5 jurisdictions / 4 days

Why it matters

Three updates and one trend-line. (1) Baltimore's <10 MW carveout is the first observed case of a major US city pre-architecting an exemption that lets a specific institutional applicant (Johns Hopkins) advance during a moratorium period. This is structurally different from the moratoria covered yesterday and previously — most prior moratoria are uniform, applying to all proposed data centers without exception. The Baltimore template ('moratorium with named-or-effectively-named institutional carveout') gives Cliff's customers a tactical playbook for hostile-jurisdiction sites: if a co-tenant or anchor is an institutional user (university, hospital, government) at sub-threshold MW size, structuring the campus to qualify for the carveout becomes a viable approval path. The 10 MW vs 25 MW threshold negotiation also matters — at 10 MW, only small enterprise / institutional projects qualify; at 25 MW, mid-size colocation facilities qualify too. Cliff's regulatory-knowledge graph needs to track moratoria with carveouts as a separate sub-tier (alongside flat moratoria and ban-pathway moratoria like Bulloch GA from yesterday). (2) Northampton County's 32-month duration is the longest county-level moratorium observed to date — longer than Sedgwick KS's two-stage 90+90, longer than Apex NC's 12, longer than Bulloch GA's December 2026 (which is ~20 months from passage). The 32-month duration is structurally similar to a *de facto* ban for a typical hyperscale planning cycle: most active hyperscaler site-selection cycles run 12-18 months from initial site identification to contracted offtake, and a 32-month pause forces re-routing rather than waiting. Cliff's regulatory-risk module should treat any moratorium of 24+ months as 'effective ban' for hyperscaler underwriting, even if the statute is technically temporary. (3) Ferguson MO's 3-3 tied vote with one abstention is also a fresh data point — the 'tied vote kills the project' outcome is functionally similar to a No vote for the developer, but the procedural posture (tied vote) leaves the door open for a future re-vote with a single-vote shift. SSL Investments' $1.8B proposal is now stalled rather than dead, and the developer's next move (reconsidered terms, additional concessions, possibly walking away) is a watch item over the next 30 days. (4) The trend: 14 net-new bans March-April + 8 jurisdictional actions May 6-7 + 5 more in 4 days (this period) = the cluster cadence is now confirmed at roughly 4-6 actions per week and accelerating. Cliff's daily monitor should officially move from 'daily batch' to 'rolling watch' for moratorium tape; the cadence is now too high for once-a-day catch-up to be timely for any customer who's mid-application in a state where neighboring counties are moving.