How a Crusoe-Google campus became Texas’ co-location test case

Key event: Texas PUCT adopted strict net-metering rules for co-located data centers (threshold: 75 MW) and is using Crusoe/Google’s Goodnight campus as the first test case; the campus pairs a 265.5 MW Serena wind farm with a wider buildout of 933 MW of off-grid gas turbines (expected ~4.5M tCO2/year). New rules demand 100% curtailment within 30 minutes on grid emergency, prohibit participation in paid demand-response, and impose financial liability for 'underutilized' transmission assets (defined as ≥25% lower expected transmission). Crusoe’s co-located portion was sold to Ensign in mid‑March; developers warn these requirements materially worsen project economics, risk stranded assets, and will disincentivize co-location, while regulators argue the trade-off preserves reliability and protects ratepayers.

Analysis

Texas’ hardline co-location rules reprice the optionality of behind-the-meter arrangements: developers who counted on fast, low-cost grid access now face a multi-year liability tail tied to utilization and forced curtailment exposure. That shifts competitive advantage to firms with balance-sheet capacity to absorb transmission attrition or to vertically integrated players who own dispatchable assets — expect spreads in valuation multiples between well-capitalized incumbents and asset-light entrants to widen over 6–24 months.

A second-order beneficiary is the market for fast-acting grid services (storage, synthetic inertia, contracted ancillary services). By excluding co-located loads from paid demand response and creating uncompensated curtailment obligations, the rules raise the marginal value of firm, dispatchable capacity and contracted reserves in ERCOT; this will compress merchant returns on intermittent projects unless paired with storage or firming contracts within 12–36 months.

Catalysts that could materially change outcomes are binary: (1) FERC or federal DOE intervention that standardizes a less punitive net-billing framework (3–18 months), and (2) a high-profile litigation or commercial renegotiation that forces compensation for curtailments (months to years). Tail risks include rapid load-growth triggering reliability events that lead to even stricter curtailment enforcement or, conversely, grid investments that reduce the 25% underutilization cliff, materially altering developer economics.

Key Decisions for Investors

• Long AES (AES) or NextEra (NEE) — 12–24 months: buy shares or 12–18 month call spreads to capture upside from higher demand for storage and contracted ancillary services; R/R: moderate upside (15–30%) vs execution/regulatory risk if policy shifts blunt merchant opportunity.
• Pair trade — short CyrusOne (CONE) / long Equinix (EQIX) — 6–12 months: CONE is more exposed to marginal co-location projects and utilization liabilities while EQIX has pricing power and can internalize transmission costs; target 1:1 notional, take profits at 20% net move or cut losses at 12%.
• Long regulated transmission/utility exposure (AEP) — 6–18 months: buy AEP to capture potential transmission upgrades and reduced stranding risk as the state funds reliability; expected defensive upside 8–18% vs regulatory execution risk.
• Event hedge: buy 6–12 month puts on a data-center developer ETF or concentrated names with heavy co-location exposure (where available) to protect against adverse PUCT rulings or a surprise enforcement action — pay small premium for asymmetric downside protection over the next 3–9 months.