FERC is expected to rule today on the single question every state ratepayer-protection bill is built around — who pays for the transmission upgrades data centers force on the grid — while the moratorium streak hits seven counties in seven days and a new exemption pattern appears in Utah, where Box Elder commissioners passed a 180-day pause but carved out the 9 GW Stratos project because it lives inside the state's Military Installation Development Authority and never went through county land-use review at all.

Why it matters

This is the single most important federal docket of the year for site readiness. Three downstream consequences to track. (1) If FERC adopts cost-shifting onto the requesting load, the state ratepayer-protection bills become redundant in their main goal and pivot to the harder questions (community benefits, water, noise), while site selection has to start carrying a federal interconnection-cost line item that doesn't exist today. (2) If FERC formalizes co-location, the Behind-The-Meter pathway becomes federally blessed instead of legally murky — that directly expands the BTM SKU surface Cliff is sized around. (3) Whatever FERC says about federal-vs-state jurisdiction is the part that will get litigated, and the litigation outcome is what actually matters for two to three years. The corpus needs to ingest the order text the moment it drops and tag every state proceeding that cites or relitigates it — the value of a regulatory graph compounds the most around new federal anchors, because every state docket below it will reference back.

Why it matters

Two operational implications. (1) The DC+BESS bundling pattern is now established in two states within 48 hours — that means the regulatory corpus has to be queryable by 'large industrial energy load' as a category, not just 'data center,' and Cliff's own land-and-expand path into BESS and other high-capex industrial categories is being validated from the opposition side as well as the demand side. (2) Iowa now has seven counties with active moratoriums, which is enough density that 'where in Iowa can a new data center actually be permitted in 2026?' becomes a hard parcel-level question — exactly the kind of cross-jurisdiction query a regulatory graph answers and a spreadsheet cannot.

Why it matters

Two signals worth tracking. (1) 18 months puts Gage County between the standard 6-12 month pause and the new 24-month long-window bans (Anderson TN, Jawando's proposed MD bill). That duration is long enough to kill a project's financing timeline outright — a corpus that ranks moratoriums needs to encode duration as a continuous variable, not a yes/no flag. (2) The pattern of a project's headline load being framed as a multiple of the host city's total load ('3x Lincoln,' the Tom Marge / Counterflow framing of 'two Manhattans plugged in next month' from yesterday's structural picture) is becoming the dominant resident-side argument. That makes load-as-multiple-of-host-city-load a number worth pre-computing per prospective site, because if a developer doesn't run the comparison, the local opposition will.

Why it matters

The Salix → Woodbury annexation pattern is now the second case in 48 hours of a sub-1,000-resident municipality being used as a jurisdictional vehicle to move farmland out of township/county control for a data center. That is a parcel-by-parcel question — annexation status, annexing municipality population, township-vs-county jurisdictional layer — that no human consultant tracks at a queue-wide scale. Two implications for the regulatory graph. (1) Annexation-as-of-date is now mandatory metadata on every parcel record for the moratorium tracker, because a project that closed annexation in April escapes a moratorium passed in June, but a project that hasn't yet is caught. (2) The pattern is asymmetric — annexation INTO a city escapes a county pause (Boone IN, Woodbury IA pending), but city-level pauses can sometimes be triggered by the same annexation (when the township pushes back). The corpus has to be able to answer 'who currently has jurisdiction over this exact piece of land, and which neighboring entity has paused permitting?' as a single composed query.

Why it matters

MIDA is a brand-new exemption pattern for the corpus and the most strategically important one to encode, because it points at a category of state-level special authorities the corpus is not currently watching. The same archetype exists elsewhere: industrial development authorities, port authorities, military adjacency zones, special economic development districts, even some state-charted technology authorities. The pattern is universal — a sub-state authority is given land-use approval power that bypasses normal county/city zoning. For Cliff that means three things. (1) Site-readiness scoring has to ask 'is any part of this parcel inside a state-level special authority?' as a first-order field, because if yes, county moratoriums are irrelevant. (2) Conversely, a developer can deliberately seek out parcels inside these authorities as a moratorium-proof siting strategy, and the corpus should be able to surface that inventory. (3) Tracking the political durability of each authority (MIDA is currently a target of a Republican primary fight per the Deseret News) is now part of the risk model — an authority that gets restructured by a state legislature can re-expose all of its projects to county-level review. None of this is in any spreadsheet today; it requires reading the actual enabling statute for each authority and tracking amendments, which is exactly the regulatory-text ingestion work Cliff is built around.