Cliff Daily — May 19.

Why it matters

Denver is the second large city in the Mountain West to enact a moratorium in the last week (after Reno NV on May 14). The cluster is now real and the political pattern is durable: in any city with an existing visible data center footprint near residential neighborhoods, a one-year pause is now the default political ask, and the votes are not close. For Cliff's site-readiness map, the Denver moratorium effectively closes the city of Denver to new data center primary-use applications for 12 months, but Adams County (where the bulk of Denver-region data center activity already sits, e.g., the Aurora/Brighton corridor) is unaffected. Adams County is the substitute jurisdiction any blocked Denver project will route to, and Adams County has no moratorium discussion on the agenda. Add a 'substitute jurisdiction' column to the moratorium tracker — for each moratorium city, the relevant 5-county metro substitute is the actual go-forward siting target.

Why it matters

Charlotte is the first city this year where a moratorium proposal has been visibly slowed by an executive (mayor) rather than killed by a developer-side council majority — the council itself is split exactly 5-5. The pattern matters because it provides a counter-template: in cities with strong-mayor systems and split councils, the pause-vs-build fight can be slowed at the procedural level for weeks or months without ever being explicitly defeated. For Cliff's framework, the takeaway is that 'is there a moratorium' is now a 3-state field, not a 2-state field: {moratorium adopted, no moratorium in motion, moratorium in motion but procedurally slowed by mayor or other executive}. The third state is now significant enough to be tracked separately because it implies a multi-week regulatory window during which the developer can still move forward with permitting but the moratorium risk is non-zero. The actionable variable is the May 26 hearing and the June 8 earliest-adoption date — any North Carolina site under consideration this month should price in a 60% probability of a 150-day Charlotte moratorium being in force from mid-June onward (based on the 5-5 split, the mayor's procedural-not-substantive 'no,' and the general direction of NC HB 6 debate at the state level).

Why it matters

Cheyenne is the most informative outcome of the four because it isolates one variable: when the local construction workforce is large enough to dominate council public comment, moratorium proposals stall at the committee level. This is a real structural variable for any metro with high data center construction concentration — Loudoun County VA, Quincy WA, Council Bluffs IA, and the Phoenix West Valley would show similar dynamics. For Cliff's framework, add a 'data center construction workforce as % of metro construction employment' field; values above ~15% likely correlate with weak moratorium passage probability. The bigger story: 'moratorium adopted' is not a regional phenomenon driven by climate / political ideology; it's a hyperlocal phenomenon driven by whether existing data center facilities are visible to residential neighborhoods AND whether the local construction workforce shows up to public comment. The Mountain West cluster narrative from yesterday is correct in direction but overstated — Denver passed, Reno passed (different state), Cheyenne stalled. Three different outcomes in three cities in one week.

Why it matters

The Cave City pattern (small town, pre-emptive moratorium triggered by mere developer interest, fast 4-1 first reading, expected second-reading pass within 48 hours) is the procedural archetype the moratorium-advocacy network is now exporting. There is now a working template for any rural town councilor who wants to pre-empt data center siting before the developer even files an application — a 30-day-to-effect path from 'developer expresses interest' to 'one-year pause enforceable.' For Cliff's framework, this changes the site-readiness math for rural sites in unfamiliar jurisdictions: 'no current moratorium' is no longer a reliable indicator that a site is open for development, because a moratorium can now be enacted in under a month from the moment the developer's interest becomes public. This implies a much stronger argument for keeping site interest confidential through the early diligence phase (which most data center developers already do via shell-company land options, but small-town councils are increasingly able to identify the underlying developer from public records). Add 'developer-disclosed interest → moratorium' as a tracked metric — the average days from disclosure to first-reading moratorium is now ~30 in Kentucky and trending shorter. Note: 'Discovery' as the developer name needs primary-source verification — could be a code-named LLC or a misreport; flag for follow-up before any GTM artifact references it.

Why it matters

Two updates. (1) The hyperscaler 'build your own power plant and put the data center next to it' strategy is the single biggest economic alternative to waiting in PJM's transmission queue, and it remains in regulatory limbo. The accepted parts of the February filing are favorable enough that some configurations (large plant + smaller grid-export service request) are workable today, but the rejected parts mean the definition of 'co-located' itself is still contested. Any PJM-zone build-vs-buy-power analysis for data center customers needs an 'FERC ruling pending' caveat in the cost model. The cost difference between buying transmission service at full PJM rates vs. avoiding it via co-location is on the order of $40-80/MWh delivered (a real number — PJM transmission service for new large loads runs $30-60/MWh on transmission alone, plus capacity at $333/MW-day for 2026/2027 ≈ $24/MWh on a 60% load factor). For a 500 MW data center, that's $175M-$350M per year of tariff cost exposure. (2) For Cliff's regulatory tracking, the FERC co-location proceeding (EL25-49 et al.) is the single highest-stakes federal docket in the data center space and the most likely to produce material site-readiness changes for any PJM customer in 2026-2027. Add the FERC docket to the live-docket-intelligence wedge — this is exactly the kind of multi-stage federal proceeding that the regulatory knowledge graph is designed to track in real time, with each filing, comment, and order generating a structured event that affects downstream site economics. The next inflection points: (a) protest/comment deadline on PJM's May 18 filing (typically 21 days after filing, so ~June 8); (b) FERC ruling on the new compliance (60-180 days); (c) any rehearing requests after FERC rules.

Why it matters

Texas is the highest-velocity state for data center interconnection volume (238.6 GW in queue, 225 new requests in 2025), and SB 6 + the PUCT rule + ERCOT Batch Zero are the three layers that together define the cost and timeline of any new Texas data center. For Cliff's beachhead wedge — Layer 5 live-docket + interruptible-load — this is the central regulatory artifact and the single richest source of structured filing data in the next 12 months. PUCT's draft rule had a closed comment period (April 17), so the next inflection point is final rule adoption. Once final, every developer in the ERCOT queue has to comply, which generates a high volume of filings, financial security documents, and site control disclosures — all structured, all primary-source, all queryable. The non-refundable per-MW fee structure ($50,000-$100,000/MW) creates a clean economic discriminator: it filters speculative requests out of the queue, which means the surviving queue is much higher-quality information about real data center demand. Update strategy/ercot-pclr-batch-zero.md with the SB 6 rule status: comment period closed April 17, final rule pending, PUCT is the implementing regulator, and any Texas siting analysis post-rule-adoption needs an SB 6 financial-security column in addition to the existing ERCOT Batch Zero PCLR column.