DTE Energy

Phase A — Understand the business

Lens 1 · Company Overview

DTE makes money the way every vertically-integrated regulated utility does: it invests capital in poles, wires, pipes, and generation; a state commission (the Michigan Public Service Commission, MPSC) sets rates that let it recover those costs plus an authorized return on the equity portion of the rate base; and it earns that return more or less mechanically so long as it keeps spending prudently and the commission stays constructive. The "product" is regulated reliability; the "moat" is a legal monopoly franchise.

Five reportable segments:

• Electric (principally DTE Electric) — the crown jewel. 2025 net income $1,158M (79% of total), on operating revenues $6,935M. Generates, purchases, and distributes electricity to 2.3M customers in southeastern Michigan. Generation mix: coal (retiring fast), a pumped-storage plant, the Fermi 2 nuclear plant (~15% of generation, ~1,170 MW), plus growing wind and solar, supplemented by purchased power.
• Gas (principally DTE Gas) — 2025 net income $295M, revenues $2,052M. Distributes natural gas to 1.4M customers statewide (corp. organized 1898).
• DTE Vantage — non-utility renewables (RNG projects), custom energy solutions, and a steel/coke business (EES Coke). 2025 net income $154M. Earnings are heavily tax-credit-driven ($84M of PTCs/ITCs in 2025).
• Energy Trading — physical/financial power, gas, and environmental marketing. 2025 net income $123M but with enormous mark-to-market noise (revenues swung $3.8B → $6.5B YoY). This is the segment that makes GAAP EPS diverge violently from operating EPS.
• Corporate & Other — holding-co interest and tax; a $(268)M net loss drag in 2025.

Contract structure / payment terms. The utility revenue is not take-or-pay but is protected by a battery of regulatory cost-recovery mechanisms — Power Supply Cost Recovery (PSCR), Gas Cost Recovery (GCR), infrastructure recovery mechanisms (IRM), securitization, EWR, RDM decoupling — that pass fuel and specific program costs through to customers with limited earnings risk. ~10% of retail sales are open to alternative suppliers under Michigan's capped retail-access program; the rest is captive. Customer concentration is nil — the filing explicitly states no single customer is material to DTE Electric — though the incoming 3-4 GW of hyperscaler load will, for the first time, create meaningful large-load concentration if it lands.

Lens 2 · Supply Chain

Commercial-layer files for the energy topic are missing (kb/energy/wiki/* not compiled), so this lens is grounded in the 10-K + web. A regulated utility's "supply chain" is fuel-in / electrons-and-molecules-out:

Upstream (inputs):

• Coal — long-term contracts for ~6.9M tons low-sulfur western coal + ~1.3M tons Appalachian coal delivered 2026-2027. A shrinking input: all 11 coal units at Trenton Channel, River Rouge, and St. Clair are retired; the last 5 units (Belle River unit + 4 at Monroe) retire in 2026/2028/2032.
• Natural gas — purchased for both the gas utility (GCR pass-through) and gas-fired generation; the replacement fuel as coal exits.
• Nuclear fuel — for Fermi 2 (uranium/enrichment; the chokepoint the broader nuclear-fuel-cycle names like Centrus/Cameco sit upstream of).
• Renewables equipment — wind turbines, solar panels, and increasingly battery storage (Trenton Channel is being repurposed to a BESS in 2026). Exposed to the same panel/module supply chain as First Solar/Nextracker peers, and to international steel prices (the 10-K flags steel-price volatility as a risk to the EES Coke steel business).

The company (midstream): DTE Electric's generation fleet → its own transmission/distribution grid → meters. DTE Gas's storage + pipeline network → distribution mains → meters.

Downstream (end customers): 2.3M electric + 1.4M gas Michigan retail accounts (residential/commercial/industrial), plus — the new node — hyperscaler data centers. Named buyers now on the chain: Oracle (1.4 GW, under construction), Google (1 GW, filed for MPSC approval, decision ~Sept 10 2026), and ~2 GW of unnamed hyperscalers in advanced negotiation.

Chokepoints / single-source dependencies: (1) The MPSC is the single most important node — it is a regulatory chokepoint, not a supplier, but every dollar of return flows through its orders. (2) MISO grid interconnection and capacity — the 10-K flags "potential capacity constraints in the MISO region." (3) Fermi 2 is a single large asset (~15% of generation); an extended outage would force expensive purchased power. Names or it didn't happen — the chain is: {western + Appalachian coal miners, Enrico Fermi 2 nuclear fuel suppliers, gas producers, wind/solar/battery OEMs} → DTE Electric/DTE Gas → {2.3M/1.4M Michigan retail accounts + Oracle + Google + ~2 GW hyperscalers via MISO}.

Lens 3 · Competitive Advantages (moats)

For a regulated utility, "moat" and "regulatory franchise" are the same thing.

• Legal monopoly (the dominant moat). DTE Electric and DTE Gas hold exclusive service franchises for their Michigan territory. No competitor can string a competing distribution grid; retail access is legislatively capped at ~10%. This is the deepest, most durable moat in the equity universe — it is granted by statute, not earned in a market.
• Regulatory relationship / constructive commission. DTE describes "a constructive regulatory environment and solid relationships with their regulators". Michigan is generally rated a constructive-to-average regulatory jurisdiction; forward test years are permitted (the 2025 electric case used a projected 12-month period ending Dec 2026), which is a favorable feature. But the moat is only as good as the ROE the commission grants — and at 9.9% electric / 9.8% gas, DTE's authorized ROEs sit at the lower end of the peer range (many peers are 9.9-10.5%), and the MPSC declined to raise the electric ROE in the Feb 2026 order.
• Scale + rate-base momentum. A $54B asset base and a $36.5B five-year capex plan create a self-reinforcing flywheel: capex → rate base → authorized return → earnings → capacity to fund more capex. Bigger rate base = bigger absolute earnings even at a fixed ROE.
• Bargaining power. Over customers: near-total (captive monopoly), constrained only by the affordability ceiling the commission enforces politically. Over suppliers: moderate — DTE is a large fuel/equipment buyer but is a price-taker on commodity coal, gas, uranium, and increasingly on solar modules and batteries where the whole industry is bidding for the same constrained supply.
• The new, contingent moat — grid position for AI load. DTE happens to sit on transmission capacity and generation in a state courting data centers. A hyperscaler cannot easily build 1.4 GW of dedicated power anywhere else in southeastern Michigan; DTE is the counterparty. This is a location moat that only matters now because AI made megawatts scarce.

Moat verdict: Wide and durable on the regulated franchise; the differentiator vs. peers is below-average (lower authorized ROE, Michigan storm/reliability political heat), offset by an above-average new catalyst (data-center load).

Lens 4 · Segments

Segment net income by year:

Reading the trend:

• Electric is the story — net income climbed from $772M (2023) → $1,158M (2025), driven by implementation of new rates (+$194M rev in 2025, +$338M in 2024), interconnection sales (+$231M in 2025), PSCR, and favorable weather. This is the rate-base flywheel working. Operating income $1,214M → $1,437M → $1,661M.
• Gas is a smaller, weather-whipsawed rate-base story; the 2024 dip was unfavorable weather, 2025's recovery was favorable weather (+$119M) plus new rates.
• The two non-utility segments are noise, not signal. Energy Trading's net income collapse (336→123) is timing-related mark-to-market on gas structured transactions — it will keep whipping GAAP EPS around. DTE Vantage lives on production/investment tax credits ($84M in 2025) and a fading steel/coke business (Steel revenue −$100M in 2025 on weak demand).
• Corporate & Other is a growing drag (−$268M) as higher interest expense on a debt-funded capex program and higher taxes (One Big Beautiful Bill + Illinois tax-law changes) bite. This is the balance-sheet cost of the growth showing up in the P&L.

Geography: Effectively 100% Michigan for the utilities; DTE Vantage/Energy Trading operate across the US but are immaterial to the thesis. There is no meaningful international exposure (steel-price sensitivity aside).

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print: Q1 2026, reported 2026-04-30)

Headline: GAAP net income $247M, GAAP diluted EPS $1.19 — down sharply from Q1 2025's $445M / $2.14. But this is a GAAP head-fake. On an operating basis, operating earnings were $407M, operating EPS $1.95, up 21% YoY — a clear beat.

Why the divergence: The GAAP decline was driven by the Energy Trading and DTE Vantage segments plus Corporate & Other — i.e. the mark-to-market timing swings on gas structured transactions that DTE excludes from operating EPS. The core regulated business strengthened.

What drove the good number: DTE Electric segment net income jumped $71M to $218M in Q1 2026, on favorable tax timing (RNG/renewable credits), implementation of new rates, and colder weather. This is exactly the composition a utility bull wants — the regulated engine up, the volatile trading book down.

Guidance & tone: Management reaffirmed 2026 operating EPS guidance of $7.59-$7.73 (6-8% growth over the 2025 midpoint), and signalled bias to the high end on RNG tax credits. Tone was notably more bullish than the 10-K, because the data-center pipeline expanded between the two filings.

Balance-sheet flags: Cash from operations $3,409M (down $234M YoY on working capital); capex ~$4.4B utility; net cash from financing +$2,057M (the tell — the capex program is debt/equity funded, not self-funded). $1.4B of long-term debt matures within 12 months. Qualified pension underfunded by $127M. A $1.5B ATM equity program was filed Dec 2025 (undrawn at year-end), with $500-600M of equity issuance planned in 2026 and again in 2027-2028 — dilution is a structural feature of the funding plan.

Market reaction / what's priced in: The stock trades ~$153 against a Street consensus target ~$159 — essentially fairly valued, ~4% implied upside. The market is not yet fully paying for the 3-4 GW pipeline; it is paying for the reliable 6-8% grower.

Unusual vs. own history: The GAAP/operating gap ($1.19 vs $1.95) is unusually wide this quarter — a reminder that DTE's reported GAAP EPS is a poor quarterly signal because of Energy Trading. Judge this name on operating EPS and segment income, not GAAP.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on the research shelf (transcripts/ empty), so this is web-grounded and should be deepened on the next refresh.

Tone trajectory (Q3 2025 → Q4 2025 → Q1 2026): A clear, monotonic escalation of the data-center narrative, which is the single most important sentiment shift for this name.

• Q3 2025: Provided "2026 early outlook"; data center framed as optionality.
• Q4 2025 (Feb 2026): Set formal 2026 guidance ($7.59-$7.73), unveiled the $36.5B capex plan, and disclosed the 1.4 GW deal as secured.
• Q1 2026 (Apr 2026): The pivot. Management quantified a 3-4 GW hyperscaler pipeline (Oracle 1.4 GW under construction; Google 1 GW filed; ~2 GW in advanced talks) and stated explicitly that data-center wins could push EPS growth above 8%.

Recurring phrases management leans on: "constructive regulatory environment," "long-term 6-8% operating EPS growth," "customer affordability," "strong balance sheet," "bias to the high end." The affordability drumbeat is defensive — it is the political counter-pressure to a large capex ask, and management repeats it to pre-empt the MPSC.

What they've stopped saying: Coal is fading from the narrative (the fleet is nearly retired); the story has shifted decisively from "energy transition / decarbonization" (the 2021-2023 framing) to "load growth / grid investment / data centers" (the 2025-2026 framing). That is a favorable narrative rotation — from a cost/compliance story to a growth story.

Lens 7 · Comps

Regulated integrated-utility peers. Multiples are `` with source/date; where I could not source a specific figure I mark it n/a rather than fabricate. All caps in USD.

5-year average ROE: not sourced for the peer set — do not fabricate. DTE's current ROE is 10.4%, modestly above its 9.9% authorized (utilities typically earn slightly above authorized via off-system sales and non-utility income).

Read: DTE at ~19.3x forward is squarely mid-pack — cheaper than WEC (~22x) and Southern (~19-22x), roughly in line with CMS (its closest comp, the other Michigan utility) and a shade above Duke (~18.7x). The 3.06% yield is at the low end of the 2.8-3.2% peer band. Nothing here screams cheap or expensive — DTE is priced as an average regulated grower. The mispricing case, if there is one, is that ~19x does not yet capitalize the above-trend growth (>8%) that a 3-4 GW data-center win would unlock; the comp set is priced for ~6-7% growers. Best single comp is CMS — same state, same regulator, same ~6-8% algorithm — and DTE trades at a similar-to-slight-discount multiple to it.

(Provenance caveat: the peer forward P/Es above are single-source web reads at slightly different dates and should be treated as directional, not precise. A rigorous comp table would pull a same-date consensus screen.)

Lens 8 · Stock-Price Catalysts (5-year, >5% moves)

Mostly `` + inference; utilities are low-beta (DTE beta 0.38) so >5% single-day moves are relatively rare and almost always regulatory or storm-driven, not earnings-surprise-driven.

• July 2021 — DT Midstream spin-off. Structural re-rating event: DTE shed its commodity-exposed midstream business, leaving ~90% regulated. Reduced beta, simplified the thesis, and is the reason the current name is a "pure-play utility."
• Feb 2023 — the ice storm. Major multi-day outages triggered political/MPSC backlash on reliability. Reliability became the dominant regulatory and reputational risk; DTE responded with hardening capex (upgraded circuits saw 33% fewer outages 2H23). Storm-cost recovery and prudence disallowances (e.g. the ~$33M disallowance in the Feb 2025 order) trace to this.
• Dec 2016 / ongoing — Fermi 2 license renewal to 2045. Removed a large-asset tail risk; the plant (~15% of generation) is a clean baseload anchor through 2045.
• Rate-case orders (recurring, the real catalysts). These are the earnings for a utility. Most recent: Feb 19, 2026 electric order — +$242M annual revenue, ROE held at 9.9% (vs the $574M / 10.75% requested). A partial win (revenue yes, ROE bump no) — the kind of "constructive but not generous" outcome that keeps the stock range-bound rather than re-rating.
• 2025-2026 — the data-center pivot. The emerging catalyst: each MPSC approval of a hyperscaler contract (Oracle done; Google decision ~Sept 10 2026) is a discrete, datable upside event. The Google order in September 2026 is the next hard catalyst.

Pattern the market reacts to: For DTE, the tape responds to (1) rate-case outcomes, (2) storms/reliability, and (3) — newly — data-center contract approvals. It does not react much to GAAP quarterly EPS (too much trading noise) or to macro beyond the rates channel (utilities sell off when the 10-year yield spikes, because they're bond proxies — beta to rates, not to the economy).

---

Phase C — Judge people & books

Lens 9 · Management

• Track record. CEO Joi Harris (age 55) took the seat Sept 8, 2025 after a multi-year, well-telegraphed succession; a 34-year DTE lifer (since 1991), ex-President/COO of the parent and previously President/COO of DTE Gas. As COO she is credited with a large reliability improvement (press framing: "70% reliability boost") — directly relevant given the 2023 ice-storm scar. Jerry Norcia (CEO 2019-2025, at DTE since 2002) moved to Executive Chairman and remains an advisor — continuity, not rupture. This is a stability archetype: an internal operator handed a running machine, not a turnaround artist or an empire-builder.
• Tenure & skin in the game. Deep bench tenure (both CEO and Chair are 20-35 year DTE veterans). Insider ownership data not on the shelf (insider-transactions.csv absent) — not sourced; for a utility this size insider ownership is typically low single-digit % and not a material alignment signal. The alignment mechanism is comp tied to operating EPS growth + reliability + safety metrics, standard for the sector.
• Capital allocation. The defining recent decision — the 2021 DT Midstream spin — was value-accretive and disciplined: it de-risked the story and let the market re-rate DTE as a pure utility. Ongoing allocation is textbook regulated-utility: plow $36.5B into rate base (2026-2030), fund with a debt/equity blend, grow the dividend (~$3.88→$4.15→$4.44→$4.66 per share; >100 consecutive years of dividends), target a pure-play payout ratio (~77% currently). ROE ~10.4% and stable. No value-destroying M&A, no buyback-at-the-top games — the non-utility investments (DTE Vantage) are deliberately capped and criteria-gated ("start with a limited investment, evaluate, expand or exit").
• Red flags. Modest. (1) The EES Coke steel/coke subsidiary carries Clean Air Act litigation and estimated litigation penalties ($13M booked in 2025) — a non-core distraction. (2) The persistent, growing Corporate & Other loss (−$268M) shows the interest cost of the growth plan. (3) The reliance on operating (non-GAAP) EPS as the headline metric — legitimate for this business, but it means investors must trust management's adjustments (mostly Energy Trading MTM, which is defensible).
• Founder vs. professional manager. Purely professional managers — a career-utility executive team. For a regulated monopoly at this stage, that is the right archetype: you want steady operators who execute rate cases and keep the lights on, not visionaries. The one place vision matters — capturing the data-center wave — the team is demonstrably moving fast (Oracle signed, Google filed).

Lens 10 · Forensic Red Flags

Acting as a forensic analyst. For a rate-regulated utility, the accounting risk surface is different from an industrial — the biggest items are regulatory assets/liabilities (does the balance sheet carry recoverable costs that might get disallowed?) and derivative mark-to-market (Energy Trading).

• Regulatory assets/liabilities. Net regulatory liabilities of ~$2,881M and regulatory assets embedded in rate base. These are recoverable if the commission continues to allow it — the 10-K's own critical-accounting note flags that a change in the regulatory environment could force write-offs. The ~$33M storm-cost disallowance (Feb 2025 order) is a small, concrete example of this risk crystallizing. Watch item, not a red flag — Michigan remains constructive.
• Cash flow vs. earnings. FY2025 cash from operations $3,409M against GAAP net income $1,462M — CFO comfortably exceeds earnings (normal for a heavily-depreciating utility: D&A was ~$1.8B+ across segments). No earnings-quality red flag here — the gap is depreciation, not accruals. The genuine tension is that CFO ($3.4B) does not cover capex ($4.4B) + dividends ($871M) — the ~$1.9B annual external-funding gap is the real financial fact, met by debt and equity issuance. This is structural for a growth-phase utility, not fraud, but it is why the balance sheet and dilution matter.
• Receivables / inventory vs. revenue. No flags surfaced; utility receivables track billing and are managed via uncollectible reserves (sensitive to low-income assistance funding, which management calls out).
• SBC flattering non-GAAP. Immaterial — utility SBC is small; the non-GAAP bridge is dominated by Energy Trading MTM timing, not SBC.
• Goodwill / intangibles. Annual impairment test Oct 1, 2025 used discount rates 6.1-8.9%; each reporting unit's fair value substantially exceeded carrying value — no impairment. Clean.
• Pension. Qualified pension underfunded $127M; OPEB overfunded $513M; plan assets $5.4B. Nominal required contributions expected. Low risk.
• Derivatives / Energy Trading. The one area of genuine opacity — gas structured transactions are marked to market while the offsetting non-derivative contracts are not, producing large timing swings ($70M unfavorable timing change in 2025; $167M in 2024). This is disclosed, understood, and reverses on settlement — but it makes GAAP quarterly EPS unreliable. Judge on operating EPS.
• Leverage. Debt/equity 2.19x; long-term debt $23.8B on $12.3B equity (~66% debt capitalization). High in absolute terms but normal for a regulated utility (rate base is largely debt-funded, and the regulator sets an equity-thick ratemaking capital structure). The collateral-posting trigger on a below-investment-grade downgrade ($483M) is a standard tail risk; DTE is solidly investment grade.

Regulatory findings (required sub-section):

• SEC Litigation Releases / AAERs: None. Verified via SEC EDGAR EFTS (LR + AAER) for 2021-07-06 → 2026-07-06 — total_sec_findings: 0.
• Item 3 (Legal Proceedings), most recent 10-K: DTE cross-references Notes 9 (Regulatory Matters) and 18 (Commitments & Contingencies); it discloses environmental proceedings only where government sanctions ≥$1M are expected. The material live item is EES Coke Battery, LLC — the Michigan coke subsidiary — facing EPA Clean Air Act Findings of Violation (SO2 emissions, 2018-2019 and a challenge to its 2014 permit), with the Sierra Club and City of River Rouge intervening and DTE Energy parent entities named as defendants to share potential penalties. DTE booked ~$13M of estimated litigation penalties in 2025 (including $8M for EES Coke). Material to EES Coke, immaterial to DTE Energy consolidated.
• Non-SEC enforcement (web search): No material FTC/DOJ/CFPB actions surfaced against DTE Energy. The relevant regulatory heat is MPSC/EGLE/EPA environmental and reliability oversight (ice-storm accountability, coal-plant compliance) — routine for a Michigan utility, not enforcement-grade fraud.
• Conclusion: No material accounting or securities-fraud findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3/Note 18 as of 2026-07-06. The only live legal exposure is the EES Coke Clean Air Act matter, quantified in the low tens of millions and non-core.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (operating EPS, FY2026-FY2028)

Built bottom-up from FY2025 operating EPS of $7.36 and management's reaffirmed FY2026 guidance of $7.59-$7.73. DTE's algorithm is explicit: 6-8% long-term operating EPS growth, funded by $36.5B of rate-base capex, with ~$500-600M/yr equity dilution.

Input assumptions (each labeled):

• FY2026 base = midpoint of guidance ≈ $7.66.
• FY2027-28 growth = 7% base (mid-point of the 6-8% algorithm), 8% bull (data-center load lands, RNG credits, high end), 5% bear (adverse rate-case outcomes, higher-for-longer rates lift interest drag, dilution runs hot).
• Dilution ~1% shares/yr from the ATM/DRIP (already inside management's per-share guide).

Operating EPS projection:

What moves it: The single biggest swing factor is whether the 3-4 GW hyperscaler load converts — management says data-center wins push growth above 8%, i.e. beyond the top of this range. Rate-base capex is the mechanical driver; interest expense (Corporate & Other drag) is the mechanical headwind. Weather and Energy Trading create quarterly noise but wash out annually in the operating number.

Valuation cross-check: At $153 and base FY2026 EPS $7.66, DTE trades at 20.0x current-year / ~18.6x FY2027 operating EPS — consistent with the 19.3x forward-P/E web read and mid-pack vs. peers. A re-rate to ~21-22x (WEC-like, justified if >8% growth becomes the base case on data-center wins) on FY2027 base EPS $8.20 implies ~$172-180 — roughly the Street high end. The bear (5% growth, no multiple help) holds the stock near $140-145.

Forecast NOT logged to forecast.ts (per --watchlist rule: skip the Brier create step in the breadth loop). If promoted to a thesis, log: "DTE FY2026 operating EPS >= $7.66, p≈0.75, resolves 2027-02-28."

Lens 12 · Bull vs Bear

Bull case. DTE is a rare thing: a low-beta, monopoly-moated, dividend-growing regulated utility that has accidentally become an AI-infrastructure play. The regulated engine already delivers a dependable 6-8% operating-EPS CAGR off a $36.5B rate-base program in a constructive Michigan jurisdiction. On top of that sits 3-4 GW of hyperscaler load — Oracle (1.4 GW) already under construction, Google (1 GW) filed with a September 2026 decision, ~2 GW more negotiating. Large-load additions grow rate base and spread fixed costs across more sales, which is doubly accretive and helps affordability — the political win-win. Management has said explicitly that data-center wins push EPS growth above 8%, which the current ~19x multiple does not capitalize. The Fermi 2 nuclear anchor (to 2045) and near-complete coal exit remove tail risks. You are paid 3% to wait, with dividend growth and >100 years of continuity. The contrarian bull line: the market still files DTE under "boring bond-proxy utility" when it has quietly become the pick-and-shovel landlord for Michigan's data-center build-out.

Bear case (2-3 things that could permanently impair or de-rate).

1. Regulatory ceiling. The moat is only as good as the ROE the MPSC grants — and the Feb 2026 order held ROE at 9.9% and awarded $242M vs $574M requested. If Michigan turns less constructive as customer bills climb (affordability backlash — the recurring theme in management's own script, and the 2023 ice-storm scar is fresh), the $36.5B plan earns a shrinking return and the growth algorithm breaks. This is the core structural risk.
2. Balance-sheet / rate sensitivity. 66% debt capitalization, a ~$1.9B/yr external funding gap (CFO doesn't cover capex + dividends), $500-600M/yr of dilution, and a bond-proxy valuation. In a higher-for-longer rate world, interest expense (the growing Corporate & Other drag) compounds and the multiple compresses as the yield gap to Treasuries narrows. Utilities were the worst S&P sector in 2022-2023 rate spikes for exactly this reason.
3. Data-center load is a filed hope, not banked earnings. Only Oracle is under construction; Google and the ~2 GW are contingent on MPSC approval and hyperscaler capex plans that can slip or shrink. If the AI-capex cycle cools before the contracts close, the ">8%" upside evaporates and DTE is just a fully-priced 6-7% grower.

Pre-mortem (it's Jan 2028 and the thesis broke — what happened?): Most likely story — a hostile rate-case cycle. Customer bills rose with the capex; a populist MPSC (or ballot-driven pressure) cut authorized ROE toward 9.5% and disallowed a chunk of storm/grid spend as imprudent; simultaneously the 10-year yield sat at 5%+, so the multiple compressed from 19x to 15x even as EPS grew. The Google/hyperscaler deals got trimmed as the AI-capex cycle paused. DTE still grew EPS ~4% but the stock fell 20% on multiple compression. Second, lower-probability story: a major reliability failure (another ice storm) triggered penalties and a reputational/regulatory crackdown.

Are multiples too high? No — 19x forward is mid-pack and defensible for a 6-8% grower with an embedded free option on >8%. It is not cheap either. The asymmetry is modest: capped upside from re-rating (+15-18% to the Street high), real downside from rate compression (~-15-20%).

Contrarian view — what the market refuses to see: Both directions. Bulls under-appreciate that the data-center load is a double-edged affordability sword — if not structured carefully (special large-load tariffs, take-or-pay), residential customers subsidize hyperscaler grid build, inviting exactly the regulatory backlash that caps the ROE. Bears under-appreciate that DTE's authorized ROE being low (9.9%) is actually a source of safety — there's less room for the commission to cut, and a low authorized ROE plus a fat rate base still compounds absolute earnings reliably. The market is treating "utility + data center" as unambiguously bullish; the truth is it's a regulatory-structuring problem whose resolution (the Google order, the large-load tariff design) is the real catalyst to watch.

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

Dismantling the bull case.

• What structurally breaks the model? DTE doesn't "make money" in a market — it collects an ROE the MPSC permits. The entire equity value is a leveraged bet that a political body keeps granting ~10% on an ever-growing asset base while customer bills rise. That's not a moat, it's a permission — and permission can be revoked at the ballot box or by a commission responding to affordability rage. The 2023 ice storm proved DTE has negative political capital in Michigan when the lights go out.
• Where is revenue concentrated, and what if it shifts? Today: diffuse (2.3M captive accounts). But the bull thesis deliberately concentrates it into a handful of hyperscalers. If Oracle/Google build their own generation (behind-the-meter gas or SMRs — which every hyperscaler is now exploring), or if one data-center project is cancelled, a chunk of the "growth" reverses. You'd be underwriting 3-4 GW of counterparty risk to three tech companies whose capex plans change quarterly.
• Why is the moat weaker than bulls think? Because the return on the moat is capped and shrinking relative to the capital deployed. A wider rate base at a flat/declining ROE, funded 66% by debt and topped up with perpetual equity dilution, is a treadmill: EPS grows but per-share value creation is thin once you net out dilution and the rising interest drag (Corporate & Other already −$268M and worsening).
• Most dangerous competitor bulls underestimate: Not another utility — it's behind-the-meter self-generation. If hyperscalers, frustrated by interconnection queues and utility timelines, increasingly co-locate their own gas/nuclear power (the Talen/Constellation/SMR model), the utility gets disintermediated from the exact load growth it's counting on. DTE is racing to sign these deals precisely because the alternative is the hyperscaler routing around it.
• Worst capital-allocation / governance items: Minor but present — the EES Coke Clean Air Act mess (why does a regulated utility still own a coke battery?), the reliance on non-GAAP operating EPS as the headline (which conveniently smooths a genuinely volatile trading book), and a growing holding-company loss that flatters segment optics.
• What must hold for today's price? (1) MPSC keeps ROE ≥9.9% and approves the $36.5B plan into rate base; (2) rates don't spike (the multiple is a duration bet); (3) at least Google + ~1-2 GW of the pipeline actually energizes; (4) no reliability catastrophe. Break any one and the 6-8% algorithm slips toward 4-5% while the multiple compresses.
• If growth disappoints 20-30%: A 6-8% grower becomes a ~5% grower; on a bond-proxy multiple that also compresses 2-4 turns, the stock is a -20% to -25% name, and you're left holding a 3.5% yield with sub-inflation-plus growth — a bond with equity risk.
• Single scenario that permanently impairs: A sustained Michigan affordability/reliability political crisis that structurally lowers authorized ROE and imposes recurring prudence disallowances — turning DTE from a 10% ROE machine into an 8.5% one. Plausibility: moderate — Michigan is constructive today, but the trend of rising bills + AI-load-subsidy optics + a fresh storm memory is exactly the fuel for it.

Lens 14 · Management Questions (ordered by information value)

1. On the 3-4 GW data-center pipeline — what is the contract structure? Are these take-or-pay / minimum-demand tariffs that protect other ratepayers and DTE if a hyperscaler walks, or merchant load DTE is building rate base against on faith?
2. How will the large-load tariff be designed so residential customers do not subsidize hyperscaler grid investment — and have you pre-cleared that structure with MPSC staff to avoid an affordability backlash?
3. What authorized ROE do you realistically expect over the next two rate cycles, given the Feb 2026 order held you at 9.9% against a 10.75% ask? At what ROE does the $36.5B plan stop creating per-share value net of dilution?
4. Walk through the funding stack for 2026-2030: how much incremental debt vs. equity, at what assumed cost of debt, and what's your plan if the 10-year sits above 5%?
5. What is the probability and timeline you'd put on the Google 1 GW MPSC approval (Sept 2026) and the ~2 GW in negotiation converting to signed contracts?
6. If a hyperscaler chooses behind-the-meter self-generation (gas or SMR) instead of DTE supply, how much of the load-growth thesis is at risk, and how are you defending against disintermediation?
7. How do you de-risk the coal-to-clean transition cost — specifically the Monroe retirements (2028/2032) and the Trenton Channel battery conversion — against stranded-asset and cost-recovery risk?
8. What is your credit-rating floor commitment, and how much equity dilution are shareholders signing up for through 2028 to defend it?
9. Reliability post-2023 ice storm: what specific, measurable SAIDI/SAIFI targets have you committed to the MPSC, and what's the penalty exposure if you miss?
10. On Energy Trading — given the persistent GAAP/operating divergence, why keep this segment at all? What's its strategic rationale versus the earnings-quality noise it creates?
11. Why does DTE still own EES Coke, and what is the full quantified tail exposure on the EPA Clean Air Act litigation including the Sierra Club intervention?
12. What is the realistic long-duration storage / SMR / hydrogen roadmap you referenced — capital, timeline, and cost-recovery mechanism — or is it aspirational?
13. What's the plan if Michigan's 100% clean-energy-by-2040 mandate (2023 legislation) proves more expensive than modeled in the next IRP (2026)?
14. Where can you take O&M cost out to fund affordability without cutting reliability — quantify the efficiency program?
15. If you had to choose between defending the dividend-growth track record and funding the full capex plan in a capital-constrained scenario, which gives?

---