The News section surfaces recent news articles mentioning US power plants and interconnection queue projects. Articles are sourced from Google News via the Serper API, then deduplicated into stories, classified by an LLM into categories, and enriched with entity extraction. Available on plant detail pages, project detail pages, and as a dedicated news hub at /news.
Coverage
News articles are fetched for plants with 10 MW or greater nameplate capacity and for active interconnection queue projects. Larger plants (100 MW+) use exact-match search queries for precision; smaller plants use broader queries for recall. Currently covers ~8,000 plants and ~900 queue projects with ~112,000 articles grouped into ~39,000 unique stories.
Story clustering
The same news event is often reported by multiple outlets. The platform groups articles about the same story into clusters using normalized title matching. Within each cluster, one article is marked as the primary (displayed to the user) and the rest are treated as sources. A source count badge (e.g., "3 sources") appears on articles covered by multiple outlets, with a collapsible list showing all source URLs.
This deduplication reduces ~112,000 raw articles to ~39,000 unique stories — eliminating repetitive noise while preserving source attribution for verification.
Classification
Each article is classified by an LLM into one of six categories, plus subcategories and named entity extraction.
Deals
PPAs, M&A activity, project finance, green bonds, tax equity, and debt financing.
Hazards
Wildfires, hurricanes, tornados, hail, flooding, extreme heat/cold, and ice storms affecting plants.
Regulatory
FERC orders, state PUC decisions, federal policy changes, permitting updates, and rate cases.
Grid
Outages, curtailment events, interconnection queue updates, and capacity additions.
Industry
Earnings reports, partnerships, technology announcements, and corporate news.
Development
Project milestones, construction progress, delays, EPC contracts, and commercial operation announcements. Primarily covers queue projects and plants under construction.
Entity extraction identifies companies, agencies, and people mentioned in each article along with their role (buyer, seller, developer, regulator, etc.). These entities power the "Entity Mentioned" filter in the news hub sidebar.
Plant and project news
On each plant's detail page, the News section shows up to 4 recent articles with category badges, entity pills, source count badges, and publication date. Category filter pills appear when a plant has articles across multiple categories. A "View all articles" link navigates to the news hub filtered to that plant.
Queue project pages also show relevant news articles. The pipeline separates plant-level news (mentioning operating facilities) from project-level news (mentioning development, permitting, and construction activity) using the article's primary entity context. Currently 909 queue projects have associated news.
News hub
The dedicated news page at /news provides a full-featured news feed with a sidebar containing 9 independent filters.
Category & Subcategory
Filter by classification category. When a category is selected, its subcategories appear inline for further refinement.
Entity Mentioned
Filter by companies, agencies, or people extracted from article text. Searchable list of the most frequently mentioned entities.
Owner / Operator
Filter by the plant's owner or operator (from the ownership dimension, not article text).
Other filters
Fuel type, state, ISO/RTO, date range, plant capacity (MW), and news source. All filters combine with AND logic.
Data sources
Articles are fetched from Google News via the Serper API. Classification and entity extraction are LLM-based. Articles are refreshed monthly; classification results are cached and persist across rebuilds. Story clustering and deduplication run automatically during each build.
