A movement to ban data centers gains steam across the U.S.

Sen. Bernie Sanders will introduce legislation to bar construction of new data centers until Congress enacts federal AI safety and wealth-sharing rules, a move echoed by Rep. Alexandria Ocasio-Cortez and mirrored by local and state pauses in roughly a dozen states and dozens of municipalities. The proposal raises tangible regulatory and political risk for data-center and AI-related firms, with industry groups planning to spend "tens of millions" in politics and Anthropic underwriting pro-industry efforts with $20 million. Expect increased scrutiny and potential delays to data-center deployments that could affect regional energy demand, development timelines, and the political landscape ahead of the midterms.

Analysis

Local and state moratoriums on new data centers create a sharp, idiosyncratic liability for wholesale data-center landlords and tower-builders: projects that were under construction or in permitting can see cashflows deferred 6–24 months, turning queued capital expenditures into frictions that compress FFO and leverage metrics. Incumbent hyperscalers and cloud providers, by contrast, have optionality — surplus capacity in global footprints, negotiated long-term power contracts, and balance-sheet flexibility — so regulatory friction disproportionately reduces the relative value of commodity colocation vs. integrated cloud supply.

Second-order supply effects matter: generator and diesel backup demand, large transformer orders, and civil-construction windows are lumpy; a few states pausing builds can create bid/ask shocks for equipment OEMs and selective regional labor underutilization, compressing near-term bookings for firms with concentrated exposure. Semiconductor demand for high-end accelerators is path-dependent — a 6–12 month moratorium on new facilities won’t erase multi-quarter GPU backlogs, but it can move the inflection point for incremental unit absorption into 2027–2028 and magnify downside to consensus estimates if moratoria broaden.

Key catalysts to monitor are threefold and timeline-bound: (1) municipal/state ordinances in the next 3–9 months that set precedents, (2) federal action or preemption signals around the midterm election window (3–12 months), and (3) industry political spending and community concessions that can reverse momentum over 1–3 quarters. A practical arbitrage emerges between asset owners who rely on perpetual pipeline growth and large cloud operators who can internalize regulatory costs — the market has not fully priced in the asymmetric policy execution risk at the municipal level.

Key Decisions for Investors

• Initiate a 6–12 month relative-value pair: short Digital Realty Trust (DLR) -10% notional vs long Microsoft (MSFT) +10% notional. Rationale: DLR faces concentrated project/permitting risk and leverage sensitivity; MSFT benefits from optionality and pricing power. Risk/reward: if moratoria broaden, expect DLR underperformance of >15% vs MSFT; tail risk is federal preemption that narrows the spread quickly.
• Buy a 3–6 month DLR put spread (buy 1 ITM put ~10% down, sell 1 OTM put ~25% down) to cost-effectively express downside while capping premium. Timeframe: 3–6 months to capture municipal/state ordinance waves. Reward: asymmetric downside capture if local bans proliferate; risk: premium loss if legislation fizzles.
• Establish a 3–9 month long NVDA straddle (buy ATM call + buy ATM put) sized as a hedge against policy-driven demand variance for datacenter GPUs. Rationale: policy headlines and state moratoria can create >20% revenue re-velocity risk for GPU demand; options capture convexity. Risk/reward: paying theta for event convexity — hedge portion of AI exposure while keeping core long positions intact.
• Overweight regulated transmission/utility play with stable returns (example: NextEra Energy, NEE) on a 6–18 month horizon but size modestly. Rationale: utilities can negotiate recovery of sunk grid investments and extract higher value from existing load; downside is muted if load growth slows. Risk/reward: defensive yield with modest upside from regulatory rate cases; downside if long-term load growth structurally weakens.