# GB Carbon-Aware Compute – Forecast & Actual Intensity — workflow

## Choose this tool

Platform engineers, sustainability analysts and developers evaluating flexible GB compute workloads.

Useful output columns: `title`, `intervalStart`, `intervalEnd`, `forecastGco2PerKwh`, `actualGco2PerKwh`, `intensityIndex`, `forecastRankWithinWindow`, `forecastAboveWindowMinimumGco2PerKwh`, `forecastRevisionGco2PerKwh`.

## Practical workflow

Start with four upcoming intervals or select the lowest forecasts within the next 24 hours. Check units, UTC boundaries, forecast rank and the source's actual estimates separately. Feed candidate intervals to your own scheduler alongside capacity, cost and deadlines; repeat a named watchlist to capture source revisions.

## Check the fit

National Great Britain electricity-generation intensity in gCO2/kWh, not data-centre measurements or regional emissions. Individual forecast-ranked half-hours are not necessarily continuous and do not guarantee realized carbon savings. Future actual estimates are normally null. No generation mix, pricing, carbon offsets or workload execution is included. Default output selects 4 intervals; source failures, output caps and deadlines are explicit in SUMMARY and preserve previous history.

## Real output sample

One selected row from a 2026-11-21 run. Selected original scalar fields from one actual saved observation; nested data and private artifact links omitted.

Run: z0uPwxxjSL2YmJubn. Summary status: ok. Incomplete: false.

Dated bounded private release evidence for this exact input/build. This does not establish current values, whole-source coverage or fleet health.

```json
{
  "recordId": "GB:2026-11-21T20:00:00.000Z",
  "title": "GB carbon intensity 2026-11-21T20:00:00.000Z",
  "intervalStart": "2026-11-21T20:00:00.000Z",
  "intervalEnd": "2026-11-21T20:30:00.000Z",
  "forecastGco2PerKwh": 207,
  "actualGco2PerKwh": null,
  "intensityIndex": "high",
  "forecastRankWithinWindow": 48,
  "forecastAboveWindowMinimumGco2PerKwh": 121,
  "forecastRevisionGco2PerKwh": null,
  "comparisonStatus": "not_requested",
  "sourceUrl": "https://api.carbonintensity.org.uk/intensity/2026-11-21T20:00Z/fw24h",
  "url": "https://api.carbonintensity.org.uk/intensity/2026-11-21T20:00Z/fw24h",
  "observedAt": "2026-11-21T20:10:27.166Z"
}
```

## Combine with companion tools

Each Actor runs separately with its own inputs, authorization and pricing. These instructions do not configure automatic execution.

### [GPU Cloud Rate Monitor – Runpod & Lambda Price Changes](https://mayd-it.com/data-tools/actors/gpu-cloud-rate-monitor/)

Combine GB carbon-intensity observations with separately published GPU rental offers.

1. Review forecast intervals for a flexible GB workload and keep actual estimates separate.
2. Collect supported GPU cloud offers and confirm the provider region yourself.
3. Use both datasets as inputs to your own scheduling evaluation.

The carbon feed is GB-specific and does not prove a cloud provider's power mix. Lower forecast carbon intensity does not guarantee lower cost or emissions.

### [Azure Retail Price Comparator – Regions & Meter Changes](https://mayd-it.com/data-tools/actors/azure-retail-price-comparator/)

Bring regional cloud retail prices into a separately scoped compute-planning worksheet.

1. Save the GB carbon-intensity intervals and source observation time.
2. Run the Azure comparator with your chosen region and meter filters.
3. Keep retail price assumptions and carbon estimates separate when evaluating a workload.

Azure list prices exclude negotiated discounts and workload utilization. National grid intensity is not a provider-specific carbon accounting figure.
