CenterPoint Energy, Inc. (NYSE: CNP) — Houston’s Load Boom at a More Forgiving Price

Report date: September 3, 2026
Market data through: September 2, 2026
All dollars are U.S. dollars unless noted.

⚡ Claude’s Take

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

Verdict: ACCUMULATE / start with a first tranche, medium conviction. The prior report’s accumulation condition has arrived: a defensible entry-and-fair-value zone is approximately $37–$42, or roughly 18–20× estimated 2027 adjusted EPS of about $2.05–$2.10. At $39.51, CNP is finally inside that zone rather than leaning on its upper boundary. This is a quality-compounder-at-a-price setup, not a deep-value trade and not a falling knife.

The call changes from HOLD / accumulate only on weakness in the July 3 report. Since then, the shares have fallen 11.4%, the company’s own-history composite valuation percentile has compressed from the 95th to the 71st, Q2 earnings exceeded the prior year, 2026 guidance held, the ten-year capital plan increased without a higher stated equity need, and Moody’s reportedly removed the negative outlook. A director also bought at $40.70. Those facts make the high-$30s a reasonable place to begin paying for Houston’s superior growth territory.

This is not a full-position call because the market still grants CNP a 13% forward-P/E premium to AEP, Xcel and Duke while the central load claim remains unclassified. Texas ordered a statewide data-center audit after the Q2 call, and ERCOT twice delayed Batch Zero classifications. Meanwhile, first-half simple free cash flow was negative $1.5 billion, contracted forward shares equal about 3.8% of the current count, and the latest credit-supportive hybrid costs at least 6.4%. Near-term price momentum is negative—the stock sits below its 21-, 50- and 200-day averages—but positive five-year compounding and low stock-specific volatility argue against a distress framing.

Conviction and flips: medium. The bullish flip is public evidence that most of the management-estimated 14 GW qualifies with durable minimum-charge and credit protections, while FFO/debt reaches at least 14% without incremental equity. The bearish flip is material load disqualification or regulatory recovery lag that forces EPS growth below 6% while leverage remains near downgrade thresholds.

Tag: the territory is exceptional; the funding toll is not optional.

📈 Stock Price Action — Five-Year Event Map

CNP’s adjusted share price rose from a five-year low of $21.45 in September 2021 to a $44.99 intraday high on July 28, 2026, before closing at $39.51 on September 2. The current price is 12.2% below that high and sits within a trailing-52-week intraday range of $35.79–$44.99. Price moves below are facts; causal attributions are interpretations unless tied directly to a disclosed event. Price data, accessed September 3, 2026.

1–2. CenterPoint’s December 2021 disposal of its Enable Midstream interest completed the strategic pivot toward regulated utilities, plausibly helping CNP outperform the utility index into April 2022; the following decline largely matched the sector as higher rates compressed bond-proxy valuations. CenterPoint release, December 2, 2021.

3–5. The next eighteen months were predominantly a utility-duration cycle: CNP outperformed during the first recovery, matched XLU almost exactly in the 2023 selloff, and then lagged slightly in the rebound. That pattern cautions against attributing every move to company-specific execution.

6. Hurricane Beryl was different. CNP fell 18.1% while XLU rose 5.1%, and the largest one-day decline followed public anger over prolonged outages. The divergence makes the storm response the dominant company-specific explanation, though early-August macro volatility also contributed. CenterPoint Beryl response update, July 30, 2024.

7–8. The subsequent 84.6% advance reflected restored regulatory confidence, resiliency execution and an expanding growth algorithm. The latest retracement began before the Q2 print; CNP gained 0.2% on July 28 and then fell 9.9% through September 2 versus XLU’s 6.3% decline. Broad utility weakness explains most, while Texas’s data-center audit and ERCOT’s classification delay plausibly explain the residual. CenterPoint 2025 investor update, September 29, 2025; ERCOT notice, August 31, 2026.

🔄 Changes Since the July 3, 2026 Report

The prior bull test—sustained 8–9% EPS growth with improving credit—is tracking on earnings and the reported rating outlook, while underlying adjusted credit metrics remain mixed. Its most important load test is not yet passed because the state audit interrupted the expected August classification. The prior bear test—multiple compression from an extreme own-history premium—has partly occurred, but the business delivered through it. That combination, rather than any single earnings beat, explains the changed opening view.

1. Executive Summary

CenterPoint is now one of the cleanest regulated-utility growth stories in the United States. It owns the electricity transmission-and-distribution network across greater Houston, electric and generation assets in southwestern Indiana, and natural-gas distribution systems in Texas, Minnesota and Indiana. Following the completed Louisiana/Mississippi gas disposition and the pending Ohio transaction, essentially all earnings come from regulated monopolies. The economic formula is straightforward: invest approved capital, place it into rate base, earn a regulator-authorized return on the equity layer, and recover operating, depreciation, tax and financing costs from customers.

The franchise moat is real but not conventional. No rival can economically duplicate the local wires or gas network, and no other electric T&D utility operates inside Houston Electric’s certificated territory. Yet customers do not choose CenterPoint’s network, brand does not create pricing power, and regulators determine the return. The moat protects the right to serve; it does not guarantee that every dollar of capital will be accepted, recovered immediately or earn above the cost of funding. CenterPoint’s differentiated advantage is therefore the growth and density of its service territory, especially Houston, plus its ability to maintain regulatory trust and finance a very large construction program.

Operating evidence improved in Q2. Revenue rose 10.7%, operating income 28.1% and diluted EPS 23.2% year over year. Adjusted EPS increased to $0.40 from $0.29, led by regulatory recovery and growth, and management reaffirmed $1.89–$1.91 for 2026. First-half electric net income rose to $377 million from $279 million as rate and transmission recovery outweighed higher depreciation, interest and operating costs. Q2 2026 Form 10-Q, July 28, 2026; earnings release, July 28, 2026.

The growth opportunity is unusually large. CenterPoint raised its 2026–2035 capital plan to $66.7 billion, including $47.5 billion of electric investment and $19.0 billion of gas investment, with more than $10 billion of additional opportunities excluded. It submitted more than 17 GW of large load into ERCOT’s Batch Zero process and estimated about 14 GW would qualify as base or studied load. That would be transformational against Houston’s roughly 21 GW system peak. However, Texas ordered a comprehensive data-center audit on August 3, ERCOT delayed classifications twice, and no public classification was located by the September 2 evidence cutoff. The honest description is management-estimated eligible submissions, not approved or energized load. Q2 presentation, July 28, 2026; Texas Governor directive, August 3, 2026.

The balance sheet is the binding constraint. During 2021–2025, CenterPoint generated $10.3 billion of operating cash flow but spent $21.4 billion on capital projects and another $2.4 billion on dividends. First-half 2026 simple free cash flow was negative $1.5 billion. Debt reached $24.6 billion, contracted forward equity equals about 3.8% of current shares, and a convertible could add as much as another 2.3%. Unadjusted FFO/debt improved while company-adjusted measures slipped slightly; the reported stable outlook is important, but a 6.4%-floor junior subordinated issue shows the price of maintaining credit. This remains a recurring external-capital model.

At $39.51, CNP trades at 20.8 times the midpoint of 2026 adjusted EPS guidance, 2.2 times book value and about 13.0 times mechanically calculated enterprise value to trailing EBITDA. The forward P/E remains about 13% above the median of AEP, Xcel and Duke, but the own-history composite percentile has fallen sharply and the yield has risen to 2.43%. The current price underwrites more than generic utility economics, yet no longer demands that both exceptional execution and an extreme multiple persist.

Verdict. CNP combines a durable legal franchise, unusually strong organic territory growth and credible operating execution. The disconfirming evidence is equally clear: load qualification is pending, returns are regulator-capped, recovery can be delayed, and growth is funded externally. The thesis is strongest when expressed as a per-share regulatory execution story, not as a headline-GW or headline-capex story.

2. Business Overview

The economic machine

CenterPoint is a utility holding company, but the operative businesses are local regulated networks. Houston Electric transmits and distributes electricity; it does not sell the commodity to end customers and owns no conventional generation other than the temporary emergency-generation fleet being removed from rate base. Retail electric providers bill competitive supply, while CenterPoint earns regulated delivery charges. At December 31, 2025, Houston Electric served 2,859,313 metered premises through 67 retail electric providers, and the 10-K states that no other electric T&D utility operates in its certificated territory. 2025 Form 10-K, filed February 19, 2026.

The remaining portfolio combines Houston’s wires network, Indiana Electric, Texas gas, Minnesota gas and Indiana gas. Ohio gas is classified as held for sale and is expected to transfer in the fourth quarter. The company’s own 2026 estimate allocates roughly $33 billion of rate base across the jurisdictions:

These are company estimates, not substitutes for individual regulatory orders. Applying the stated Houston equity ratio and allowed ROE to its rate base produces a simple pre-adjustment allowed-equity-return pool of about $772 million; the comparable figure for combined Indiana gas is about $174 million. Neither is segment earnings: regulatory lag, riders, taxes, depreciation, financing, disallowances and holding-company costs intervene. The arithmetic merely shows why Houston dominates the value proposition. Q2 presentation, July 28, 2026.

How revenue becomes earnings

The recurring revenue base is unusually stable because the networks are essential and territory-exclusive. Approved tariffs recover a return on rate base plus operating costs, depreciation, taxes and financing costs. Fuel and purchased-gas costs are generally passed through rather than a source of commodity margin. Usage and weather can move timing—especially gas volumes—but the long-run earnings driver is invested capital and regulatory recovery, not spot power or gas prices.

This distinction makes consolidated revenue a noisy KPI. Fuel pass-through, weather, regulatory true-ups and divestitures can raise or lower sales without proportionately changing economic profit. From 2021 through 2025, revenue grew only 2.9% annually while operating income grew 11.5%. The better operating dashboard is rate base, customer count, recovery timing, earned versus allowed ROE, financing cost and diluted EPS.

Houston’s model also limits both upside and downside from large loads. A data center directly funds many interconnection modifications; CenterPoint earns from incremental demand charges and systemwide transmission or distribution projects that regulators approve. Management said in the Q1 call that each energized GW could contribute about $6 million of demand charges per month, but that is a management estimate and depends on tariff design, utilization and curtailment rules. CenterPoint avoids the generation-construction and commodity risks that Duke or Xcel may assume, but it also captures less rate base per GW.

Gas ballast and seasonality

Indiana gas served about 785,458 customers at year-end 2025: 669,295 in the north and 116,163 in the south. Purchased-gas costs are pass-through, while weather affects throughput and customer bills. Alternative energy is the principal substitution threat, but no ordinary competitor can lay a duplicate local distribution system economically. The gas networks supply recurrence, infrastructure-replacement spending and jurisdictional diversification; they do not match Houston’s demand runway. 2025 Form 10-K, February 19, 2026.

The portfolio is domestic, asset-heavy and readily understandable. Foreign low-cost labor cannot undermine a local certificated network. Brand matters mainly as a proxy for political trust after storms, not as a source of customer acquisition. There is no foreign currency exposure, ADR structure, master-limited-partnership tax form or K-1 complexity for common shareholders.

Portfolio simplification

The last five years removed most non-core complexity. The midstream exit, the $1.2 billion Louisiana/Mississippi gas disposition completed in March 2025, and the pending $2.62 billion Ohio gas disposition concentrate capital in Houston and Indiana. The Ohio consideration consists of $1.42 billion cash at closing and a $1.20 billion seller note due after 364 days, so not all proceeds are immediately cash. Ohio contributed $52 million of pretax income in the first half and remains in continuing operations because the disposition does not meet the accounting definition of a strategic shift. That detail matters when comparing reported growth before and after closing. Q2 Form 10-Q, July 28, 2026.

Verdict. The business is a simple, durable collection of regulated local monopolies, increasingly concentrated in its best territory. Its earnings are more recurring than consolidated revenue suggests, but the model converts capital into value only after regulatory approval and external financing. Houston improves the growth rate; it does not repeal utility economics.

3. Industry Dynamics

A structurally protected but capped industry

Regulated electric and gas delivery is structurally attractive because entry is almost impossible. The network is a natural monopoly: duplicating wires, pipes, easements, control systems and service crews would be uneconomic, while state commissions grant exclusive territories. Demand is essential and largely domestic. Customer churn affects which retail supplier bills Houston residents, not which grid carries the electricity. These features create very stable asset utilization and low risk of conventional market-share loss.

The same compact that blocks entry caps returns. Regulators approve rate base, capital structure, depreciation lives, recovery mechanisms and allowed ROE. They can defer costs, exclude assets that are not used and useful, demand customer credits or reduce the allowed equity layer. A utility cannot respond to rising demand with unconstrained pricing. Its “pricing power” is the ability to prove costs and secure tariff recovery.

That makes the capital cycle inverted relative to an ordinary competitive industry. More industry investment does not automatically destroy prices because regulators can authorize a return on necessary infrastructure. But more capex creates shareholder value only when five conditions hold: the asset is needed; construction stays near budget; it enters service; regulators admit it into rate base without material lag; and the blended funding cost is below the earned return. CenterPoint’s $66.7 billion plan is therefore an opportunity set, not value by declaration.

Texas: exceptional demand, exceptional scrutiny

ERCOT’s large-load queue illustrates both the opportunity and the danger of headline numbers. In June, ERCOT tracked more than 438 GW of large-load requests, nearly 89% data centers, while warning that not all requests become built projects. The figure later cited by the Governor exceeded 474 GW—more than five times Texas peak demand. A queue larger than the system cannot be treated as forecast consumption. ERCOT release, June 18, 2026; Texas Governor directive, August 3, 2026.

Batch Zero is meant to impose discipline by grouping qualifying projects of at least 75 MW for joint reliability analysis and statewide transmission planning. CenterPoint submitted more than 17 GW and estimated about 10 GW would be classified base load, 4 GW studied load and 3 GW incremental or ineligible. Management projected cumulative energization of roughly 3 GW in 2027, 9 GW in 2028, 11 GW in 2029, 13 GW in 2030 and 14 GW in 2031. It also cited signed Facilities Extension Agreements, end-user commitments and about $900 million of customer cash or security.

Those protections make CNP’s projects plausibly higher-quality than an undifferentiated queue, but they do not equal approval. The 10-Q says actual GW, timing, usage and energization depend on ERCOT approval, construction, regulation and materials, and that customer funds may be returnable in some circumstances. After Governor Abbott ordered project-level verification, ERCOT obtained an exception to the August 7 classification deadline, targeted August 31, and then delayed again for additional validation. ERCOT notices, August 21 and August 31, 2026.

The evidence ladder should remain explicit:

1. More than 17 GW submitted.
2. About 14 GW management-estimated eligible as base or studied load.
3. Conditional and final ERCOT classification.
4. Studied-load allocation and approved transmission plan.
5. Approved facilities and tariffs.
6. Construction and customer completion.
7. Energized, billed demand.

Public evidence clears only the first two stages. The audit is primarily a timing risk until individual projects are rejected, but it can become a volume and capital-plan risk if classification is materially below management’s estimate.

Affordability and curtailment

Large loads can improve affordability by spreading fixed network costs across more billing demand. CenterPoint estimates more than $5 billion of residential and commercial customer savings over a decade and says Texas infrastructure charges rose only slightly above 1% annually from 2014–2025. That is a useful hypothesis, not regulator-validated evidence. Savings depend on large customers paying enough fixed charges through construction, ramp and curtailment. CenterPoint Customer Savings Initiative, August 11, 2026.

The rules remain unfinished. ERCOT scheduled an October 6 workshop on SB6 registration and curtailment, including treatment of co-located large loads and customers with backup generation. If a data center is curtailed during grid stress, the economically decisive question is whether minimum charges still cover shared infrastructure. ERCOT notice, September 2, 2026.

Recovery mechanisms work, but not automatically

Texas offers efficient trackers, yet the first 2026 DCRF is a concrete reminder that speed is conditional. Houston Electric requested $108 million of annual revenue on about $2.2 billion of distribution investment. Settlement reduced the request by $6.2 million and deferred $52.3 million of certain resiliency investment; the July 9 order required that deferral upfront. CenterPoint filed a $101.4 million compliance tariff and then a second DCRF seeking $73 million on about $2.8 billion of capital through May 31. Q2 Form 10-Q, July 28, 2026.

Deferral is not disallowance, but it moves cash and earnings timing. It also demonstrates why political trust matters. The temporary emergency-generation fleet acquired before Beryl became a symbol of poor capital matching: a settlement would remove 15 large units from rates effective May 2025 and five medium units effective January 2026, lowering the current revenue requirement by $112 million if approved. The Q2 results excluded a $19 million after-tax impact associated with removed units. The proposed cleanup is constructive, but the episode shows that monopoly status cannot protect imprudent investment. PUCT Docket 57980, accessed September 2, 2026.

Indiana: slower, steadier and more lagged

Indiana gas provides a different regulatory profile. In June, the north and south utilities requested three-year compliance projects totaling $839 million and seven-year TDSIC plans totaling $1.11 billion. April CSIA updates sought $17 million of annual increases on $148.6 million of rate-base additions, with 80% recovered currently and 20% deferred to the next rate case. CenterPoint expects combined north/south general rate cases by year-end, but no requested ROE or revenue increase was public by the evidence cutoff.

Historical outcomes show affordability pressure. The latest completed northern base case produced a $5.9 million overall reduction, while the southern case allowed a $20.5 million increase but reduced the residential fixed charge. Those cases are not forecasts, yet they show why Indiana should be treated as regulated ballast with tracker-supported reinvestment rather than unconstrained growth. Indiana OUCC case summaries, accessed September 2, 2026.

Capacity constraints

The industry’s competitive bottlenecks are labor, transformers, breakers, poles, rights-of-way, generation adequacy and financing capacity. CenterPoint competes with every large utility and data-center developer for equipment and contractors even though no utility competes for its territory. Transmission planning stretches to 2027, while customers seek power sooner. The mismatch can delay in-service dates and force capital spending before recovery.

Interest rates are another industry price. Utility earnings can grow while shareholder returns disappoint if the cost of debt and equity rises faster than authorized returns. CNP’s 6.4%-floor hybrid is a visible example. The regulatory model passes reasonable financing costs through over time, but lags and capital-structure decisions decide who absorbs the interim burden.

Conventional total-addressable-market analysis is not useful for an exclusive local utility. The relevant market size is certificated rate base, connected customers and coincident load inside the territory. On that basis, Houston’s $18.5 billion estimated rate base, 2.86 million premises and potential load step-up define CNP’s domestic opportunity; expansion outside that footprint would require acquisition or a new certificate rather than ordinary customer wins.

Verdict. Regulated delivery is a structurally good industry for preserving capital and a structurally limited one for earning excess returns. CenterPoint owns a compelling demand territory, but that attracts more scrutiny, construction demand and financing needs. The industry structure removes competitive entry risk; it concentrates regulatory, affordability and execution risk.

4. Competitive Position

Moat type: a government franchise, not customer captivity

CenterPoint’s moat begins with a legal monopoly. A competitor cannot enter Houston, build a parallel grid and take share. Replacement cost, rights-of-way, operational complexity and certification reinforce the statutory barrier. Similar logic applies to local gas pipes. This is stronger than brand, patents or ordinary switching costs because the customer has no network choice.

It is also shallower than a classic excess-return franchise. The PUCT and Indiana regulators set the economic bargain. If CenterPoint lowers costs, customers eventually share the benefit; if it overspends, the regulator can defer or disallow recovery. The allowed ROEs clustered around 9.65–9.80% illustrate the cap. A monopoly protects the denominator—the asset base and customer connection—while regulation limits the numerator—the return earned on equity.

The durable differentiators within that system are territory quality, scale, regulatory execution and cost of capital. Houston offers customer growth, industrial expansion, port and petrochemical demand, electrification and data-center clusters. Density can lower unit operating cost and spreads storm-hardening investment across more demand. The company’s scale supports 24/7 operations, procurement and capital-market access. None of these permits unregulated pricing, but each affects how quickly rate base grows and whether the authorized return becomes realized EPS.

Territory quality versus contract quality

CNP’s 14-GW estimate is larger than the disclosed Xcel and Duke examples cited here, but it does not come with the strongest public contract disclosure. Xcel has described 15-year large-load agreements with minimum bills, exit fees and credit support, plus potential utility-owned generation. CenterPoint’s Houston wires-only model reduces generation and commodity exposure in that territory, and management cites facilities agreements and $900 million of security. Yet it has not publicly disclosed the same project-level term stack, and the state audit delayed the classification on which transmission needs depend.

AEP is the closest listed analog for Texas, transmission and large-load exposure, but it spans eleven states and owns generation. Duke is a high-quality regulated benchmark with nuclear and generation assets and a slower 5–7% EPS algorithm. Xcel differs through its disclosed wildfire exposure, large-load contract structure and equity needs. CenterPoint’s premium can reflect a 7–9% long-duration EPS target and wires-only Houston load leverage, but it also prices meaningful Batch Zero and rate-recovery execution before classification is public.

Financial evidence of the moat

A moat should appear in outcomes. CNP’s operating income increased from $1.36 billion in 2021 to $2.11 billion in 2025 despite portfolio shrinkage, an 11.5% annual rate. First-half 2026 electric operating income rose 30% year over year. Customer connections are stable and rate recovery is recurring. Those results demonstrate protected demand and a growing asset base.

They do not demonstrate unregulated excess returns. Authorized ROEs across the disclosed jurisdictions cluster around 9.65–9.80%, and the consolidated filing does not provide a clean jurisdiction-by-jurisdiction earned-ROE calculation. The result one should demand is consistent earned ROE around the allowed level with per-share growth after dilution, not technology-company ROIC. Q2 presentation, July 28, 2026.

Scale has not eliminated capital intensity. Capex exceeded depreciation by roughly three times in each of the last four years, and free cash flow remained negative. The legal monopoly is therefore visible in earnings stability, not in self-funded cash conversion. If recovery slows or financing costs rise, the moat does not prevent shareholder dilution.

Regulatory trust is an asset—and can be impaired

Hurricane Beryl revealed the least tangible but most important competitive variable. More than ordinary customer satisfaction, restoration performance determines political permission to invest and recover. CenterPoint’s own August 31 tracker reported 82,393 storm-resilient poles or equipment units, 13,727 vegetation-management miles, 818 automation devices and 884 undergrounded miles. That is evidence of work completed, not of better outcomes under a comparable storm. CenterPoint community progress tracker, data through August 31, 2026.

The next severe event is the real test. Independent improvement in SAIDI, SAIFI, customer-minutes interrupted and restoration tails would strengthen regulatory trust. Another operational failure after billions of resiliency spending would damage the growth plan even if the physical network remains monopolistic.

What does not constitute an advantage

Brand does not drive network selection. Retail customers can change power suppliers but not the wires utility. Customer switching cost is therefore statutory and physical, not experiential. Network effects are absent: each additional customer may improve density economics, but it does not make the service more valuable to every existing user in the platform sense. Data and automation help reliability but can be purchased by peers. Management’s growth targets are not a moat. Nor is a large queue, because project requests can be duplicated, speculative or unaffordable.

Verdict. CenterPoint has a durable, wide legal moat and a strong growth territory, but regulation deliberately prevents that moat from becoming uncapped economic power. Its relative advantage is Houston load density and wires-only exposure in that territory; its disconfirming weaknesses are thin credit headroom, a low authorized equity layer and still-incomplete evidence that headline load will become billed demand.

5. Growth History and Forward Opportunities

Historical growth: better in operating income than revenue

Revenue was distorted by fuel pass-through and disposed businesses, while regulated investment drove operating income. Diluted shares rose throughout the period, so enterprise growth did not fully reach each share. That gap is the correct lens for the next decade: rate-base growth can exceed EPS growth because depreciation, interest, regulatory lag and equity issuance absorb part of the increase.

The latest half-year bridge supports the core algorithm. Electric revenue rose to $2.58 billion from $2.26 billion and net income to $377 million from $279 million. Customer rates and design work added $131 million of pretax revenue, transmission mechanisms added $89 million and customer growth $12 million; higher depreciation, interest and O&M absorbed much of the gross benefit. Weather and usage subtracted $16 million. Q2 Form 10-Q, July 28, 2026.

Natural Gas revenue fell to $2.54 billion from $2.60 billion because the Louisiana/Mississippi disposition removed $148 million, while net income still increased to $330 million from $314 million. Portfolio shrinkage therefore makes consolidated revenue growth look weaker than continuing economics.

The $66.7 billion plan

The 2026–2035 plan now totals $66.7 billion: $47.5 billion electric, $19.0 billion gas and about $0.2 billion corporate. Spending for 2026–2030 is $34.2 billion, including roughly $6.8 billion in 2026. At least $10 billion of potential transmission, resiliency, advanced-meter and Indiana large-load projects remains outside the plan. Q2 presentation, July 28, 2026.

The latest $1.2 billion increase allocated about $800 million to Houston grid upgrades for large loads and $400 million to downtown Houston. Importantly, management did not increase the current equity-financing guide. That is constructive only if Ohio proceeds, customer contributions, operating cash flow and debt capacity really absorb the addition. A plan increase without explicit equity is not proof that per-share economics improved; it shifts the proof burden to financing execution.

Large-load optionality

Management’s 14-GW estimate is valuable because it is supported by more than expressions of interest. Signed facilities agreements and customer security distinguish it from the 474-GW statewide headline. Nearly all estimated eligible demand is targeted by 2030, with about 3 GW in 2027. If even a meaningful majority energizes, fixed network costs spread across a much larger billing base, demand charges rise and transmission needs expand.

But the range of outcomes is wide. A base-load classification affects whether a project enters the coordinated study; studied load awaits further allocation; customer construction can lag utility construction; and SB6 curtailment can change billed usage. Interconnection funding protects CenterPoint from some direct stranded costs, but customer security may be refundable. The valuable KPI is not announcements or submitted GW. It is classified GW, approved system investment, contracted minimum charges, construction milestones and billed demand.

Resiliency as both necessity and growth

Houston’s rapid physical-hardening program can improve reliability and expand rate base. The company forecasts 150 million fewer outage minutes by year-end 2026, but installed units are inputs and the forecast is not a measured outcome. A comparable severe storm will decide whether the spending improves restoration enough to rebuild trust. If it does, resiliency capex is unusually high-quality because it serves both customer value and regulatory legitimacy. If it does not, future recovery becomes harder.

Indiana infrastructure and optionality

Indiana gas has nearly $2.0 billion of proposed multi-year compliance and TDSIC work across the two jurisdictions. The 80/20 current-versus-deferred tracker split provides meaningful visibility, though the deferred portion and looming rate cases create lag. Indiana Electric adds generation and large-load optionality distinct from Houston’s wires-only model, but the public growth case is currently less developed than Texas.

Quality of growth

High-quality growth should meet three tests: it is driven by customer or reliability need rather than acquisition; regulators approve it; and after-dilution EPS grows while credit improves. CenterPoint clearly meets the first test. The second is strong but imperfect—DCRF deferral and the temporary-generation settlement are counterexamples. The third is tracking operationally, but the capital structure remains a constraint.

Verdict. CNP has genuinely organic, above-sector growth led by Houston, with unusually large load and resiliency opportunities. The growth is not yet fully de-risked: regulatory classification and financing determine how much enterprise investment becomes per-share value. The right base case is a meaningful fraction of the 14 GW, not all of it and not zero.

6. Financial Quality

Q2 and first-half earnings

Q2 adjusted EPS benefited by $0.10 per share from growth and regulatory recovery and $0.02 from O&M, partly offset by $0.01 weather/usage and $0.01 interest. The bridge matters because it shows investment recovery outpacing the financing burden in the quarter. Interest expense including securitization debt still rose 34% to $261 million, and first-half interest rose 25% to $540 million. Earnings release, July 28, 2026.

GAAP and adjusted results should both be retained. Q2 GAAP net income of $244 million reconciled to $268 million adjusted after offsetting $119 million of equity-security marks against $117 million of indexed-debt marks, plus M&A and temporary-generation adjustments. The ZENS marks mostly offset economically. The $19 million temporary-generation adjustment is different: it is a real consequence of capital that regulators are removing from rates, even if management excludes it from the forward earnings algorithm.

Cash flow: structurally negative by design, not irrelevant

CenterPoint generated $10.334 billion of operating cash flow from 2021–2025 and spent $21.367 billion on capex, a cumulative $11.033 billion shortfall before $2.406 billion of cash dividends. That is not a sign of operating failure: growth utilities routinely invest ahead of recovery. It is nevertheless economically important because debt and equity providers fund the gap. A utility can report rising EPS while shareholder value stalls if new capital is expensive or recovery lags.

The series also rejects a simplistic cash-conversion claim. Operating cash flow was only $22 million in 2021 after a roughly $2.3 billion Winter Storm Uri regulatory-asset build. It reached $3.9 billion in 2023 when CNP and CERC received about $1.1 billion of Texas customer-rate-relief bond proceeds. Those bonds are not corporate obligations, and the proceeds relieved Uri regulatory assets. The swing is regulatory timing, not a change from terrible to exceptional underlying collection. 2025 Form 10-K, February 19, 2026.

First-half 2026 cash from operations rose $90 million, but capex increased $401 million, widening simple free cash flow to negative $1.508 billion. Another $4.088 billion of estimated second-half capex implies about $6.66 billion for the year, close to the rounded $6.8 billion plan. The dividend consumed $302 million despite the cash deficit. The payout is funded by recurring earnings and future recovery, not by current free cash flow.

Balance sheet and credit

At June 30, net PP&E was $35.383 billion and regulatory assets $3.874 billion, up $1.327 billion and $699 million from year-end. GAAP borrowings rose to $24.642 billion from about $22.980 billion, while equity increased to $11.720 billion. Debt divided by debt plus equity was 67.8%, compared with 67.3% at year-end. The credit-agreement definition was 59.2% against a 65% maximum. These figures correct the prior report’s unsupported 92% figure, but 67.8% is still substantial leverage.

Liquidity was about $3.5 billion in July against $4.0 billion of revolving capacity, with about $0.5 billion drawn. Near-term liquidity is adequate. The issue is the recurring volume and price of capital. First-half financing included $3.441 billion of long-term debt and loan proceeds, $1.271 billion of repayments and $165 million of net common issuance. The company also issued restoration securitization bonds, 4.85% Houston Electric bonds, a 2.875% convertible and an SOFR-linked term loan. Q2 Form 10-Q and presentation, July 28, 2026.

The post-quarter $700 million junior subordinated note carries a 6.4% coupon through 2033 and then resets at the five-year Treasury yield plus 188.5 basis points, subject to a 6.4% floor. Interest can be deferred for up to ten years, but common dividends and repurchases are generally restricted during deferral. Rating agencies may give the security equity content, improving credit metrics, yet the economic cost is plainly higher than traditional utility debt. Form 8-K, filed July 31, 2026.

The unadjusted Moody’s-method FFO/debt comparison improved to 13.2% trailing from 12.5% in 2025. On the company’s one-time-adjusted view, however, the Moody’s series declined to 13.4% from 13.8%; the adjusted S&P series edged down to 13.0% from 13.1%, even though the unadjusted S&P comparison improved to 13.0% from 12.4%. Separately, Moody’s reportedly changed CNP and Houston Electric outlooks to stable from negative, affirmed Baa2/Baa1 and upgraded CERC senior unsecured debt to A2. No public Moody’s release or later CNP filing was located, so the rating action remains secondary-sourced. The reported outlook change lowers immediate downgrade pressure, but the long-term 14–15% target is not yet achieved. Investing.com report of Moody’s action, August 20, 2026.

Dilution and per-share quality

CNP issued 4.481 million shares in May for about $165 million by physically settling prior forwards. Separate forwards cover 24.865 million shares at an initial $36.26 and would provide about $910 million if physically settled by February 2027, equal to roughly 3.8% of June shares. The $650 million convertible due 2029 has a $53.61 conversion price and permits up to about 15.155 million shares, another 2.3% at maximum conversion. Outstanding shares already increased 0.9% year to date to 658.701 million.

These instruments do not make the EPS plan impossible; management’s guidance should already reflect expected settlement. They do explain why roughly 11% planned rate-base growth translates to 7–9% EPS growth. A $1 billion ATM remained unused at quarter-end, but the option exists if capex expands or credit weakens.

Accounting quality, obligations and latent exposures

No restatement, auditor qualification or material weakness appeared across the five-year annual/quarterly filing review. The November 2025 10-Q amendment merely added jointly reporting subsidiaries omitted from the original, without changing financial statements. That supports control quality but does not prove conservative accounting. Regulatory accounting is appropriate for approved cost recovery, but the $3.874 billion regulatory-asset balance necessarily embeds the assumption that future customers will pay; its 22% six-month increase makes aging, approved recovery and disallowance the relevant tests. Conversely, the exclusive territorial franchise is the principal valuable asset not separately recognized on the balance sheet: accounting records the network, not the economic value of the certificate protecting it from entry.

Undiscounted minimum purchase obligations total $6.401 billion: $3.968 billion natural-gas supply, $2.000 billion electric supply, principally long-term PPAs, and $433 million equipment and IT. Some costs are pass-through. CNP also retained about $426 million of measurable maximum exposure on legacy Energy Systems Group guarantees plus uncapped guarantees that cannot be estimated; management views a material payment as remote and records no liability. If all ZENS were exchanged at quarter-end, roughly $964 million of deferred taxes could be payable before offsets, although management expects carryforwards to absorb most cash tax.

The franchise and regulated assets make total loss remote under ordinary conditions. A catastrophic combination—major storm liabilities, systematic disallowance, a downgrade-driven collateral call, failed asset sales and inaccessible capital markets—could destroy substantial equity value, but the essential networks and cost-recovery compact make literal zero unlikely absent insolvency or extraordinary state intervention.

Returns on capital

Authorized ROE near 10% is consistent with allowed-return utility economics, but the consolidated filing does not calculate a clean earned ROE for each jurisdiction. The more informative test is whether actual jurisdictional returns remain near authorization after interest and dilution while customer affordability remains acceptable. Q2’s earnings bridge was directionally supportive because regulatory recovery outweighed higher financing and depreciation costs; it did not, by itself, prove full earned-ROE parity. The five-year simple-FCF record shows that the test must be passed repeatedly with external funding.

Verdict. Earnings quality improved and the Q2 bridge is credible, controls appear sound, and liquidity is adequate. Financial quality is capped by structural negative free cash flow, rising debt, contracted dilution and limited credit headroom. Scale improves operating income; it does not yet improve self-funding. The balance sheet, not current operations, remains the weak link.

7. Capital Allocation

Reinvesting into the best territory

The strongest feature of CenterPoint’s capital allocation is strategic coherence. Management exited midstream, sold slower-growth gas jurisdictions and concentrated investment in Houston and Indiana. The Louisiana/Mississippi disposition closed for $1.2 billion in March 2025. The pending Ohio disposition at $2.62 billion is particularly attractive as a recycling transaction because consideration is roughly 1.9 times its 2024 rate base and the proceeds fund higher-growth regulated assets. The caveat is timing: $1.20 billion arrives through a 364-day seller note rather than closing cash, and the disposed business’s earnings remain in continuing operations until closing.

The $66.7 billion plan is primarily organic and regulated, which is preferable to acquisition-led growth. It includes $47.5 billion in electric networks and $19.0 billion in gas systems. Posey Solar, acquired and placed in service for $357 million in 2025, subsequently obtained Indiana regulatory recovery. That is a reasonable example of acquiring an asset only when the regulatory path is visible. By contrast, the temporary emergency-generation fleet demonstrates how even safety-motivated spending can destroy value if assets are operationally mismatched and removed from rates.

The capital-allocation scorecard should therefore distinguish gross investment from admitted rate base. The ideal project has documented customer need, customer contribution where appropriate, a tracker or settled rate mechanism, short construction duration and financing below the allowed return. A speculative queue-driven transmission project with uncertain classification scores much lower even if its nominal capex is larger.

Financing the plan

Management expects roughly $4 billion of common equity financing over 2026–2035, including about $1.1 billion represented by forward sales and roughly $3 billion in later years. The Q2 plan increase did not change that guide. That restraint is positive, but there are already meaningful claims on the share count. The 24.865 million forward shares are contracted, and the convertible can dilute further if the price exceeds the conversion terms.

Debt issuance is similarly continuous. The 6.4% junior security improves rating-agency equity credit but imposes a high hurdle: after tax, its cost still consumes a large portion of a roughly 9.65% allowed equity return before operating and regulatory risk. Traditional subsidiary debt is cheaper because it is closer to rate base and more directly recoverable. Holding-company hybrid capital protects the rating at a real economic price.

The practical capital-allocation rule is that a plan increase is attractive only if incremental return exceeds incremental funding cost on a per-share basis. Management’s target of 100–150 basis points above downgrade thresholds and 14–15% FFO/debt is sensible. Until achieved, balance-sheet repair competes with every additional growth project.

Dividend and repurchases

The quarterly dividend increased from $0.23 to $0.24, a 4.3% sequential increase and consistent with a roughly 6% annual growth target. At current guidance, the annualized $0.96 represents a 50.5% payout. That leaves accounting earnings for reinvestment, but not enough cash to cover capex. The dividend is economically supported by expected future regulatory recovery and capital access.

No material discretionary repurchase program or activity was identified in the reviewed filings, and there is no economic repurchase case while simple free cash flow is deeply negative and shares are issued to fund growth. Repurchasing equity while issuing debt or forwards would be circular. The appropriate priority is maintaining investment-grade credit, funding approved projects and growing the dividend no faster than sustainable per-share earnings.

Incentives and governance

CEO Jason Wells received $12.1 million of 2025 compensation. The annual incentive weighted adjusted EPS at 70% and paid 159% overall after EPS reached the maximum goal. Customer metrics including frequent interruption and voice-of-customer earned zero. The current long-term program consists of 70% performance shares—35% relative TSR and 35% cumulative adjusted EPS—plus 30% time-based restricted shares. Carbon goals belong to legacy awards, correcting the prior report’s description. 2026 proxy, filed March 4, 2026.

The design aligns management with earnings and relative shareholder returns, but it underweights capital efficiency. The relative-TSR tranche also carries a minimum 75% payout when CNP’s P/E ranks in the peer-group top quartile. That can reward multiple expansion independently of operating value creation. Reliability and customer metrics exist, but their lower weight is uncomfortable after Beryl. A better design would give more weight to earned ROE versus allowed ROE, FFO/debt, outage-duration improvement and per-share growth net of issuance.

Shareholders strongly supported say-on-pay, but about 39% opposed the officer-exculpation amendment and roughly 25% opposed director Elinor Cloonan. The CEO has also served as chair since October 2025, balanced by an independent lead director. These are governance warnings rather than evidence of capital misallocation. Annual meeting Form 8-K, filed April 17, 2026.

Insider behavior

A complete two-year Form 4 review found 3,700 shares purchased for $141,473, all by director Laurie Fitch: 2,700 shares in May 2025 and 1,000 at $40.70 on August 17, 2026. Discretionary code-S dispositions totaled 28,839 shares for about $1.0 million; code-F withholding transactions were tax payments, not market dispositions. Directors and executives collectively own only about 0.19% of shares. Fitch Form 4, filed August 18, 2026.

The repeated director purchases are directionally positive but too small to carry the thesis. More meaningful evidence would be several senior executives purchasing with personal capital while the company issues shares.

Verdict. Capital allocation is strategically sensible: simplify the portfolio, recycle mature assets and invest organically in higher-growth regulated territories. The weaknesses are funding dependence, costly hybrid capital, earnings-heavy incentives and one vivid example of poorly matched emergency-generation spending. Management has chosen the right destination; regulators and capital markets determine the realized return on the journey.

8. Changes and Headwinds — Last Two Years

From Beryl crisis to accelerated hardening

Hurricane Beryl in July 2024 was the defining negative event. Prolonged outages triggered political scrutiny, damaged customer trust and exposed weak communications and vegetation management. The stock’s 18% decline while utilities rose showed that investors viewed the problem as company-specific. CenterPoint responded with accelerated hardening, automation, vegetation work and undergrounding, and the physical installation metrics have advanced materially.

The improvement is operationally plausible but not yet proven against a comparable storm. Reliability investment produces value only when outage frequency, duration and restoration tails improve. This is the most important non-financial falsification test over the next two hurricane seasons.

Portfolio simplification completed another leg

The March 2025 Louisiana/Mississippi disposition removed lower-growth gas assets and provided $1.2 billion. The Ohio agreement extends the process, reducing geographic diversification but increasing capital available for Houston and Indiana. Concentration makes the equity story easier to understand and increases sensitivity to Texas politics and weather. The portfolio is higher growth, not lower risk in every dimension.

Growth targets expanded

The September 2025 capital update introduced a record plan and 7–9% adjusted EPS growth through 2035. The July 2026 update added $1.2 billion and moved the large-load claim into ERCOT’s formal Batch Zero process. Q2 execution supported the near-term algorithm, with adjusted EPS up 38% and full-year guidance intact.

The qualitative change since July is that the load pipeline became more structured but also more externally scrutinized. Customer agreements and security improved evidence quality. The Governor’s audit and ERCOT delays made clear that management cannot control classification. Both developments should be incorporated; choosing only one produces either promotional optimism or reflexive skepticism.

Regulatory repair and continuing friction

The proposed removal of temporary emergency generators from rate base is a necessary trust repair and lowers customer charges. It also caused a $19 million after-tax Q2 adjustment and may reduce the revenue requirement by $112 million. As of September 2, the PUCT docket still lacked a final order because of clarification over rate-case-expense recovery. CenterPoint elected to defer those expenses to a future proceeding. PUCT filing, September 1, 2026.

The DCRF deferral is smaller in earnings terms but analytically important. It demonstrates that tracker recovery is not a blank check after Beryl. Regulatory treatment remains constructive enough to support investment, while affordability oversight can change timing and create exclusions.

Rating outlook improved, adjusted metrics did not

The reported Moody’s outlook change removes an overhang from the prior report. At the same time, debt increased and the latest subordinated capital carries a 6.4% floor. Those facts are not contradictory: rating support often requires expensive equity-like financing. Investors should treat the outlook as proof of manageable credit, not proof of abundant balance-sheet capacity.

No hidden accounting or corporate event

Only two material 8-Ks appeared after the July 3 cutoff: Q2 earnings and the junior-subordinated issuance. The five-year filing review found no new acquisition, leadership change, restatement or material weakness through September 2. The thesis changed because of price, operating execution, financing and Texas policy, not because of an undisclosed corporate restructuring.

Verdict. The last two years strengthened CNP’s long-term opportunity and improved strategic focus, but raised execution stakes. Beryl accelerated valuable investment while reducing political margin for error; Batch Zero expanded potential demand while placing it under a stricter test. The net change is positive for long-term earnings capacity and mixed for risk.

9. Risk Analysis

Correlated downside matters more than isolated downside

The most dangerous scenario is not one delayed project. It is a feedback loop: load classification falls short, planned capital produces less near-term revenue, equity needs increase, a credit metric weakens, financing costs rise and regulators become more sensitive to affordability. Because CNP begins with negative free cash flow and a valuation premium, the same event can reduce both earnings expectations and the multiple.

Storm risk creates a second feedback loop. A severe hurricane causes restoration expense and capital needs, while poor performance damages regulatory trust and delays recovery. Securitization can ultimately recover extraordinary costs, but political and cash timing remain material. The Beryl price reaction shows that essential-service status does not eliminate equity drawdowns.

What mitigates the risks

The networks are essential, geographically diverse across electric and gas operations, and supported by established recovery mechanisms. Customer security protects some interconnection spending; Ohio proceeds add financing capacity; liquidity is about $3.5 billion; and the reported stable rating outlook reduces immediate downgrade risk. Many costs can be deferred or securitized rather than permanently lost.

These mitigants reduce insolvency risk more than valuation risk. An outcome can be manageable for creditors and still produce weak per-share returns if dilution rises or the premium compresses.

Verdict. The risk profile is moderate for the operating franchise and high for the translation of enterprise growth into per-share value. No single disclosed liability threatens the company’s existence. The central downside is a compound of load timing, regulatory lag and external financing that lowers EPS growth while compressing the premium.

10. Valuation Discussion (Embedded Expectations)

Current valuation bridge

At $39.51 and 658.720 million legally outstanding shares, common equity value is about $26.03 billion. Adding June 30 carrying-value debt of $24.642 billion and subtracting June 30 unrestricted cash of $49 million produces an approximate $50.62 billion enterprise value using September 2 market capitalization plus quarter-end net debt. It is not a same-date balance-sheet measure; the later hybrid initially adds both debt and cash, while subsequent cash use is unknown. Restricted cash is not netted. The pending Ohio consideration should not be subtracted before closing and without also removing Ohio earnings.

The 13.0-times figure uses approximately $3.901 billion of trailing EBITDA, reconstructed from trailing operating income plus depreciation and amortization in the 2025 Form 10-K and Q2 2026 Form 10-Q. Even so, EV/EBITDA is not a clean takeover multiple for a regulated utility. Depreciation represents real replacement needs, securitization debt corresponds to pass-through assets, and utility value depends on jurisdiction-specific rate base and returns. P/E, price-to-book, earned ROE and funding needs are more informative.

Own-history and peer context

The own-history composite, P/E, P/B and P/S percentiles are 71.3%, 71.3%, 55.3% and 87.3%, respectively. In July, the comparable sequence was 95th, 94th, 91st and 99.9th. The premium is no longer extreme, but neither is the stock statistically inexpensive. P/S is particularly elevated, though revenue is the weakest utility valuation denominator. Valuation history, data through September 2, 2026.

The three-peer median is 18.4 times, leaving CNP at a 12.9% premium. Some premium is rational: CNP targets faster growth than Duke and owns a wires-only Houston model with exceptional demand density. Yet AEP offers a similar growth range at a lower multiple, and Xcel has disclosed stronger minimum-bill and exit protections. The premium already recognizes meaningful success.

Peer guidance comes from AEP’s July 30 update, Xcel’s July 30 earnings release and Duke’s August 4 earnings release. September 2 closing prices come from dated adjusted-price datasets for AEP, Xcel and Duke, accessed September 3.

Scenario framework: what the current price requires

The scenarios below express annualized total-return math, not future dollar values. They isolate the interaction among EPS growth, valuation, dividends, dilution and credit.

Bear mechanics. Rate-base growth slows to 8–9%, earned ROE falls into the high-8%/low-9% range, Batch Zero projects slip, and DCRF/Indiana lag forces more equity. Five percent EPS growth plus roughly a 2.5% dividend contribution is largely consumed by annualized multiple compression from 20.8 to 17 times.

Base mechanics. Rate base grows near 10–11%, management delivers the midpoint of 7–9% EPS growth, a meaningful but not complete portion of Houston load qualifies, and credit improves gradually. The multiple normalizes to 19 times, so earnings and dividends create a moderate return rather than an exceptional one.

Bull mechanics. Most management-estimated load survives the audit, projects energize close to schedule, rate-base growth reaches the low double digits, earned ROE stays near authorization and funding occurs without issuance beyond the committed and planned amounts. Ten percent EPS growth and a maintained premium support a low-teens return.

Reverse-engineering market expectations

If the forward multiple converges to the 18.4-times peer median over three years, CNP needs roughly 10% EPS growth plus the dividend to approach an 8% annualized return. If 20.8 times persists, approximately 5% EPS growth plus the dividend can reach the same neighborhood. The current valuation therefore underwrites either near-top-end growth with ordinary multiple compression or merely adequate growth with a durable premium.

The market is correctly recognizing the exclusive franchise, Houston territory, strong Q2 and long runway. It may be underestimating how quickly classified loads improve affordability and transmission need. Conversely, it may be overestimating the certainty of 14 GW, the speed of tracker recovery and the ability to fund $66.7 billion without more dilution.

Book value supplies a useful cross-check. At 2.22 times book and roughly 10% ROE, the simple earnings yield on book is only about 4.5% before growth. The premium works because retained earnings and new equity fund a rapidly expanding rate base, not because the existing book earns an unusually high return. That reinforces the importance of growth duration and funding cost.

Verdict. Valuation has improved from extreme to demanding-but-defensible. The price no longer requires a historically maximal multiple, but the 13% peer premium still assumes superior growth and competent financing. Expected returns are most sensitive to Batch Zero classification and the multiple, not to one quarter of earnings.

11. Variant Perception

Market-implied belief

The market-implied positive view is easy to state: CNP is a direct public Houston electricity-demand vehicle, its Houston wires-only model limits generation risk in that territory, a $66.7 billion regulated plan supports 7–9% EPS growth through 2035, and large customers can lower bills for everyone else. Q2 execution and the reported stable credit outlook reinforce that view.

The market’s 20.8-times multiple confirms that CNP is not treated as an average utility. Yet the decline from the June high and collapse in own-history percentile show that confidence is no longer one-way. The stock has become a debate about evidence quality and funding rather than a simple load-growth momentum trade.

Analyst questions on the two latest calls concentrated on exactly the right fault lines: how much of the load was approved rather than merely committed, how customer charges and transmission economics work per GW, whether the larger capital plan changes equity needs, and how Indiana load and generation affect the plan. Management’s answers supplied useful facilities-agreement, security and demand-charge detail, but the subsequent state audit overtook its expected August classification date. Q1 and Q2 call transcripts, April 23 and July 28, 2026.

Strongest positive case

The best positive argument is not that all 14 GW energizes. It is that CenterPoint needs only a meaningful portion. Signed facilities agreements and $900 million of cash/security make its queue higher-quality than statewide headlines. Because Houston Electric owns wires rather than generation, direct customer contributions limit stranded capital. New billing demand spreads fixed costs, improves affordability and creates second-order transmission investment. If credit reaches 14–15% FFO/debt, the company can fund the plan with less equity than skeptics expect.

The additional evidence would be public classification of most base load, durable minimum-charge terms through curtailment, approved transmission projects and actual 2027 energization. Independent storm-performance improvement would strengthen regulatory trust at the same time.

Strongest negative case

The strongest negative argument is capital arithmetic. The company spent twice its operating cash flow on capex over five years, owes dividends, carries $24.6 billion of debt and has contracted dilution. A 6.4% hybrid is costly against a sub-10% authorized equity return. If large-load timing slips, capital may be spent before billing demand arrives; if regulators protect affordability, revenue recovery can lag; if credit weakens, more equity follows. The business can remain healthy while per-share returns disappoint.

The negative case also notes that a 13% peer premium leaves little room for simultaneous disappointment. The own-history percentile is lower but still above average, and the dividend yield is modest for a utility. A decline in EPS growth toward 5% could be compounded by a multiple reversion toward the high teens.

Factor positioning and what the tape prices

Three- and six-month total-price returns are negative 4.3% and 9.2%, while twelve-month performance remains positive 6.5% and five-year annualized return is 11.7%. The stock sits below short-, medium- and long-term moving averages, but its five-year maximum drawdown is only 22.8% and stock-specific volatility is about 10.1%. That is broken near-term momentum, not distressed price behavior.

The factor model identifies CNP primarily as a utility and low-volatility dividend exposure: Utilities loading +0.918, Low Volatility +0.309 and a utility-dividend basket +0.290, with negative Growth and Quality loadings in the broad model. Its 0.643 R-squared means common factors explain much, not all, of variation. The utilities factor was weak over the last month and quarter, which supports the conclusion that macro de-rating explains most of the recent decline. FactorsToday CNP loadings, accessed September 3, 2026.

The variant is therefore not “the market missed data-center demand.” The market sees it. The differentiated view is that submitted load has better contractual evidence than a generic queue but worse public protection than some peers, and that customer-cost spreading may be more valuable than direct capex. The second variant is that the funding toll can consume much of the enterprise growth even when the demand thesis is right.

Assumptions that matter most

1. A meaningful majority of management-estimated eligible load survives audit and classification.
2. Minimum charges and customer security cover shared infrastructure through delay or curtailment.
3. Earned ROE stays close to authorization despite affordability pressure.
4. FFO/debt reaches at least 14% without equity materially above the stated plan.
5. Resiliency spending produces measurable storm-performance improvement.

Verdict. The market is appropriately skeptical after an extreme premium, but still prices superior execution. The best analytical edge is definitional: separate submitted, classified, built and billed GW; separate rate-base growth from after-dilution EPS; and separate creditor safety from shareholder return. Momentum offers no independent reason to dismiss or embrace the fundamentals.

12. Fact vs. Interpretation Table

13. Open Questions

1. How much of the estimated 14 GW receives conditional and final base-load status, and how much remains studied or ineligible?
2. What minimum-bill, exit-fee, credit-support and curtailment terms apply to each major load, and can customer funds be returned after utility spending begins?
3. How much incremental transmission and distribution rate base follows from approved load, rather than customer-funded interconnection work?
4. Will the October SB6 process preserve fixed-cost contributions when large users are curtailed?
5. Can FFO/debt reach 14–15% while the company spends about $6.7 billion in 2026 and settles forward equity?
6. What revenue increase, ROE and equity layer will CenterPoint request in the combined Indiana gas cases, and how much of the $1.95 billion infrastructure proposal is approved?
7. Do the resiliency projects reduce outage duration and restoration tails in a comparable severe storm?
8. What is the final accounting and cash use of Ohio proceeds, including the seller note and foregone earnings?
9. Does the next incentive design increase the weight on reliability, customer outcomes and capital efficiency?

14. What Must Be True

Positive case requirements

The positive thesis does not require perfect load conversion. It requires enough qualified demand to validate system investment, spread customer costs and support the upper half of the EPS range without balance-sheet deterioration. The most important proof point is the combination of classified GW and financing, not either in isolation.

Negative case requirements

The negative thesis is falsified if CNP converts classification into billed load, sustains near-top-end EPS growth, strengthens credit and limits dilution. A merely healthy utility is not enough to disprove it; the issue is whether superior enterprise growth reaches each share.

Overall analytical verdict. CenterPoint’s franchise and territory are proven. Its large-load magnitude, recovery timing and funding efficiency remain partly prospective. The evidence supports a better risk-reward balance than in July, but the company must still prove that record capital deployment produces record per-share value rather than record external financing.

15. Public Source Appendix

Primary company and SEC sources

• CenterPoint Energy 2025 Form 10-K, SEC/CenterPoint Energy, filed February 19, 2026, annual regulatory filing: document.
• CenterPoint Energy Q2 2026 Form 10-Q, SEC/CenterPoint Energy, filed July 28, 2026, quarterly regulatory filing: document.
• Q2 2026 earnings release, CenterPoint Energy/SEC, July 28, 2026, Form 8-K Exhibit 99.1: document.
• Q2 2026 investor presentation, CenterPoint Energy/SEC, July 28, 2026, Form 8-K Exhibit 99.2: document.
• Q1 2026 earnings-call transcript, CenterPoint Energy IR, April 23, 2026, company transcript: document.
• Q2 2026 earnings-call transcript, CenterPoint Energy IR, July 28, 2026, company transcript: document.
• Junior subordinated notes Form 8-K, SEC/CenterPoint Energy, filed July 31, 2026, financing filing: document.
• 2026 definitive proxy statement, SEC/CenterPoint Energy, filed March 4, 2026, governance filing: document.
• 2026 annual-meeting results, SEC/CenterPoint Energy, filed April 17, 2026, Form 8-K: document.
• Laurie Lee Fitch Form 4, SEC, filed August 18, 2026, insider-transaction filing: document.
• Laurie Lee Fitch Form 4, SEC, filed May 13, 2025, insider-transaction filing: document.
• Customer Savings Initiative, CenterPoint Energy, August 11, 2026, company release: document.
• Community Progress Tracker, CenterPoint Energy, data through August 31, 2026, operating dashboard: document.
• Dividend declaration, CenterPoint Energy, July 16, 2026, company release: document.
• Enable Midstream exit release, CenterPoint Energy, December 2, 2021, historical company release: document.
• Hurricane Beryl response update, CenterPoint Energy, July 30, 2024, historical company release: document.
• 2025 investor update, CenterPoint Energy/SEC, September 29, 2025, Form 8-K exhibit: document.

Regulators and public authorities

• PUCT approves ERCOT Batch Zero framework, ERCOT, June 18, 2026, grid-operator release: document.
• Governor Abbott directs comprehensive data-center audit, Office of the Texas Governor, August 3, 2026, government directive: document.
• Batch Zero market notices M-A080326-01, -02 and -03, ERCOT, August 3, 21 and 31, 2026, grid-operator notices: notice 1, notice 2, notice 3.
• SB6 workshop notice M-A090226-01, ERCOT, September 2, 2026, grid-operator notice: document.
• PUCT Docket 57980, Public Utility Commission of Texas, filings through September 2, 2026, regulatory docket: docket.
• Vectren North and South gas rate summaries, Indiana Office of Utility Consumer Counselor, accessed September 2, 2026, regulatory case summaries: North, South.

Peer and third-party cross-checks

• AEP Q2 2026 earnings update, American Electric Power, July 30, 2026, company release: document.
• Xcel Energy Q2 2026 earnings release, SEC/Xcel Energy, July 30, 2026, company filing exhibit: document.
• Duke Energy Q2 2026 earnings release, SEC/Duke Energy, August 4, 2026, company filing exhibit: document.
• CNP and peer adjusted price histories, AZI Trading, data through September 2, 2026, third-party market data: CNP, XLU, AEP, Xcel, Duke.
• CNP valuation history, AZI Trading, data through September 2, 2026, third-party valuation dataset: dataset.
• CNP factor loadings and performance statistics, FactorsToday, data through September 2, 2026, third-party statistical model: loadings, performance, specific volatility, factor returns.
• Moody’s action on CenterPoint ratings, Investing.com reporting Moody’s Investors Service, August 20, 2026, secondary financial news: article.

Third-party market, valuation and factor data are cross-checks; SEC filings, company releases and regulator documents control where figures conflict. Historical source dates are retained when necessary to explain the five-year price cycle.