Three days to 'June 18 day' — PUCT Batch Zero vote and Pine Island injunction hearing both land on the same calendar day, while EPA's 'Begin Actual Construction' rule and PJM's reliability-backstop FERC filing reshape the federal floor underneath them.

Why it matters

Cliff's stated near-term beachhead is Layer 5 — live docket intelligence + interruptible-load math, driven by the April–July 2026 ERCOT deadlines. June 18 is the decisive event that converts the Batch Zero framework from 'pending' to 'effective.' Three concrete product moves: (1) The wiki page for every Texas docket-tracker entry needs a status flag that flips from 'pre-vote' to 'effective' the moment PUCT closes the meeting Thursday — schedule a refresh job to run Thursday afternoon, not the next-morning batch. (2) Any developer who is on the fence about PCLR election now has 36 days from June 18 to July 24 to make a binding decision — Cliff's PCLR decision-tree page (the one drafted yesterday) needs an explicit countdown widget so a developer reading it on July 20 sees '4 days to PCLR deadline.' (3) The PUCT order will be public the same day as the vote; the order's exact language on deliverability methodology is the kind of regulatory text the LLM-ingested wiki should absorb within hours, not weeks — that's the test of whether the ingest pipeline actually delivers the 'live docket intelligence' promise.

Why it matters

The Pine Island ruling is the single most important regulatory event for hyperscale data centers in mini-NEPA states this quarter. If MCEA wins, three things follow within 30-60 days: (1) Plaintiff lawyers file copycat AUAR / mini-NEPA challenges on the top 10-15 hyperscale projects in WA, OR, CA, MN, NY, MA — Cliff's regulatory knowledge graph should pre-flag which Cliff-tracked sites would be vulnerable to the same theory (water-use modeling, cumulative-impact analysis, customer-disclosure timing). (2) The 'water + power + cumulative-impacts' analytical framework MCEA used becomes the standard plaintiff theory — Cliff's mini-NEPA-state site pages need an explicit 'AUAR-challenge exposure' line item before this becomes the consensus underwriting variable. (3) The MCEA win cements the prediction in yesterday's note that moratorium pressure routes into mini-NEPA challenges in Home Rule states — and creates a discrete SKU opportunity ('review-challenge audit') for the 30 most exposed sites. If MCEA loses, Pine Island becomes a defensive case study and the wiki page describes how Pine Island defended the AUAR — useful but less commercially urgent.

Why it matters

Air-permit / interruptible-load math is one of Cliff's named structurally-empty layers — and this rule is the single biggest federal-level change to that math since the 1980 NSR regulations. Three immediate product moves: (1) The comment period closes ~June 25; Cliff should track which trade associations, hyperscalers, and environmental groups file comments — the comment record itself becomes part of the regulatory knowledge graph and gets cited in any final-rule challenge. (2) Every Cliff air-permit page on a BTM gas project needs a new 'proposed-rule timing scenario' that re-runs the construction schedule under the EPA proposal vs. the status quo — for the Ohio Socrates South template, that scenario probably shows 6-9 months earlier substantial completion. (3) The proposed rule is a meaningful piece of evidence for Cliff's 'AI-native' positioning thesis: a Trinity-tier consulting firm has to manually re-read every air permit, EPA guidance memo, and state delegation agreement to figure out what 'begin actual construction' means in each state — that's exactly the unstructured regulatory text Cliff's LLM ingestion can absorb in one pass. The thesis itself becomes more demonstrable the moment the rule finalizes.

Why it matters

PJM's RBP is the first concrete example of a data-center-specific capacity mechanism in a US ISO — every other ISO will reference it, copy parts of it, or define themselves against it. Three Cliff-specific moves: (1) The Phase I bilateral-contract window creates a new SKU: 'PJM bilateral counter-party matching' — Cliff already tracks which new generation projects are progressing in the PJM queue, and which data center sites are at which interconnection stage; the new product is the brokerage-grade match. Worth scoping next week whether that's a Cliff feature or a Helix-Digital function (today's note's competitive context). (2) For the regulatory knowledge graph, the PJM RBP filing itself — the actual FERC docket text — is the kind of dense, multi-hundred-page regulatory document where LLM ingestion compounds value: hyperscaler developers will need a structured summary within hours, not the standard 2-week law-firm memo cadence. (3) The FERC RM26-4 ruling (expected end of June) and the PJM RBP filing may end up cross-referencing each other; the wiki needs to track them as linked pages, not separate entries — that linking-of-related-dockets-into-one-narrative is what generic AI assistants can't do well.

Why it matters

Nassau County extends the moratorium tracker into Florida, which had been underrepresented vs. Texas, Virginia, and the Midwest. Two regulatory-graph updates: (1) The 'charter vs. non-charter Florida county' distinction is the equivalent of the Dillon's Rule vs. Home Rule distinction in other states — Cliff's per-state legal-authority page needs Florida split into the 20 charter counties and the 47 non-charter counties, because moratorium durability follows that line. (2) The inclusion of 'cryptocurrency mining operations' in the same ordinance language is the pattern most non-coastal counties will copy — the political coalition opposing data centers includes residents opposing crypto for separate reasons, and bundling the two raises the political ceiling on these ordinances. For the wiki, every county-level moratorium page should track whether crypto is bundled into the prohibition language.

Why it matters

Monterey Park is the first permanent voter-enacted data center ban, but it will not be the last — California, Oregon, Washington, Arizona, and Colorado all allow citizen-initiated General Plan amendments, and once the Monterey Park template exists, organizing groups will replicate it. Three implications for Cliff's site underwriting: (1) Every CA / OR / WA / AZ / CO site within 50 miles of a population center with active anti-data-center organizing should now carry an explicit 'voter-initiative risk' line item — the 6-12 month qualification timeline means a developer with a permitted but unbuilt site is exposed to a permanent ban midway through construction. (2) The moratorium tracker now needs four distinct status fields rather than one; the visualization above is the right schema to ship into the regulatory knowledge graph this week. (3) Cliff's near-term beachhead — air-permits + interruptible-load math in ERCOT and PJM — is largely insulated from the voter-initiative pattern (Texas and most PJM states don't have citizen-initiated General Plan amendments at the city level). That's a useful structural reason the beachhead is durable while the West Coast site-selection layer becomes more uncertain — and a talking point for the strategy memo this week.