Google weighs joining Van Buren Township data center project

Google is evaluating whether to join Project Cannoli, a hyperscale data center in Van Buren Township being advanced by Panattoni Development. The company said it is working with the project's developer but has not committed and provided no timing, investment size or approval details. Confirmation would support demand for hyperscale data center capacity and the local development, but market impact is limited until a formal commitment is announced.

Analysis

If a hyperscaler commits to a greenfield campus, the immediate second-order winners are hyperscale-focused hardware and networking vendors and heavy electrical contractors rather than generic real estate landlords. A 100–200 MW build typically represents $0.8–1.5bn of capital spend and the procurement of tens of thousands of servers/accelerators and megawatts of switching/transform gear over 12–36 months, concentrating demand into a small set of suppliers (Arista, Cisco, select transformer/UPS vendors) and shortening lead times across the supply chain.

The main risks are binary and horizon-staggered: municipal permitting and transmission upgrades (3–12 months of uncertainty) and then multi-year construction/commissioning (12–36 months) that can be derailed by capex retrenchment or supply-chain shocks. A reversal is most likely from a) utility capacity constraints or b) a corporate capex pullback if AI/cloud demand growth cools; either would favor REIT lessors in the short run and penalize vendors exposed to forward bookings.

Consensus frames this as uniformly bullish for the anchor tenant; the overlooked angles are local fiscal and labor effects and the landlord-versus-owner dynamic. Locally, large builds can tighten skilled labor and housing for 24–36 months, pressuring municipal services and potentially increasing near-term municipal spend; strategically, every hyperscaler that self-provisions reduces available leasing demand growth for public data-center REITs, but REITs with diversified tenants and edge footprints could outlast a single hyperscaler’s footprint decision.

Key Decisions for Investors

• Go long GOOGL via a 12–36 month LEAPS call spread (buy Jan-2028 LEAPS, sell a higher strike) — rationale: asymmetric upside from cloud/AI infra economics while defined downside via sold leg; target 2–4x return if Google locks multiple campus builds, risk = option premium (~100% loss of premium if thesis fails).
• Pair trade: long GOOGL equity (or calls) / short DLR or EQIX (equal notional) for 12–24 months — thesis: ownership-led hyperscale builds compress future leasing demand; target 20–30% relative outperformance, stop-loss if REIT spreads tighten by >150bp or REIT reports >5% same-store lease expansion.
• Long ANET (or buy 6–18 month call options) — exposure to hyperscale switching demand concentrated in the near term; upside if campus procurement accelerates, downside capped to equity move/option premium.