A running log of new records, features, infrastructure changes, and data quality improvements to the AI Data Center Index.
+
Added
GeoJSON preview endpoint at `/api/geojson.json` returns the 25 largest geo-located facilities as a standard FeatureCollection — ready for Mapbox, Kepler.gl, QGIS, or any GIS tool
↑
Improved
Redesigned the homepage with a spatial layout, data-driven hero signals showing operational footprint, construction activity, and announced pipeline at a glance
New index ledger section surfaces total facilities, countries covered, operators tracked, and known capacity directly on the homepage
Key findings section highlights country concentration, operator concentration, the construction wave, and energy mix with links to drill deeper
Featured facilities grid expanded from 3 to 4 cards
Switched from npm to Bun for faster dependency installs and static builds
Tightened the visual design of data center record pages with sharper typography, rectangular tag chips, and inline metric cards
Refined the site-wide design system with subtler background gradients, softer shadows, and more precise nav spacing
✓
Fixed
SERP title truncation resolved site-wide by removing the redundant brand suffix from pages that already contain "AI Data Center" in their title, ensuring Google shows full titles with facility counts and capacity numbers
+
Added
Public API preview endpoints at `/api/countries.json` and `/api/operators.json` now expose country-level and operator-level market rollups, including facility counts, capacity totals, and status breakdowns
The `/datacenters/` directory now shows a status pipeline breakdown and a top markets by capacity view, with links into the matching status and country pages for deeper research
Public API preview endpoint at `/api/facilities.json` returns the 25 largest AI data centers by capacity with operator, location, status, and coordinates — giving developers and analysts a tangible sample before purchasing the full dataset
New `/datacenters/` directory page for browsing every tracked facility with filters for country, operator, status, energy type, and minimum capacity
`/map/` now includes a filterable facility table for browsing individual data centers
Operator pages now include portfolio briefs summarizing each operator's footprint
Capacity pipeline visualization on all taxonomy slice pages (countries, operators, continents, etc.) — stacked bar showing operational/construction/planned/announced MW distribution with percentages
`/map/` page — interactive AI data center map targeting map and location queries
`/hyperscale/` page tracking 500MW+ AI data centers, accessible from the nav More menu and command palette
`/new/` page covering the 2024–2026 AI data center buildout wave
Dynamic comparison pages: any two entities in the same dimension now resolve to a live comparison
5 more comparison pages: Meta vs AWS, Equinix vs Digital Realty, Nuclear vs Fossil energy, OpenAI vs Google, CoreWeave vs Google
10 comparison pages covering major operator matchups (Google vs Oracle, OpenAI vs Microsoft, Nvidia vs Google, AWS vs Microsoft) and country matchups (France vs Germany, India vs China, Australia vs Japan, and more)
Market-intelligence summary API at `/api/summary.json` with status breakdown, energy mix, top countries and operators by facility count
6 additional data center records: Oracle Cloud UAE North, Equinix LD8 London, Vantage Paris AI Campus, Google Cloud Warsaw, Microsoft Azure Hungary, Africa Data Centres Kigali
Changelog page at /changelog/ with structured timeline view of all product updates
9 new data center records across Europe, Japan, and the UK: OVHcloud Gravelines, Hetzner Falkenstein, Digital Realty AMS13, Sakura Internet Ishikari, Google Cloud Tokyo, KDDI Musashino, Microsoft UK West, Oracle UK South, and Ark Corsham
Tokenized search and country aliases in the command palette for faster navigation across all taxonomy dimensions and reports
Taxonomy links in the mobile card view of the datacenter table
Taxonomy links on the AI Focus column in the datacenter table and on hero chips for each datacenter record page
↑
Improved
Data center detail pages now use stronger search snippets that can include a facility's global capacity rank and a clearer call to the source-backed profile, helping the largest search surface explain its value more directly in results
Homepage headline and page title now lead with the tracked facility count and country coverage, making the index clearer in search results and on first view
Country, continent, operator, and other taxonomy slice pages now show live facility counts and known capacity totals directly in their page titles, with refreshed spatial data keeping those figures current
Country, operator, and continent hub pages now surface matching intelligence reports and quick links to the full directory, market comparisons, data access, and the index so visitors can move more easily from browsing into deeper research
All data center detail pages now include the `/datacenters/` directory in their breadcrumb trail and breadcrumb structured data, making it easier to move back to the full facility directory
The `/datacenters/` directory now lets you filter facilities by AI focus and minimum megawatts, and preview matching exports before requesting the dataset
Deploys now trigger a Radicchio remote build from the repo itself instead of relying on uploaded build artifacts
Smoke verification can now target a live deployed URL directly, which makes release checks work with the remote build flow
Added CollectionPage and ranked ItemList structured data to the `/datacenters/` directory so search engines can better understand the facility list
Rebuilt the map and location surfaces on a new spatial foundation, with a generated spatial artifact and registry-backed data pipeline behind the experience
Meta descriptions were refreshed across map, data center, and taxonomy pages to better match search intent
Sitemap coverage and route records now include the dedicated facilities directory page for list and database search queries
Country, operator, continent, and other taxonomy pages now link directly to relevant head-to-head comparison pages, making it easier to jump from category overviews into side-by-side research
Map filters now cover all countries and operators, and the result count stays accurate as filters change
Taxonomy hub pages now show visible FAQ sections, and FAQ structured data now covers taxonomy hubs, the About page, and all 344 data center detail pages
Homepage now shows a dataset freshness timestamp, uses a tighter mobile hero, and data center page titles now include clearer site branding for search results
Mobile homepage: tighter first-viewport experience — shortened lede, reduced action buttons to primary 3, signal pills and market sidebar now visible above the fold on 390px screens
Meta descriptions updated across all pages for better search click-through rates
Data export form shown inline on relevant pages — no external navigation required to request datasets
Report purchase flow: replaced mailto links with inline request form on all report pages and data access page — no mail client required, works on mobile, tracks submissions via FormSubmit.co
FAQPage structured data extended to `/market/`, `/pipeline/`, `/largest/`, `/locations/`, and all operator, country, and continent report pages
End-to-end test coverage added for the `/map/` and `/new/` pages
FAQPage structured data added to homepage, `/map/`, `/reports/`, `/hyperscale/`, `/new/`, `/compare/`, and all taxonomy slice pages
All 41 pre-built comparison pairings now visible; aws-vs-oracle gap filled; 5 more country pairings added
Dataset structured data added to `/data/`, meta descriptions refreshed across key pages, sitemap expanded
Homepage title retargeted to "AI Data Center Map & Database" to better match high-impression search queries for map and database use cases
`/data/` page now links to the summary API endpoint for programmatic access
SEO meta titles and descriptions added to capacity, AI focus, cities, and states taxonomy hub pages
Sourcing depth strengthened on 8 existing records: Digital Realty Georgia, ByteDance Johor, AirTrunk MEL2, Project Bunkhouse, Google Hamina, and AI workload details for Kuwait, Aragon, and Rio entries
Meta titles and descriptions sharpened across key pages for search click-through rate
Country slice market briefs enriched with deeper intelligence for key markets
Sitemap expanded with lastmod dates, feed pages, and 7 additional comparison routes
Metadata sharpened for map pages and taxonomy hub pages
Globe on the Locations page now lazy-loads on scroll; table sort state is synced to the URL
Map, list, and country hub page metadata sharpened for search intent
High-impression hub page titles and descriptions refreshed for ranking relevance
✓
Fixed
The `/datacenters/` directory now uses the correct megawatt field when calculating capacity totals
Map filtering smoke coverage was hardened to reduce flaky test runs
Deploy scripts and manifest paths were normalized for Linux runners, and the deploy workflow toolchain was corrected to prevent broken releases
Report request pages now fall back to a working inline submission flow instead of a broken mail link
Directory filter and hero search end-to-end tests are more stable, and a dead smoke test file was removed
Dataset JSON-LD corrected to reference the real API endpoint; flaky hero search end-to-end test stabilized
FAQPage JSON-LD removed sitewide to recover SERP click-through rate
Prerender concurrency reduced and Node memory cap lowered to prevent OOM build failures
Sourcing caveats strengthened on Khazna QAJ1 Ajman, Jamnagar AI mega campus, Ethiopia national AI factory, and Lordstown Stargate records
HTTPS canonical host enforced sitewide; smoke test contract refreshed to match
Data page API endpoints section now uses flex layout for proper visual alignment
+
Added
Hero search bar on the homepage for searching datacenters by name, operator, or location
Status legend on the SliceMap geographic visualization to explain color coding
Individual intelligence report pages for operators, countries, and continents
Nearby Facilities section on each datacenter record page showing geographically close facilities
11 new pre-rendered comparison pages covering regional and operator matchups with FAQ structured data
FAQ structured data and a visible FAQ section on all 344 datacenter record pages
FAQ structured data on taxonomy hub pages and the about page for rich snippet eligibility
Pipeline Tracker page showing the full AI data center development funnel from announced to operational
Pre-rendered comparison pages for high-value regional and operator matchups (14 pages total)
Search filter on taxonomy hub tables for faster navigation across large lists
Buildout wave visualization on the homepage showing year-by-year construction acceleration
Geographic distribution and operator breakdown sections on intelligence report pages
Error page with navigation links for better recovery from dead-end routes
Data readiness indicator on directory table entries to signal evidence quality
Breadcrumb navigation and structured data on about, feed, data access, and submit pages
Investment Tracker added to SiteNavigationElement structured data for search visibility
↑
Improved
Taxonomy slice pages now use a lightweight dot map instead of the 3MB globe, cutting page weight significantly
Comparison pages, pipeline coverage, and facility overview pages added to the sitemap
Homepage FAQs rewritten to target high-value search queries for better SERP click-through
DatacenterTable now has sticky column headers and an aggregate summary row
DatacenterTable cells link directly to taxonomy pages for faster navigation
Footer includes cross-links to Silicon Score, GGUF VRAM Calculator, and InferCost
Homepage meta title now targets "map" and "locations" queries for broader search coverage
Navigation consolidated to a single line with a More dropdown to reduce header clutter
Mobile homepage above-the-fold layout tightened for better first-impression density
Coverage page aligned to site design system with visual completeness bars
·
Changes
Added source count column to the directory table to show how many sources back each entry
·
Changes
Added keyboard navigation for all directory tables
Added AI focus filter to the directory and workload breakdown on the market page
Added data freshness signal to table views
Improved sitemap coverage and structured data across all pages
Added RSS autodiscovery link to the site
Improved JSON-LD dateModified across all key pages
·
Changes
Fixed 3D globe reliability: bundled globe.gl instead of loading from CDN
Improved globe performance: split vendor libraries into a separate chunk
·
Changes
Added intelligence report pages with preview and purchase flow
Added report call-to-action links on datacenter record pages
Added geographic footprint map to taxonomy slice pages
Added market concentration analysis section to the homepage
Added key market findings section to the homepage for featured snippet targeting
Added data-driven summary text to taxonomy slice pages
Added investment breadcrumbs and structured schema
Added print styles for PDF intelligence brief extraction
Added sortable Operator column to the directory table
Added mobile card views to the locations, directory, and investment tracker tables
Added slash-key shortcut to jump to search
Improved meta titles and descriptions for better search result click-through rates
Added FAQ sections with structured data to the largest, market, and locations pages
Improved initial load performance: isolated 1.5MB catalog data from the main app bundle
Fixed social preview image
·
Changes
Added suggested comparisons to the compare page empty state
Added geographic footprint and peer city ranking to the compare tool
·
Changes
Improved hosting reliability: migrated Radicchio API to Hetzner
·
Changes
Improved homepage: redesigned with clearer information architecture and aligned primary navigation
Fixed legacy URL routes: collapsed into compatibility redirects to preserve existing links
·
Changes
Added Nuxt-powered site: migrated all informational and taxonomy pages to Nuxt Vue components
Added data coverage transparency instrument and CSV export surfaces in the data hub
Added structured data markup to record and taxonomy pages for richer search results
Improved data quality: backfilled missing energy type and AI focus fields across 98 entries; migrated 160 legacy capacity fields to megawatt schema
·
Changes
Improved taxonomy page layout: unified slice surfaces across the site
Fixed build: removed broken dist symlink; excluded build output directories from version control
·
Changes
Added SEO meta tags and mobile filter sheet overlay
Fixed page title duplication; improved meta descriptions and added FAQ to the about page