Facility directories are common; what is not is pairing them with interconnection-queue depth and live ISO telemetry, because that is what turns 'is there a site here' into 'when could it actually energise'. The per-ISO verdicts are the project's own scoring, not raw filings, so the methodology resource is worth reading before quoting one. Note also that queue depth for large loads exists only where the ISO publishes it, which is a smaller set than the map suggests.
A hosted MCP server for data-centre and power-grid intelligence. It combines a facility database, the physical infrastructure around it and live grid telemetry into one surface, aimed at site selection, energy analysis and market research — the questions where the answer is normally spread across a regulator's filings, an ISO's dashboard and a PDF.
- 18,800+ data centre facilities across 170+ countries with operator, capacity, location and fiber connectivity, and 126,000+ substations with voltage class and capacity estimates
- Live load and generation mix across the 7 US ISOs and 40+ EIA balancing authorities, plus Great Britain, 24 European ENTSO-E zones, Taiwan and Australia, refreshed roughly every 5 min
- Interconnection-queue snapshots with per-ISO BUILD, CAUTION or AVOID verdicts, including live large-load queue depth where the ISO publishes it
- Transmission lines, gas pipelines and fiber routes; NEPA filings for upcoming federal energy and data centre projects; tax incentives by state with eligibility
- 1,900+ tracked M&A transactions, an AI capacity index and a hyperscaler deal tracker; 300+ markets scored daily with the same BUILD/CAUTION/AVOID verdicts
- 82 MCP tools across facility search, market intel, grid and interconnection, renewables, site analysis, deals, fiber routing and infrastructure
- 13 guided prompts that appear as slash commands in the client — `/dchub:analyze-site`, `/dchub:pick-a-market`, `/dchub:power-availability`, `/dchub:site-report`, `/dchub:compare-markets`, `/dchub:fiber-plan`, `/dchub:grid_and_queue`, `/dchub:water_risk`, `/dchub:hyperscaler_activity`, `/dchub:whats_changed` among them
- Citable resources for the methodology and coverage behind an answer — `dchub://about`, `dchub://methodology`, `dchub://data-sources`, `dchub://coverage`
Nothing to start: add `https://dchub.cloud/mcp` over HTTP transport and anonymous access allows 5 calls/day with no key at all. An email signup gives a free key with 50 calls/day, and paid tiers run from Starter at $9/mo for 200 calls/day up to enterprise volumes. Data sources are named — EIA hourly RTO data, the HIFLD substation database, OpenStreetMap, PeeringDB, regulations.gov NEPA filings, USGS, EPA eGRID and FEMA NRI. The transport layer is open source; the data and business logic are hosted.
One command — npx -y @smithery/cli install @azmartone67/dchub --client claude
