Skip to main content
Rase AI Intelligence Newsletter

Transportation, manufacturing, agriculture & energy Industry development

Google moved from AI cooling recommendations to supervised control

DeepMind reported in 2016 that machine learning helped reduce cooling energy at Google data centers. In 2018 it documented an automated control system working under operator supervision. [1][2]

Development status
Google reported an internal trial followed by supervised automated control at multiple data centers. [1][2]
Sector
Data-center operations and cooling. [2]

What happened

The first report describes recommendations in Google’s own facilities. The later system sent actions to a local controller that checked them before applying them, with human operators retaining supervision. [1][2]

Why it matters

Cooling uses energy beyond the servers themselves. This deployment shows why operating boundaries and independent equipment checks matter when software shifts from advice to control. [1][2]

Key dates

  1. DeepMind reported initial cooling energy results [2]
  2. DeepMind documented supervised automated control [1]

Risks and open questions

DeepMind reported cooling energy savings of up to 40% in its studied facilities. The figures are specific to those systems and use cooling energy as their denominator. [2]

Sources and verification

Catalog evidence checked How we verify records

  1. Primary source DeepMind’s 2018 account of supervised autonomous cooling
  2. Primary source DeepMind AI Reduces Google Data Centre Cooling Bill by 40% — Google DeepMind

Related profiles

White Apollo Go RT6 robotaxi with a dark roof and roof-mounted sensors on a street.
An Apollo Go RT6 on Heping Avenue in Wuhan, China, in December 2025. S5A-0043 / Wikimedia Commons (resized for display) Image source CC BY 4.0

Transportation, manufacturing, agriculture & energy

Apollo Go

Baidu

The service operates automated ride services or tests in markets covered by local deployment arrangements.

Medium confidenceEvidence checked

Transportation, manufacturing, agriculture & energy

Pony.ai robotaxi

Pony.ai

The service lets users hail rides in the driverless bZ4X robotaxi at the documented Guangzhou deployment.

Medium confidenceEvidence checked

Transportation, manufacturing, agriculture & energy

TYM Level-3 autonomous tractor

TYM

The reported prototype performs supported driving functions for agricultural equipment in the demonstrated setting.

Medium confidenceEvidence checked