# Making room in a data center

Google DeepMind reported that an AlphaEvolve-discovered scheduling heuristic had been in production for more than a year and continuously recovered an average 0.7% of Google's worldwide compute resources.

Canonical: https://brightaifuture.com/discoveries/alphaevolve-data-centers
Format: discovery
Source publication: 2025-05-14
Bright publication: 2026-09-07
Substantive update: None recorded
Evidence and review: Deployed; confidence: unassessed; approved; ai-assisted. AI-assisted editorial comparison with the cited primary source; result, setting, source date and limitations retained. Independently checked within the research team. Publication authorized by the site owner; no human source review is claimed.

## The human problem

Data-center capacity is finite, and idle or poorly allocated compute leaves less capacity for useful work on the same infrastructure.

## The prior constraint

A production scheduler must make dependable allocation decisions across changing workloads at very large scale.

## AI’s actual role

AlphaEvolve used language models, automated evaluation, and evolutionary search to produce a scheduling heuristic for Google's Borg system.

## The documented result

Google says the production heuristic continuously recovers, on average, 0.7% of its worldwide compute resources.

## Why it may matter

Recovering capacity can allow more tasks to run on an existing computational footprint.

## Limitations

This is a company-reported, Google-specific operational result.

The announcement does not quantify electricity, water, carbon, reliability, or independent-audit outcomes.

## Unresolved questions

Would the heuristic transfer to other scheduler designs?

What are the measured energy and reliability effects under representative workloads?

## Provenance and history

{
  "dates": {
    "eventDate": null,
    "publicationDate": "2025-05-14",
    "captureDate": "2026-09-07",
    "lastReviewedDate": "2026-09-07"
  },
  "provenance": {
    "origin": "editorial",
    "externalId": "https://deepmind.google/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/"
  },
  "revisions": [
    {
      "id": "revision:95733113a8bd2aeccf4b",
      "recordedAt": "2026-09-07",
      "summary": "Recovering capacity can allow more tasks to run on an existing computational footprint.",
      "sourceIds": [
        "source-alphaevolve-data-centers"
      ]
    }
  ],
  "corrections": []
}

## Original sources

- [AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms](https://deepmind.google/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/)

## Continue exploring

- [Energy & infrastructure](https://brightaifuture.com/worlds/energy)
- [Work & learning](https://brightaifuture.com/worlds/work)
- [Can intelligence help make energy more abundant, reliable and responsible?](https://brightaifuture.com/threads/energy)
- [When does a technical result become a public capability?](https://brightaifuture.com/threads/community)
- [Move the work. Watch the costs move.](https://brightaifuture.com/experiments/resource-shift)