~$9.0B senior-debt bar. On the cash side, the screen nets against only ~$0.9B of unrestricted cash rather than the $3.6B total shown here, because much of the balance is restricted or collateral-related; that is why net debt (~$22.5B − ~$0.9B ≈ $21.6B) is higher than a naive $9.0B − $3.6B would suggest.

For a capital-intensive power producer, 2.7x net-debt/EBITDA with ~6x interest coverage is ordinary — this is roughly where you'd expect a large generator or regulated-adjacent utility to sit. It is neither a fortress balance sheet (like Visa's 0.3x) nor a distressed one. The leverage is the cost of owning a lot of physical generation.

The relevant risk isn't default — it's that heavy capital needs (potential plant restarts, uprates, new contracts) require ongoing financing, and rising rates raise that cost. But nothing on the balance sheet undermines the equity thesis. Treat the leverage as a neutral structural fact, not a red flag.

#Capital allocation: reinvestment first, and the screen flags to note

Judging management by how they deploy cash is harder here than usual, because the free-cash-flow base has been negative in aggregate over the analyzed period, so the tidy “percent to dividends, percent to buybacks” breakdown the screen produces for a cash-generative compounder simply doesn’t compute (the totals are meaningless when cumulative FCF is negative). What I can say from the data: capital has gone overwhelmingly into the fleet, capex ran $1.7B to $2.9B a year, and CEG has been a modest net repurchaser of shares (net stock issuance was slightly negative, about -1.6% of sales in the latest year, meaning buybacks modestly exceeded issuance).

i

×

A reinvestment-first profile, with red flags that fit an IPP

This is an illustrative split of where CEG's capital energy has gone — dominated by reinvestment into the generation fleet, with debt service, modest net buybacks, and a small dividend. It is directional, not precise: because cumulative free cash flow over the five-year window was negative, a clean percent-of-FCF allocation table (the kind you'd build for Visa or Coca-Cola) can't be computed, and I won't pretend otherwise.

The red-flag screen raises three items: ROIC below 10% in three of five years, negative free cash flow in three of five years, and leverage above 2.5x in one of five years. Read through the IPP lens, none of these is a governance failure — they are the mechanical fingerprints of a capital-heavy, commodity-exposed, hedge-collateral-posting business. The negative-FCF and low-ROIC flags are the same cyclicality shown in the earlier charts.

The thing to actually watch on capital allocation is discipline as the AI-power capex cycle ramps: potential reactor restarts and uprates are large, long-dated bets. The bull case needs those investments to earn well above the cost of capital under contract; the bear case is that management chases growth capex whose returns don't justify the spend. That's the judgment call, and it's too early to score.

#The dividend is real but tiny, and it’s telling you something

CEG does pay a dividend, but it’s a small, young one, so treat it as a signal rather than a reason to own the stock. The company only started paying after its February 2022 spin-off from Exelon, and the payout has grown fast off a low base: dividend per share went $0.56 → $1.13 → $1.41 → $1.55 across 2022 to 2025, a compound growth rate near 40% that reflects a new payer ramping up, not a mature grower you’d extrapolate. The current forward rate of $1.71 yields just 0.69%, on a payout ratio of only about 14% of earnings. There’s no long “consecutive-increase streak” to cite here, because the dividend simply hasn’t existed long enough, and I won’t invent one.

The more interesting number is the yield trend. Even as CEG kept raising the dividend, the yield fell from about 0.65% in 2022 to roughly 0.44% by 2025, because the share price re-rated far faster than the payout grew. That’s the AI-power re-rate showing up in a single ratio: the market repriced the equity so aggressively that a rising dividend still shrank as a percentage of price. For an income investor, CEG is not the story. For everyone else, the shrinking yield is just another way of seeing how much growth expectation is now embedded in the stock.

i

×

A small dividend, and a yield that shrank while the stock ran

The left axis is dividend per share (the rising bars-as-line), the right axis is dividend yield. CEG began paying only after its 2022 Exelon spin-off, so the payout is young: it climbed from $0.56 to $1.55 per share in four years, a ~40% growth rate that reflects a new payer scaling up from near zero, not a durable long-term rate you should project forward.

Watch the yield line, though. Despite steady dividend hikes, the yield fell from ~0.65% to ~0.44% over the same window. That happens only when the share price rises faster than the dividend — which is exactly what the AI-power re-rate did. The market repriced the equity so hard that a growing dividend still became a smaller slice of the price.

The practical read: the dividend is not the reason to own CEG, and I deliberately de-emphasize it. At a ~0.69% forward yield and a ~14% payout ratio, there is ample room to keep raising it, but the income is immaterial to the total-return case. The stock lives or dies on the growth thesis, not the cash it hands back.

#The price: how much AI-power premium is already in the multiple?

This is the entire debate. Nobody argues the fleet isn’t valuable. The question is whether ~18x forward earnings already pays for the AI-power story, or leaves room.

Start with what the multiple is. Trailing P/E is 21.4x, forward P/E is 17.6x, EV/EBITDA is 13.5x on the last twelve months and 12.3x on next-year consensus. For historical context, independent power producers used to trade in the low teens on earnings and high-single-digits on EBITDA, valued as leveraged commodity plays. CEG’s premium to that old range is the AI-power thesis, expressed in one number.

i

×

A growth multiple on a business that used to trade like a commodity

These bars compare trailing (LTM) and forward (NTM) multiples. The forward numbers are lower than trailing because analysts expect earnings to grow — consensus next-year revenue is ~$35.6B versus $25.5B reported, a ~39.5% step up. That is a very large jump for a business whose revenue has been flat (~1.5% five-year CAGR), and it is worth being skeptical about: a big share of it is a consensus assumption about power prices and merchant volumes, not signed, contracted revenue. The genuinely contracted piece — deals like the Crane/Three Mile Island restart supplying Microsoft, plus capacity-auction payments — is real but is a fraction of a $10B step-up. As earnings rise, a flat price mechanically compresses the multiple, but if the consensus power-price assumption proves optimistic, the forward multiple is understated and the stock is dearer than it looks.

The key context is history: IPPs traditionally traded around 10-13x earnings and high-single-digit EV/EBITDA, because the market treated their profits as unreliable commodity flows. CEG at ~17.6x forward earnings carries a clear premium to that legacy range. That premium is the market capitalizing the AI-power demand story — the belief that data-center contracts will make these earnings more durable and faster-growing than a normal IPP's.

So the valuation question reduces to a single judgment: is that premium earned? If long-term hyperscaler power deals genuinely convert CEG into a contracted-growth business, ~18x is reasonable and the Street's $358 target makes sense. If the demand proves cyclical or the contracts underwhelm, the multiple re-rates back toward IPP norms and the stock has meaningful downside. The bars can't tell you which; the durability of the demand does.

Now the fair value, and here I have to be explicit about method, because the data feed offers two numbers I’m going to reject outright. The reverse-DCF implies ~17% FCF growth and the illustrative DCF spits out a $77 fair value — both built on that broken free-cash-flow denominator. I’m setting them aside entirely, exactly as the classification demands. A free-cash-flow DCF on a company with three negative-FCF years is a category error, like valuing an insurer on EV/EBITDA.

Instead, anchor on earnings and EBITDA, and on the analyst mean. Consensus forward EPS is about $13.58. Put a range of multiples on it: a defensive 15x gets you to ~$204 (roughly where a de-rated IPP would sit), a middle 18x gets ~$244 (about the current price), a growth-crediting 22x gets ~$299, and a full AI-premium 26x gets ~$353, which is essentially where the 21-analyst consensus of $357.81 lands. In other words, the Street is applying about a 26x forward multiple, pricing the AI-power thesis close to fully.

But before you accept that forward EPS, look at what it assumes. Consensus next-year revenue is about $35.6B against $25.5B reported, a 39.5% jump in a business whose revenue grew roughly 1.5% a year over the last five. That is not the base I’d extrapolate blindly. A ~$10B step-up is far larger than the contracted demand on the books today: the signed pieces (the Crane Clean Energy Center supplying Microsoft, other long-term power-purchase agreements, and capacity-auction payments) are real and rising, but they are a fraction of $10B. The bulk of the consensus jump rests on an assumption about merchant power prices and volumes staying elevated, which is exactly the cyclical variable that historically capped IPP multiples. So the honest reconciliation is this: my ~18x mid-case ties the multiple to contracted demand plus a modest merchant tailwind, not to the full consensus revenue leap. If the contracted book scales into that $35.6B, the growth-crediting and full-premium cases activate. If it doesn’t, and merchant prices normalize, forward EPS comes in below $13.58 and the whole multiple ladder shifts down.

Intrinsic value · the verdict

Blended fair value

$290

Current price

$240

Margin of safety

17%

i

×

What you're looking at & why it matters

Each dot is a fair-value-per-share estimate from a different method; the dashed vertical line is today's market price. Dots to the right imply upside on that method; dots to the left mean the price is already ahead of it.

The methods disagree on purpose — a multiple built on a depressed-cash-flow year lands low, a growth DCF lands high. The honest answer is the range between them, not any single point.

Why it matters: seeing every method at once, against the price, keeps the valuation honest — it exposes how much of the outcome rides on the growth assumption instead of hiding behind one false-precise target.

Fair values from multiple methods (see above). Source: Yahoo Finance + earnings-multiple framework · as of July 2026. Illustrative, assumption-driven, not a price target.

The messy middle. I won’t pretend to a precision the business doesn’t allow. The de-rated case ($204) is what CEG is worth if the AI-power story fizzles and it re-rates back to a commodity IPP. The full-premium case ($353-358) is what it’s worth if hyperscaler contracts durably transform the earnings profile. Reality is a probability-weighted blend, and I land on an intrinsic range of roughly $260 to $320, with a midpoint around $290. That’s above today’s $240 but well below the Street’s $358, because I’m discounting the AI-power premium more heavily than consensus does, on the view that the durability is still unproven.

Two margins of safety. Against my ~$260-320 range, $240 sits at roughly a 7-25% discount — there’s some cushion, but not a screaming bargain. Against the number an investor actually watches, the $357.81 consensus, the price is about 33% below, which is a large gap that says either the Street is early or the market is skeptical. I lean toward skeptical: consensus looks to be pricing the AI-power thesis nearer to certainty than the numbers yet justify.

Which reduces the whole thing to one sentence. Is the AI-driven power-demand uplift durable enough to justify a growth premium on a business that has always been a cyclical commodity producer, when the Street is already sitting ~50% above the spot price? Clear that, contracts get signed, power prices and capacity payments stay elevated, nuclear support holds, and the earnings line smooths upward, and the premium is earned. Fail it, and CEG re-rates back toward the low-teens multiple that IPPs have always worn, and the ~$204 case is the floor. Everything else is detail.

#Tailwinds and headwinds

A valuation is only as good as the demand story behind the growth. Here’s what pushes CEG above the mid-case and what pulls it back to the commodity floor.

Tailwinds — what earns the premium:

• AI data-center power demand is the structural driver. The explosion in AI compute has turned always-on, carbon-free electricity into a scarce, contracted commodity. Hyperscalers are signing long-term power-purchase agreements to lock in supply, and CEG’s fleet is a prime source. The landmark example is the deal to restart Three Mile Island Unit 1 as the “Crane Clean Energy Center” under a long-term agreement with Microsoft, the kind of contract that, if replicated, converts merchant-price exposure into contracted, higher-margin, more durable revenue. This is the single most important variable in the thesis.
• Rising power prices and capacity payments. Tight supply-demand in the grid regions CEG serves has pushed up both energy prices and the capacity payments generators earn for being available. A structurally tighter grid lifts the earning power of the entire fleet, not just the contracted portion.
• A nuclear fleet you cannot replicate. New nuclear is extraordinarily hard to build, and existing reactors are being valued as strategic scarce assets rather than aging liabilities. That scarcity is real pricing power, and it’s the asset-level moat behind the multiple.
• Policy support for nuclear. Federal support for existing nuclear generation, notably the Inflation Reduction Act’s nuclear production tax credit (PTC), puts a price floor under the fleet’s economics and cushions weaker power-price years. The mechanism is designed to top revenue up to roughly a $43/MWh floor (the figure widely cited for the IRA nuclear PTC, though I was unable to re-verify the exact rate and phase-out schedule against primary IRS/Treasury guidance for this piece, so treat the number as indicative and confirm before relying on it). The durability of that support past its current window is a genuine tailwind worth tracking.

Headwinds — what pulls it back to the commodity floor:

• It’s still a commodity business. Strip away the contracted portion and CEG’s earnings ride power prices and mark-to-market swings, as the volatile ROIC and net-income charts show. If the AI-demand narrative cools or power prices normalize, the earnings line reverts to its jagged self and the growth premium unwinds toward the historical low-teens IPP multiple. That’s the ~$204 downside case.
• The premium may already be priced. The Street’s $358 target implies a ~26x forward multiple, which prices the AI-power thesis close to fully. If deals are slower or smaller than hoped, there’s more room to disappoint than to surprise, even if the direction is right.
• Execution and financing risk on growth capex. Reactor restarts, uprates, and new contracts are large, long-dated, capital-heavy bets. They require ongoing financing in a higher-rate environment, and the returns won’t be visible for years. Capital-allocation discipline here is unproven for this specific growth cycle.
• Counterparty and single-partner concentration. The thesis-defining contract, the Crane Clean Energy Center restart, is a ~20-year, ~835 MW deal tied to a single hyperscaler, Microsoft. That is roughly 2-3% of CEG’s ~32 GW fleet committed to one counterparty, so the direct volume concentration is modest today. The real concentration risk is narrative, not tonnage: the AI-power premium leans heavily on a small number of marquee deals with a handful of deep-pocketed but capex-cyclical hyperscalers (Microsoft, Amazon, Google, Meta). If one pulls back on data-center buildout, or a flagship deal slips or is repriced, the market’s willingness to pay 26x for “contracted transformation” can reverse well ahead of any actual revenue hit. Track how many distinct counterparties and how many gigawatts get put under contract, not just the headline names.
• Regulatory and political overhang. Nuclear economics depend partly on policy (the production tax credit and related support). A shift in the political winds, or in how data-center power deals are regulated at the grid level, could compress the very economics the bull case relies on.

The bull and bear cases share one fulcrum: whether contracted AI-power demand can turn a cyclical commodity producer into a contracted-growth business. While the deals come and power stays tight, the discount to the Street is a gift. The day the demand narrative cracks, the premium unwinds toward the low-teens multiple IPPs have always worn.

A note on sourcing: I was again unable to verify Constellation Energy’s most recent quarterly earnings-call transcript word-for-word at the time of writing (the transcript source was unavailable), so I’ve deliberately kept the forward section to well-established, structural drivers of the business, the AI-power demand story, the Microsoft/Three Mile Island restart, rising capacity payments, and nuclear policy support, rather than citing specific quarterly figures, guidance, or executive quotes I couldn’t confirm. Constellation’s next report is scheduled for August 6, 2026 (Constellation Energy Investor Relations); that’s where to verify contracted-volume, capacity-payment, and guidance trends against these assumptions.

#So — great asset, fair price?

The asset is exceptional and genuinely hard to replicate: the largest carbon-free, always-on fleet in the country, arriving at the exact moment AI made that kind of power scarce. That’s not in dispute. What is in dispute is whether an independent power producer, a business whose returns and cash flows have always swung with a commodity, deserves a durable growth multiple. The market has decided it does, and priced CEG about 50% above where it trades, on the bet that data-center contracts rewrite the earnings profile.

I’m more cautious than that. At $240 you’re paying a modest discount to my ~$260-320 intrinsic range and a large discount to the Street’s $358, but the gap between those two anchors is precisely the uncertainty: how much of the AI-power premium is real and durable versus a commodity re-rate that fades. The stock isn’t expensive on earnings; it’s conditional on a thesis that hasn’t fully shown up in the numbers.

What would change my mind? Two things I’d track every quarter. First, contracted volume: are the long-term hyperscaler power deals actually being signed, at scale, at prices that lift and smooth the earnings line? If the Crane restart is the first of many, the premium is being earned in real time. Second, the ROIC and margin lines: if returns start holding above the cost of capital through a softer power-price year instead of collapsing back toward the hurdle, that’s the signal the business has structurally changed. Until those two lines confirm it, I treat the AI-power premium as a promising bet rather than a fact, and the ~$204 commodity floor as the risk that keeps me honest.

---

What’s your read on Constellation here — is the AI-power demand structural enough to justify a permanent growth premium on a nuclear IPP, or is the Street’s $358 target pricing a transformation that’s still mostly a promise? I’d genuinely like to hear the bear case against these numbers. Find me on X or LinkedIn.

Keep reading: for the same quality-versus-price lens applied to a capital-light business at the opposite end of the spectrum, see my breakdown of Visa as the toll booth on $15 trillion in payments, and the rest of the Finance category works through more companies the same way.

Data sourced from Yahoo Finance via my local analysis tooling, as of July 8, 2026, computed from Constellation Energy’s reported financial statements. Because CEG is a capital-heavy, commodity-exposed independent power producer with negative free cash flow in three of the last four years, I deliberately reject the free-cash-flow reverse-DCF and DCF the tooling generates (they rest on a broken denominator) and value the company on forward earnings and EBITDA multiples plus a judgment on the durability of the AI-power demand catalyst. I was unable to verify Constellation’s latest earnings-call transcript word-for-word at the time of writing, so the forward section is limited to well-established structural drivers rather than unconfirmed quarterly figures or executive quotes. Figures are illustrative and may lag or contain gaps; always cross-check against Constellation’s 10-K/10-Q filings on SEC EDGAR before acting.