Uploaded June 2026 | Updated September 2026, 2 weeks ago
AI is increasing electricity demand. Used well, it may also be one of the tools we need to reduce it.
My guest is Philippe Rambach, Chief AI Officer at Schneider Electric. Philippe works at the intersection of climate tech, energy management, industrial automation, and artificial intelligence, where the challenge is not abstract: how do we cut emissions, lower energy use, support electrification, and keep increasingly complex grids stable?
That matters now because the energy transition is entering a harder phase. More renewables. More electric vehicles. More data centres. More pressure on grids that were never built for this level of distributed, variable demand. Policy, investment, and net zero plans are all running into the same brutal question: can we decarbonise fast enough without making the system harder to operate?
What changed my thinking in this conversation was Philippe’s insistence that the most important AI for climate may not be the loudest AI. Not chatbots. Not giant models looking for a problem. He points instead to AI that cuts energy waste in buildings, shifts demand away from peak hours, supports grid operators, and helps industrial companies move from endless pilots to production. He also shares Schneider Electric’s calculation that, in some energy-saving uses, the carbon emitted to run an AI model can be far outweighed by the energy saved.
We also get into the AI hype problem. Why so many pilots fail. Why domain knowledge matters as much as technical skill. Why using an LLM to optimise room temperature makes no sense. And why Europe’s strength may lie not in copying Silicon Valley’s AI obsession, but in applying AI to real-world energy, infrastructure, manufacturing, water, and industrial systems.
This is not AI theatre. It is a grounded conversation about emissions reduction, grid flexibility, electrification, climate policy, and the practical work of making decarbonisation operate in the real world.
For business leaders, sustainability professionals, policymakers, investors, and technologists trying to cut emissions without breaking systems.
If you’re using AI in real energy, industrial, or climate work, I’d be very interested in your perspective in the comments.
Links:
Climate Confident Podcast: climateconfidentpodcast.com
Subscribe on YouTube for more practical climate and decarbonisation conversations.
Follow Climate Confident wherever you get your podcasts.
Chapters / Timestamps
00:00 – AI’s 500-to-1 energy-saving claim
01:19 – Why Schneider chose a business leader for AI
04:33 – The data journey before ChatGPT
06:36 – The cheapest energy is the energy you do not need
07:21 – Why peak demand makes electricity dirtier and dearer
09:45 – How AI cuts building energy waste
12:15 – Why the energy transition may need AI
14:12 – AI’s climate upside versus its footprint
18:23 – Why LLMs are not the answer to every energy problem
19:30 – The AI pilot trap and why scaling fails
25:46 – The missing skill: AI plus domain knowledge
28:10 – Why renewable-heavy grids become harder to manage
30:53 – AI as infrastructure for grid stability
34:16 – Responsible AI starts before deployment
36:49 – AGI hype, human judgement, and energy impact
39:36 – Start from business value, not technology
AI is increasing electricity demand. Used well, it may also be one of the tools we need to reduce it.
My guest is Philippe Rambach, Chief AI Officer at Schneider Electric. Philippe works at the intersection of climate tech, energy management, industrial automation, and artificial intelligence, where the challenge is not abstract: how do we cut emissions, lower energy use, support electrification, and keep increasingly complex grids stable?
That matters now because the energy transition is entering a harder phase. More renewables. More electric vehicles. More data centres. More pressure on grids that were never built for this level of distributed, variable demand. Policy, investment, and net zero plans are all running into the same brutal question: can we decarbonise fast enough without making the system harder to operate?
What changed my thinking in this conversation was Philippe’s insistence that the most important AI for climate may not be the loudest AI. Not chatbots. Not giant models looking for a problem. He points instead to AI that cuts energy waste in buildings, shifts demand away from peak hours, supports grid operators, and helps industrial companies move from endless pilots to production. He also shares Schneider Electric’s calculation that, in some energy-saving uses, the carbon emitted to run an AI model can be far outweighed by the energy saved.
We also get into the AI hype problem. Why so many pilots fail. Why domain knowledge matters as much as technical skill. Why using an LLM to optimise room temperature makes no sense. And why Europe’s strength may lie not in copying Silicon Valley’s AI obsession, but in applying AI to real-world energy, infrastructure, manufacturing, water, and industrial systems.
This is not AI theatre. It is a grounded conversation about emissions reduction, grid flexibility, electrification, climate policy, and the practical work of making decarbonisation operate in the real world.
For business leaders, sustainability professionals, policymakers, investors, and technologists trying to cut emissions without breaking systems.
If you’re using AI in real energy, industrial, or climate work, I’d be very interested in your perspective in the comments.
Links:
Climate Confident Podcast: climateconfidentpodcast.com
Subscribe on YouTube for more practical climate and decarbonisation conversations.
Follow Climate Confident wherever you get your podcasts.
Chapters / Timestamps
00:00 – AI’s 500-to-1 energy-saving claim
01:19 – Why Schneider chose a business leader for AI
04:33 – The data journey before ChatGPT
06:36 – The cheapest energy is the energy you do not need
07:21 – Why peak demand makes electricity dirtier and dearer
09:45 – How AI cuts building energy waste
12:15 – Why the energy transition may need AI
14:12 – AI’s climate upside versus its footprint
18:23 – Why LLMs are not the answer to every energy problem
19:30 – The AI pilot trap and why scaling fails
25:46 – The missing skill: AI plus domain knowledge
28:10 – Why renewable-heavy grids become harder to manage
30:53 – AI as infrastructure for grid stability
34:16 – Responsible AI starts before deployment
36:49 – AGI hype, human judgement, and energy impact
39:36 – Start from business value, not technology
