$15bn Stargate Data Center Campus in Port Washington, Wisconsin to be Kneecapped as Residents Take to the Ballot

A $15bn Stargate AI data center campus (~1.3 GW capacity) in Port Washington, WI is facing a local referendum today that would require voter approval for city tax incentives, likely hampering future development though not outright cancelling the project. Backed by OpenAI, Oracle and federal support, the project touts ~4,000 construction jobs and sustainability features, but rapid grassroots organizing (~1,000 signatures) signals heightened permitting and political risk that could set a national precedent for opposing AI data center megaprojects.

Analysis

This vote is not a local curiosity — it is a template that converts permitting risk into electoral risk, which raises the effective hurdle for megaprojects. Expect developers to factor in an extra 12–24 months of schedule friction and to price an incremental 3–5 percentage-point financing/WACC premium for projects in politically contested jurisdictions; that uplift propagates directly into lower project IRRs and higher demand for pre-commitments from hyperscalers.

Second-order demand effects are asymmetric and regional. A cascade of successful referenda would depress near-term orders for site-specific capex (generators, substations, transformers, large cooling systems) and accelerate scarcity-driven rent appreciation for existing colo capacity — localized rents could reprice up 5–10% as buyers seek immediate installed MW. Conversely, states with pre-emptive legislative regimes (or with stronger pro-development political majorities) become incremental winners for siting, shifting growth, labor demand and tax-revenue flows there.

For coalition backers like Oracle, the immediate hit is reputational and political — not necessarily technical — raising the probability of contingent costs (community benefits, tariff concessions, legal fees) over 12–18 months. The market may overshoot: while projects rarely die overnight, the re-pricing window (and an opportunity to trade it) is short-lived — anticipate a 3–9 month period of elevated volatility as developers re-contract deals and re-file incentive requests.

Key Decisions for Investors

• ORCL — tactical 6-month put-spread (buy 6m 10% OTM puts, sell 6m 30% OTM puts) sized 1–2% NAV. Rationale: capture asymmetric downside if political contagion forces re-pricing of developer obligations while financing lines are re-cut. Max loss = premium; payoff if multiple referenda or costly concessions emerge within 6 months (~3–6x premium).
• EQIX / DLR / QTS — selective long (2–4% NAV) in established colo REITs with diversified, operating MW. Use 6–12 month horizon or buy 9–12m call spreads if leverage desired. Rationale: constrained greenfield supply should support existing rent capture; target regional rent repricing of +5–10%. Risk: hyperscaler demand remains lumpy and could pull back if macro slows.
• NVDA / AMD — short-duration hedges not outright shorts: buy 3–6 month puts sized to cover existing long exposure (10–25% of position). Rationale: near-term moderation in new-campus orders can dent quarterly accelerator demand; this is a hedge against policy-driven timeline shocks. Risk: long-term secular AI demand remains intact, so keep exposure small and time-limited.
• Event trade — monitor municipal ballot pipeline: if 2+ successful referenda occur in adjacent states within 9–12 months, increase ORCL put exposure and add long positions in colo REITs; if no contagion after 12 months, unwind puts and trim REIT longs. Rationale: this sequenced trigger converts qualitative political risk into a concrete trade cadence.