PNNL Data Center Atlas API and MCP server

API access

Your AI client reaches PNNL Data Center Atlas through one ParcelOps MCP endpoint, so there is no separate key to hold for it and no per-source client to write. Reaching it takes no credentials of your own.

What this source cannot tell you

• ⚠ Anything outside Contiguous United States coverage represented in the published atlas. A request past that edge comes back empty rather than approximated.
• ⚠ Anything after 2035.
• ⚠ Anything PNNL/DOE IM3 Open Source Data Center Atlas does not publish. ParcelOps returns this source as it stands and never fills a blank field from a different one.
• ⚠ Existing inventory is compiled partly from OpenStreetMap and may contain omissions, overlaps, or duplicate representations.
• ⚠ Projected records are scenarios, not announced projects, forecasts of a specific operator, or proof of future utility capacity.
• ⚠ Existing inventory is ODbL with attribution/share-alike obligations; projected scenario data is CC BY 4.0 with attribution obligations.

Methods (3)

A method is one call your agent can make against this source.

• searchExistingDataCenters

  read

  Search the published existing data-center points, building footprints, and campuses.

  FreshnessScheduled (source-level)
  Input schema

  {
    "type": "object",
    "properties": {
      "state": {
        "type": "string",
        "pattern": "^[A-Za-z]{2}$"
      },
      "county": {
        "type": "string",
        "maxLength": 120
      },
      "operator": {
        "type": "string",
        "maxLength": 200
      },
      "name": {
        "type": "string",
        "maxLength": 200
      },
      "facilityType": {
        "type": "string",
        "enum": [
          "point",
          "building",
          "campus"
        ]
      },
      "minSquareFeet": {
        "type": "number",
        "minimum": 0,
        "maximum": 1000000000
      },
      "bbox": {
        "type": "array",
        "description": "[west, south, east, north] in WGS84.",
        "minItems": 4,
        "maxItems": 4,
        "items": {
          "type": "number"
        }
      },
      "point": {
        "type": "object",
        "properties": {
          "lat": {
            "type": "number",
            "minimum": -90,
            "maximum": 90
          },
          "lon": {
            "type": "number",
            "minimum": -180,
            "maximum": 180
          },
          "radiusMiles": {
            "type": "number",
            "minimum": 0.1,
            "maximum": 500
          }
        },
        "required": [
          "lat",
          "lon",
          "radiusMiles"
        ],
        "additionalProperties": false
      },
      "limit": {
        "type": "integer",
        "minimum": 1,
        "maximum": 200
      },
      "offset": {
        "type": "integer",
        "minimum": 0,
        "maximum": 5000
      }
    },
    "required": [],
    "additionalProperties": false
  }

• searchProjectedDataCenters

  read

  Query one explicit 2035 growth and market-gravity scenario.

  What you must providegrowthScenariomarketGravityWeight

  FreshnessScheduled (source-level)
  Input schema

  {
    "type": "object",
    "properties": {
      "growthScenario": {
        "type": "string",
        "enum": [
          "low",
          "moderate",
          "high",
          "higher"
        ]
      },
      "marketGravityWeight": {
        "type": "number",
        "enum": [
          0,
          25,
          50,
          75,
          100
        ]
      },
      "bbox": {
        "type": "array",
        "description": "[west, south, east, north] in WGS84.",
        "minItems": 4,
        "maxItems": 4,
        "items": {
          "type": "number"
        }
      },
      "point": {
        "type": "object",
        "properties": {
          "lat": {
            "type": "number",
            "minimum": -90,
            "maximum": 90
          },
          "lon": {
            "type": "number",
            "minimum": -180,
            "maximum": 180
          },
          "radiusMiles": {
            "type": "number",
            "minimum": 0.1,
            "maximum": 500
          }
        },
        "required": [
          "lat",
          "lon",
          "radiusMiles"
        ],
        "additionalProperties": false
      },
      "limit": {
        "type": "integer",
        "minimum": 1,
        "maximum": 200
      },
      "offset": {
        "type": "integer",
        "minimum": 0,
        "maximum": 5000
      }
    },
    "required": [
      "growthScenario",
      "marketGravityWeight"
    ],
    "additionalProperties": false
  }

• getServiceInfo

  read

  Return artifact versions, schemas, licenses, attribution duties, and scenario limitations.

  FreshnessScheduled (source-level)
  Input schema

  {
    "type": "object",
    "properties": {},
    "required": [],
    "additionalProperties": false
  }

Where this source is used

• Texas grid and land screening

  Put Texas grid context next to the state parcel record and the site constraints in one screening call.

• Land research

  Parcel records, ground conditions, and what reaches the property line, for raw and rural land.

• Data center siting

  Power, water, transport, airspace, and hazard screening for a site before anyone signs anything.