A ragtag Wisconsin group is leading America’s first anti-data center referendum

Port Washington voters will weigh a referendum that could require voter approval before creating tax incremental districts (TIDs) with a base value of $10M+, potentially constraining future data center incentives tied to the proposed $15B, 1.3GW 'Stargate' campus. The project partners (OpenAI, Oracle, Vantage) cite 4,000 construction jobs, ~1,000 permanent jobs and $175M in local infrastructure investment; opponents cite concerns over water, noise, grid impacts and transparency. The measure faces an MMAC lawsuit and possible state-level legal challenge, and similar data-center referendums are emerging in other US municipalities, creating regulatory and execution risk for local AI/data-center deployments.

Analysis

A wave of municipal referenda elevates siting and political risk from a nuisance to a quasi-permanent margin in hyperscale deployment economics: expect a measurable increase in developers’ upfront non-capex costs (legal, PR, local infrastructure contributions) that will compress near-term returns on new campuses by an incremental 200–400bps of IRR for marginal projects. That favors vendors and cloud players who can internalize or monetize that overhead (software, managed services, hardware leasing) over pure play colo models reliant on rapid greenfield builds.

Second-order supply effects are underappreciated: if marginal projects concentrate into fewer permissive states, land, labor, and power capacity will see localized price inflation while transmission and infra contractors win multi-year work; conversely, utilities face political pressure to push costs onto developers or to secure larger long-term PPAs, raising the likelihood developers absorb more grid upgrade capex (shifting capital intensity back onto hyperscalers). Expect a bifurcation over 6–24 months between vertically integrated cloud providers that can flex capacity and those that must build third-party campuses.

Legally, this is a volatility amplifier with short, medium, and long arcs: court rulings on the validity of referendum expansion can move sentiment within days/weeks, legislative pre-emption or standardized statewide rules can resolve uncertainty over 6–18 months, and persistent local resistance will meaningfully slow capacity growth on a multi-year horizon if not countered by policy. Tactical positioning should therefore be event-driven and asymmetric: hedge near-term exposure to project delays while maintaining optionality for the secular AI infrastructure demand that remains intact on a 2–5 year view.

Key Decisions for Investors

• Pair trade (3–9 months): Long ORCL equity or 3–6 month ATM calls vs short MSFT via a 3–6 month put spread. Rationale: ORCL's software/hybrid infra revenue is less sensitive to greenfield siting delays; MSFT's valuation premium is more exposed to near-term capex/capacity timing. Target relative outperformance of 8–12%; size to no more than 1–2% portfolio risk; stop if pair moves against you by 4% absolute on portfolio exposure.
• Short AMZN (3–6 months): Buy a protective put spread (e.g., 3–6 month ATM/10% OTM) equal to 0.5–1% portfolio risk. Rationale: AWS faces incremental reputation and capex friction that can compress growth multiple in the near term. Risk/reward: limited downside via spread; payoff asymmetric if multiple municipal votes or legal defeats slow new capacity pacing over the next two quarters.
• Selective long GOOGL (6–12 months): Initiate a modest overweight or buy 6–12 month calls sized to 1–2% portfolio. Rationale: Google’s integrated build-and-operate model and diversified revenue buffer it against isolated siting fights; secular AI demand preserves multi-year growth. Target 10–15% upside; hard stop at 6% adverse move.
• Event hedge (days–weeks around catalysts): Ahead of court rulings or municipal ballots, buy a small volatility basket: long ORCL calls and GOOGL calls and long puts on MSFT/AMZN for 30–90 day tenors. Rationale: captures asymmetric upside from policy clarity while capping cost of downside protection; size as tactical overlay (0.5–1% portfolio).