Public urges scrutiny of DTE Saline Township data center proposal

DTE Energy has asked the Michigan Public Service Commission to fast-track approval (an ex parte motion) of a power contract to supply a massive OpenAI- and Oracle-backed data center campus in Saline Township that initially will draw 1.4 GW from the grid (with plans for an additional 3 GW later). Developers project 2,500 union construction jobs and 450 permanent roles, while opponents and some commissioners warn the project could strain the grid, raise residential rates, and pose environmental and water-use risks amid Michigan’s 2040 clean-energy mandate. The MPSC is scheduled to consider the request Dec. 5 after public comment from business supporters, labor groups and numerous residents urging greater regulatory scrutiny.

Analysis

Market structure: The Stargate campus (1.4 GW initial, potential +3 GW) creates a discrete demand shock that benefits hyperscalers (OpenAI/ORCL) and data‑center developers (Related Digital) while creating upside for a regulated utility that can recover capital and demand charges (DTE). If approved, DTE’s rate base and predictable load could increase EBITDA by a low‑single digit percentage over 2–4 years, but approval risk compresses valuation due to potential contested hearings and customer pushback. Grid supply/demand: adds material baseload-like demand equal to ~1 large nuclear unit; expect upward pressure on midcontinent wholesale power and capacity prices, and modestly higher natural gas dispatch for near‑term firming needs.

Risk assessment: Tail risks include MPSC denial/contested case causing >30–60% delay to build timelines, legal setbacks from local rezoning or water permits, or an AI demand reversal leaving stranded 4+ GW of capacity — each could drive >20% revaluation for exposed equities. Immediate window (days–weeks) centers on regulatory filings and public comment; medium term (3–12 months) is interconnection and permitting; long term (1–4 years) is construction and load realization. Hidden dependencies: transmission upgrades, PPAs, and potential carve‑outs to fossil generation undercut the clean‑energy mandate and create stranded asset/regulatory risk.

Trade implications: Tactical stance: small asymmetric bets. Buy ORCL exposure (call spread 6–9 month, +10% OTM) to capture AI revenue upside with limited premium; keep DTE equity exposure modest and hedged — establish 1–2% long DTE position but buy 6‑12 month protective puts (5–10% OTM). Consider short regional muni/utility bond basis if contested approval looks likely (benefit if spreads widen >50bp). Entry: size trades within 2–6 weeks around MPSC procedural outcomes; exit if a contested case is formally opened or approvals are granted.

Contrarian angles: Markets focus on headline approval vs. execution; they underprice multi‑year interconnection and water permitting risks that typically add 12–36 months to build timelines. Historical parallels: hyperscaler data‑center booms often saw 20–40% project schedule slippage and occasional write‑downs; if AI demand softens, stranded capacity risk materializes — making long‑only utility exposure without downside protection potentially overdone. A successful contested process could force DTE to secure PPAs/firming capacity, increasing near‑term capex and lowering shareholder returns contrary to the bullish narrative.

Key Decisions for Investors

• Establish a tactical 1–2% long position in DTE (ticker DTE) conditioned on MPSC granting ex parte approval; simultaneously buy 6–12 month protective puts (5–10% OTM) sized 50% of the equity position to cap downside if a contested case is filed.
• Buy ORCL call spread (ticker ORCL) 6–9 month expiry: long 10% OTM calls and short 25% OTM calls to express upside from increased AI compute demand while limiting premium outlay; target position size 1–2% of portfolio and reassess after public statements from OpenAI/Oracle within 90 days.
• If the MPSC signals a contested case or denial, initiate a short regional utility bond/muni basis trade: go short 5–10 year utility muni bonds or buy protection (CDS where liquid) for a notional equivalent to 2% portfolio if spreads widen >30–50 bp.
• Monitor 30–90 day catalysts (MPSC minutes, court filings, interconnection queue updates, and Related Digital permitting). If interconnection timelines extend beyond 12 months or water permits are denied, exit long DTE and rotate 2–3% into cloud/data‑center hardware suppliers or ORCL where revenue exposure is direct and less regulated.