Trump-backed hopeful to succeed DeSantis unveils plan to protect local water, electricity from data centers

This is less about one bill than a pricing signal: the marginal AI watt is moving from an embedded utility subsidy to a privately financed capital item. That shifts economics from regulated utilities to hyperscalers and their power partners, and it raises the hurdle rate for new data-center capacity in politically sensitive metros. The immediate market risk is not total AI demand, but slower site approvals and higher all-in cost per MW, which can compress returns on smaller/late-cycle AI projects first.

The second-order winners are vendors that help companies self-generate or harden supply: grid equipment, switchgear, gas turbines, backup power, and power-management names. The losers are utilities that have been marketing AI load growth as a visible rate-base tailwind; if more states emulate this stance, some of that demand becomes off-grid and less accretive to regulated earnings. Data-center REITs are a more mixed read: higher tenant costs can support pricing for scarce capacity, but permitting friction and local moratoria can slow absorption and delay new leasing.

Over the next 1-3 months, the catalyst path is legislative copycat risk in other swing states and whether any federal language starts looking enforceable rather than aspirational. Over 6-18 months, the structural issue is social-license risk: even if AI capex remains enormous, the buildout may tilt toward fewer, larger campuses with dedicated power, favoring incumbents with land, interconnects, and balance sheet. The consensus may be underestimating how quickly political pressure can force private power procurement into the model, but it may also be overestimating the chance of a near-term broad slowdown in AI spending; the pain is likely more granular and regional than sector-wide.