CMS Energy

Phase A — Understand the business

Lens 1 · Company Overview

Business model — a regulated monopoly compounder. CMS Energy Corporation (Michigan-incorporated, HQ Jackson MI; NYSE: CMS) is a holding company. It operates in three segments: Consumers electric utility, Consumers gas utility, and NorthStar Clean Energy (non-utility IPP + marketing). Consumers is the engine — a vertically integrated electric and gas utility with a "mix of primarily residential, commercial, and diversified industrial customers". NorthStar is "primarily a domestic independent power producer and marketer" with 1,665 MW of generating capacity.

How money is actually made. Consumers invests in regulated assets (generation, distribution, gas infrastructure), the MPSC sets rates that allow recovery of those costs plus an authorized ROE, and the spread compounds. 2025 segment net income to common: electric utility $719M, gas utility $409M, NorthStar $71M, corporate interest & other −$138M = $1,061M total. So the regulated utilities are ~107% of net income and corporate carry is a ~13% drag — i.e. essentially the entire equity value is the Consumers rate base.

Customers / suppliers / competitors. End customers: ~2M Michigan homes and businesses. Key contract structure on the demand side is now the large-load tariff (MPSC-approved Nov 2025): for customers ≥100 MW (data centers), a minimum 15-year contract, an 80% minimum demand-billing obligation, upfront fees, and collateral/exit-fee protections. That is the take-or-pay structure that de-risks the data-center growth. Competitors: in a regulated monopoly there are no direct retail competitors except the capped Retail Open Access (ROA) program (10% of sales, fewer than 300 customers actually using it); the relevant "competition" is other utilities competing for the same data-center developers (a named risk factor) and capital-markets competition for yield.

customers.csv is empty in the research layer — the customer-concentration data here is sourced from the 10-K narrative ``.

Lens 2 · Supply Chain

A utility's "supply chain" is its fuel/capacity stack upstream and its captive ratepayers downstream. Named stakeholders:

• Upstream fuel/capacity: natural gas suppliers and pipelines (city-gate contracts); coal supply for the soon-retiring J.H. Campbell plant; purchased power. 2025 fuel costs: fuel for electric generation $657M, purchased & interchange power $1,706M, purchased power–related parties $94M, cost of gas sold $809M. Purchased power is the single biggest cost line and rising (+28% YoY), reflecting the coal-to-PPA transition.
• Owned generation (the chokepoint being deliberately dismantled): D.E. Karn coal units (515 MW) retired 2023; J.H. Campbell coal (1,407 MW) approved for retirement but held open by U.S. DOE Section 202(c) emergency orders through Feb 17, 2026. Replacement capacity: Covert Generating Station (1,200 MW, gas, bought 2023); a new 10-year PPA with the MCV Partnership for up to 1,240 MW (effective June 2030); 850 MW of contracted battery storage (online through 2028).
• Renewable build: up to 9,000 MW solar + 4,000 MW wind (owned + PPA) under the Renewable Energy Plan; Muskegon Solar (250 MW) online 2025. March-2026 IRP preview adds >13 GW of new renewables/storage + two new gas plants ~1,500 MW in Bay & Genesee Counties.
• Downstream: ~2M captive Michigan ratepayers + the incoming data-center hyperscalers (undisclosed names; one ≥1 GW developer contracted, load ramping from 2030).
• Single-source dependency / chokepoint: MISO interconnection queue and the FERC tariff. The named operational chokepoint is interconnection delay at MISO for new generation/storage — explicitly flagged as a risk to the entire build plan. The other chokepoint is regulatory, not physical: every dollar of this chain only earns a return if the MPSC grants recovery.

supply-chain.md commercial-layer file is missing — chain mapped from filings ``.

Lens 3 · Competitive Advantages (moats)

Moat = regulated monopoly + constructive jurisdiction + a rare demand tailwind. This is the strongest moat archetype in equities (a legal monopoly) but with a capped upside (the regulator sets the return).

1. Legal monopoly / franchise. Consumers serves its territory under municipal franchises; there is no retail competitor at scale (ROA capped at 10%, barely used). Switching cost for a customer is effectively infinite (you can't choose another grid).
2. Constructive regulatory relationship (the real variable). Authorized ROEs of 9.90% (electric) and 9.80% (gas) and a track record of timely rate orders. Michigan's forward-test-year, file-and-implement framework is investor-friendly. This is the moat that can erode — see Lens 13.
3. The data-center flywheel — the differentiated 2026 edge. Unlike most utilities where load growth raises bills, here each incremental GW of large load reduces the average customer's 5-yr rate CAGR by ~2% while adding $2-5B of capital-investment opportunity. Affordability is the political constraint on rate-base growth; data-center load relaxes that constraint. That is a genuine, hard-to-replicate competitive advantage for this specific territory (Michigan land, water, fiber, and an under-loved grid).
4. Bargaining power. Over customers: total (monopoly). Over suppliers: moderate — exposed to commodity, equipment, and tariff inflation, but the PSCR/GCR mechanisms pass prudent fuel costs through to ratepayers, so commodity risk is largely socialized.

positioning.md / bottlenecks.md missing — moat assessed from filings + web.

Lens 4 · Segments

Segment net income to common (the 10-K reports segments by NI, not standalone revenue/EBITDA — every figure ``):

Geography: 100% Michigan — there is no geographic diversification (a concentration risk, Lens 13). Deliveries: electric 37.4B kWh (2025) vs 36.8B (2024); gas 311 Bcf vs 268 Bcf. Gas volume jumped on weather; the structural electric-delivery story is the 5-yr weather-normalized increase guided by management on data-center load. The acceleration is in gas (weather-driven, non-recurring) and in the corporate drag (rising rates) — the structural growth lever is electric rate base, which compounds at 8%+.

segments.csv empty — segment figures from the 10-K narrative tables ``.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print: Q1 2026, reported 2026-04-28)

The headline. Q1-2026 net income to common $338M, diluted EPS $1.10 vs $302M / $1.01 in Q1-2025 (+12% NI, +9% EPS). Adjusted EPS $1.13 vs $1.10 consensus — a beat. Revenue $2.73B vs ~$2.55B expected.

Drivers. Electric + gas rate increases, offset partially by higher service-restoration costs and increased D&A + property taxes from higher capex. This is the textbook utility pattern: rate relief in, depreciation/interest/storm costs out, net positive.

Guidance & tone. Management reaffirmed 2026 adjusted EPS guidance of $3.83-$3.90 and long-term adjusted EPS growth of 6-8%, "with continued confidence toward the high end". Tone unchanged/constructive. Notably the guidance was reaffirmed again on June 3 alongside the CFO transition — management going out of its way to decouple the org change from the numbers.

FY2025 full-year (the base). Operating revenue $8,539M (+13.6% YoY, weather + rate-driven), operating income $1,727M (+16.1%), net income to common $1,061M, diluted EPS $3.53 vs $3.33 (2024) and $3.01 (2023). Three-year EPS CAGR ≈ 8.3% — squarely inside the 6-8% guide.

Balance-sheet flags (mild, see Lens 10). 2025 operating cash flow $2,235M (down from $2,370M on working-capital + lower tax-credit proceeds); capex (cash-flow basis, ex-finance-lease) $3,824M; net cash used investing $4,038M. Free cash flow is structurally negative — this is an external-capital-funded grower. Receivables grew (+$251M) outpacing the revenue mix; inventories roughly flat.

Market reaction. Stock was roughly flat-to-down ("Why Is CMS Energy Down 0.7% Since Last Earnings Report") despite the beat — i.e. the good quarter was priced in; the stock has been driven by macro (rates) and the CFO news, not the print.

Lens 6 · Earnings Calls (sentiment trend)

transcripts/ is empty in the research layer, so sentiment is reconstructed from press releases, Q1 call coverage, and 8-Ks ``.

• Consistent management refrain (last 3-4 prints): "reaffirm guidance," "confidence toward the high end," "all-of-the-above" supply, and increasingly, the data-center pipeline as the headline growth narrative. The Feb-2026 (FY25) release was titled "Exceeds Earnings Guidance in 2025, Raises 2026 Adjusted EPS Guidance"; Q1-2026 "Reaffirms".
• What got louder: data centers / large load (now quantified at ~9 GW pipeline with staged 1.5-2 GW qualified, 4-5 GW advanced, 1-2 GW final). The "all-of-the-above" + two new gas plants language (March 2026) is new — a tonal shift toward more generation capex, not just distribution.
• What got quieter / a new wrinkle: the abrupt June-3 CFO change introduced an unscripted note — analysts flagged "NorthStar concerns" and "potential strategic changes," which management moved quickly to damp by reaffirming guidance. Net call sentiment: steady-positive on the regulated core, a small confidence ding on governance continuity.

Lens 7 · Comps

Peer set = large-cap regulated electric/gas utilities with similar 6-8% EPS-growth profiles. All multiples `` with date; market cap / EV/EBITDA where sourced, else n/a.

P/E (trailing) for CMS ≈ 20.3x on $3.53; 5-yr avg ROE n/a (compute: 2025 NI to common $1,061M / avg common equity ≈ ($8,920M+$8,006M)/2 = $8,463M → ~12.5% ROE ).

Read: CMS at ~19x forward is in line with DTE (~19x) and a turn-plus cheaper than WEC (~22x), with a comparable 6-8% growth algorithm and a slightly lower yield than WEC. It is neither the cheap nor the premium name in its cohort — fairly priced against peers, with the data-center optionality arguably under-credited vs WEC's premium.

Lens 8 · Stock-Price Catalysts (what moves CMS >5%)

The 5-yr total-return series in the 10-K tells the structural story bluntly:

CMS returned +34% total over 5 years — lagging both the S&P 500 (+96%) and even its own utility index (+64%). Over the trailing 52 weeks it returned ~7.5% vs the Utilities SPDR +13.2% and S&P 500 +27.9%. CMS has been a chronic relative underperformer — the price reacts to:

1. Interest rates (the dominant driver — utilities are bond proxies; the 10-K names rate-driven funding cost as a core risk).
2. Rate-case orders (e.g., BMO cut its target on the rate-case filing).
3. Governance/management (the June-3 CFO exit triggered the Jefferies downgrade and a confidence wobble).
4. Data-center contract milestones (the emerging positive catalyst — fair-value estimates nudged up to ~$82 on data-center optimism).

The pattern: CMS does not get the AI/data-center growth multiple yet; the market still treats it as a rate-sensitive bond proxy. That gap is the potential re-rating thesis.

---

Phase C — Judge people & books

Lens 9 · Management

• CEO — Garrick J. Rochow (age 51), President & CEO since Dec 2020. Internal promote; oversaw the coal-exit acceleration, the Covert acquisition, and the data-center pivot. Track record: delivered top-of-range EPS in 2025 and a clean 6-8% compounding algorithm; no major value-destroying M&A. Founder vs. professional manager: career operator (professional-manager archetype) — appropriate for a regulated-utility stage where capital discipline > vision.
• CFO transition — the live issue. Rejji P. Hayes retired effective June 3, 2026 after ~9 years as CFO (a widely respected, long-tenured CFO). Successor: Srikanth "Sri" Maddipati (43), a 12-year internal CMS veteran (joined 2014 as assistant treasurer) overseeing IR/Treasury/Tax/Accounting/FP&A. Assessment: orderly internal succession, not a scandal exit — but its abruptness ("immediate") surprised the Street and raised questions about NorthStar strategy. Jefferies cut to Hold (PT $88→$74) on it. Net: a continuity ding, mitigated by an internal hire and an immediate guidance reaffirmation. Worth monitoring, not a thesis-breaker.
• COO — Tonya L. Berry (EVP & COO since July 2025, 7-year+ internal). Deep bench of internal promotes — a culture signal (consistent with the "CE Way" lean operating system).
• Capital allocation. ~12.5% ROE; dividend raised to $2.28 for 2026 — the 20th consecutive annual increase, 2025 declared $2.170 vs $2.060 (2024) vs $1.950 (2023) = ~5.5% dividend CAGR, payout ~61% of EPS. No buybacks (appropriately — it issues equity to fund growth: $497M of forward sales settled in 2025 under a $1B ATM program). Capital allocation is disciplined and rate-base-focused; the only "destruction" was a small $15M renewable-development impairment in 2025.
• Skin in the game / red flags: insider ownership n/a — insider-transactions.csv not present; Part III comp incorporated by reference to the May-2026 proxy. No related-party red flags surfaced. Internal-promote-heavy bench is the dominant governance signal.

Lens 10 · Forensic Red Flags

Acting as a forensic analyst on FY2025 + Q1-2026, the accounting is clean for a regulated utility, with the usual rate-regulated nuances to watch:

• Cash flow vs. earnings. Net income $1,002M vs operating cash flow $2,235M (CFO > NI, normal for a heavy-D&A utility: D&A $1,306M). No earnings-quality red flag — earnings are if anything conservatively backed by cash. The watch item is structurally negative FCF (capex $3,824M >> CFO) funded by debt + equity issuance — financing-dependent but standard for the model.
• Regulatory assets (the utility-specific item). $104M current + $3,355M non-current regulatory assets; $4,091M non-current regulatory liabilities. These hinge on the MPSC continuing to deem recovery "probable"; a regulatory reversal would force a write-off. The J.H. Campbell emergency-order costs are NOT yet in MPSC rates and lack a FERC-approved MISO tariff for recovery — an explicit, unrecovered-cost overhang Consumers is litigating at FERC. This is the single sharpest forensic flag: real costs incurred, recovery not yet secured.
• Receivables. AR + accrued revenue $1,306M (2025) vs $1,049M (2024), +24.5% vs revenue +13.6% — receivables outrunning revenue, partly weather/PSCR-undercollection-driven. Mild flag, explained by gas-cost pass-through timing, not aggressive recognition.
• SBC / non-GAAP gap. Adjusted EPS $3.53-area vs GAAP — the adj-vs-GAAP gap is small (Q1 adj $1.13 vs GAAP $1.10), so non-GAAP is not flattering results materially. Clean.
• Self-reported error (minor). In the 2025 electric rate case, "Consumers identified an error affecting the calculation of the rate increase" — corrected via an MPSC Errata in April 2026 ($277M → $217M, no change to total revenue requirement or customer rates). Self-caught, immaterial to economics, but a small process-control note.
• Holding-company structure. CMS parent carries ~$5.9B of its own LT debt and depends entirely on Consumers' upstream dividends ($898M paid to CMS in 2025) to service it. Structural subordination is disclosed and managed (parent debt/cap 0.56 vs <0.70 covenant).

Regulatory findings (required). Per regulatory/regulatory-findings.md: No SEC Litigation Releases and no AAERs naming CMS Energy since 2021 (searched via SEC EDGAR EFTS). Item 3 (Legal Proceedings): the 10-K discloses no material standalone litigation — it points to Note 3 (Regulatory Matters) and Note 4 (Contingencies), plus "routine lawsuits…incidental to their businesses". Non-SEC enforcement web search ("CMS Energy" (FTC OR DOJ OR settlement OR penalty) enforcement): no material federal enforcement actions surfaced; the only quasi-adversarial matters are the ratepayer-advocacy challenges to the J.H. Campbell emergency orders (MI Attorney General, environmental groups) — these are regulatory/political, not enforcement. Environmental contingencies (23 former MGP sites, coal-ash/CCR remediation) are ordinary-course and rate-recoverable. Conclusion: No material regulatory or accounting-fraud findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3 as of 2026-06-23. The one genuine overhang is recovery risk on J.H. Campbell, not misconduct.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (FY2026 → FY2028, adjusted EPS)

Built bottom-up from the FY2025 base ($3.53 GAAP / ~$3.54 adj) and the 6-8% growth algorithm, anchored to the actuals + guidance and consensus. Output ``; inputs labeled.

• Inputs: rate-base growth >8%/yr; authorized ROE ~9.9%; capex $4.4B (2026) rising to $6.2B (2028); equity dilution from the ATM (~1-2%/yr); rising interest expense (LT-debt interest $798M in 2025, +14% YoY) as a headwind; data-center load adding capex (each GW = $2-5B) without raising customer rates.

At ~$74 and base FY2026 $3.87, the stock trades ~19.1x forward — and ~16.6x FY2028 base. The math: 6-8% EPS growth + ~3% yield = ~9-11% total-return algorithm, fair for a high-quality regulated utility but not a bargain. The asymmetry is to the upside only if the data-center pipeline converts faster than the 6-8% guide assumes (the bull EPS path) and the market re-rates the multiple from "bond proxy" toward "AI-adjacent grid build."

Forecast NOT logged to forecast.ts (watchlist/unattended run — Brier forecast is logged only on a committed base case via /thesis).

Lens 12 · Bull vs Bear

Bull case. A pristine regulated monopoly in a constructive jurisdiction with a once-in-a-generation demand catalyst. The ~9 GW data-center/manufacturing pipeline is the differentiator: it extends the capex runway (each GW = $2-5B) and improves affordability (−~2% to customer rate CAGR per GW), which neutralizes the political ceiling on rate-base growth and could push EPS to the high end / above the 6-8% band. 20 straight years of dividend increases, ~12.5% ROE, solid IG ratings (Moody's Baa1 CMS / Aa3 Consumers senior secured ), $25.8B 5-yr capex visibility. If the market ever re-rates Michigan grid-build as "AI infrastructure," there's multiple upside on top of the EPS compounding.

Bear case (permanent-impairment risks). (1) Regulatory de-rating — an MPSC ROE cut (toward the ~9% national drift) or a recovery disallowance (J.H. Campbell costs are not yet in rates ) would structurally lower the earned spread on the entire rate base. (2) Rates stay higher for longer — a $18.9B debt load re-financing into higher coupons compresses the equity return; utilities are the most rate-sensitive sector and CMS has already underperformed (+34% 5-yr total vs +64% for its own utility index ). (3) Data-center load disappoints — the 10-K explicitly warns the load "may not develop as anticipated," developers may pick other regions, and even contracted customers "may not fulfill" obligations; the whole upside narrative is contingent.

Pre-mortem (18 months out, thesis broke). Most likely failure: the data-center pipeline slips (hyperscalers delay/relocate or a 2027 AI-capex digestion hits), the multiple stays a bond-proxy ~17x, AND a 2026-27 rate case lands a lower ROE — so EPS grows ~6% (low end) while the multiple de-rates, and the stock dead-money's at ~$72 for two years. Secondary failure: the new CFO signals a NorthStar strategic shift that the market dislikes.

Are multiples too high? No — ~19x forward is in line with DTE and below WEC, fair for the growth. The risk isn't the multiple being expensive; it's the multiple having no upside re-rating unless data centers convert.

Contrarian view (what the market refuses to see). The market is still pricing CMS as a rate-sensitive bond proxy (hence the chronic underperformance), while the data-center pipeline quietly converts it into a demand-growth utility whose load growth is uniquely accretive to affordability. The Street credits WEC with a premium and CMS with none, despite a comparable algorithm and arguably better load-growth optionality. If even 2-3 GW of the pipeline signs firm 15-year contracts, the "boring bond proxy" framing breaks.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• The growth is regulator-granted, not earned — and the regulator can take it back. Every dollar of the 8%+ rate-base growth requires MPSC approval at a healthy ROE. ROEs nationally have drifted toward ~9%; a single adverse Michigan rate order resets the earning power of a $30.7B asset base. The 9.90%/9.80% authorized ROEs are the whole thesis and are not contractually guaranteed.
• Revenue concentration is absolute — one state, one regulator. 100% Michigan, ~96% one subsidiary. There is zero geographic or regulatory diversification. A Michigan recession, a hostile MPSC composition change, or a municipalization push (a named risk) hits everything at once.
• The data-center story is the most dangerous kind of bull narrative: unsigned and far-dated. The flagship 1 GW load doesn't ramp until 2030; the 9 GW "pipeline" is mostly "qualified/advanced," not contracted. The 10-K itself lists four separate ways it could fail (won't develop, competition steals it, zoning/permitting blocks it, contracted customers default). Bulls are paying today for 2030 optionality.
• Unrecovered J.H. Campbell costs. DOE is forcing a coal plant to run that Consumers wanted to retire; there is no FERC-approved tariff to recover those costs yet and they're not in MPSC rates. That's a live, quantifiable hole being litigated.
• Most dangerous competitor bulls underestimate: not another utility — interest rates and the alternative. At a ~3% yield and ~9-11% total-return algorithm, CMS competes with risk-free Treasuries; if the long end stays elevated, the bond-proxy bid never returns and the stock stays a relative laggard (as it has for 5 years).
• Governance wobble. A "respected, long-tenured" CFO left abruptly, triggering a sell-side downgrade and NorthStar-strategy questions. Successions this sudden sometimes precede strategy or disclosure changes.

If growth disappoints 20-30% (EPS growth falls to ~5% and the multiple de-rates to ~16x): FY2028 EPS ~$4.10 × 16x ≈ $66 — roughly −11% from $74, before dividends. The downside is cushioned by the regulated floor and the dividend, but there's real dead-money/mild-loss risk.

Single scenario that permanently impairs: a structural shift in Michigan regulation (legislated retail competition / community-solar carve-outs / a sustained low-ROE regime) that breaks the monopoly-plus-constructive-ROE compact. Plausibility: low (Michigan has been reliably constructive), but it is the only thing that kills the model rather than just denting growth.

Lens 14 · Management Questions (ordered by information value)

1. Of the ~9 GW data-center pipeline, how many GW are under firm, signed 15-year large-load contracts (vs LOIs/qualified), and what is the contracted in-service ramp by year?
2. What ROE and equity-thickness are you assuming in the 2025 electric ($217M) and gas ($240M) rate cases, and what's your read on the MPSC's appetite given affordability politics?
3. On J.H. Campbell: what is the expected timing and probability of a FERC-approved MISO tariff for emergency-order cost recovery, and what's the unrecovered exposure if it's denied?
4. Why did Rejji Hayes depart immediately, and does the CFO change signal any reconsideration of NorthStar's strategic role or potential divestiture?
5. The June-2026 IRP preview adds two new ~1,500 MW gas plants — how does new owned gas generation square with the 100%-clean-by-2040 standard, and what's the rate-base/return profile vs PPAs?
6. How much incremental capex beyond the current $25.8B 5-yr plan would 4-5 GW of "advanced-stage" data-center load require, and how would you fund it without over-diluting?
7. What is the embedded vs. authorized ROE gap today, and how much regulatory lag are you carrying on the rising-capex base?
8. At what point does the equity-issuance need (ATM/forward sales) become dilutive enough to pressure the 6-8% per-share growth, given a structurally negative FCF profile?
9. What is the sensitivity of the 6-8% EPS algorithm to a sustained +100bps in your average cost of debt?
10. How do you protect existing ratepayers if a contracted data center defaults on its 80% minimum-demand obligation — is the collateral/exit-fee structure tested?
11. What is the current consolidated rate base, and your projected rate base in 2030 under the base vs data-center-upside cases?
12. How exposed is the capex plan to equipment tariffs and solar/battery supply-chain constraints, and what's in rates vs. at-risk?
13. What's the plan for the hydro-dam sale (13 dams + 30-yr PPA) and its rate/earnings impact if MPSC/FERC approval slips beyond 12-18 months?
14. How should we think about NorthStar's earnings durability given the DIG outage volatility and the tax-equity-driven (BG/HL Solar) structure?
15. What reliability-incentive ($10M annual upside/downside) outcome is embedded in 2026 guidance, and how are you tracking the new performance metrics?

---