f4ab9f5a19cc141476…The English page renders the captured source as Markdown and preserves its images and links. Copyright remains with the official publisher.
Climate tech podcast: net positive AI
Evidence tier: A1 Evidence type: Auto-discovered official publication Source: Barclays browser-verified publication index Official publication date: 2026-07-03 Captured: 2026-07-18T16:32:06.575Z
Can AI be net positive?
The rapid growth of AI and data centres is creating a new kind of challenge for energy systems. The IEA projects global data centre electricity demand will more than double by 2030 to around 945 TWh, equivalent to more electricity than Japan uses in a year1.
This raises a key question. If AI is driving electricity demand at a speed the energy system has not encountered before, can it also help fix that system?
This is explored in the latest episode of Barclays’ Climate Tech podcast, recorded live at the Sustainable Markets Initiative’s CEO Summit, featuring:
Bernard Looney, Chair of ExpectAI and CEO and Chair of Prometheus Hyperscale
Aniruddha Sharma, CEO and Chair of Carbon Clean
Martin Soltau, Co-CEO of Space Solar
“I am absolutely convinced that are we to focus on it, we can absolutely have a world where the positives of AI vastly outweigh the negatives.
That's not to say there aren't negatives, but the positives vastly outweigh them, such that AI is a net positive for the energy system”
- Bernard Looney, Chair of ExpectAI, CEO and Chair of Prometheus Hyperscale
Rethinking the data centre
For Bernard Looney, AI’s growth cannot be separated from the energy challenge. He believes you cannot win on AI without solving energy and you cannot solve energy by building data centres the way they have always been built.
Prometheus Hyperscale is tackling this directly by redesigning the data centre model around self-sufficiency and efficiency. Its model combines behind-the-meter power generation with liquid-cooling technology that can reduce energy use by up to 50% compared with traditional air-cooled systems2. By blending natural gas with solar, wind and eventually small modular reactors, Prometheus Hyperscale is treating the data centre as part of the energy solution, not just another source of demand.
Beyond data centres, AI can also help businesses navigate the transition more practically. Barclays is trialling ExpectAI’s Una platform with clients to help them access relevant tools and support as they transition.
“Let’s try and do whatever we can do today. And let’s not get into a situation where, you know, perfect is the enemy of good.”
- Aniruddha Sharma, CEO and Chair of Carbon Clean
Decarbonising what gets built now
Carbon Clean is solving another part of the same equation on how to decarbonise the energy infrastructure built to meet AI’s rising power demand. As new gas fired generation comes online, its carbon capture technology offers a way to reduce emissions now, with modular systems that are cheaper, smaller and more easily deployable. With technology that can be installed on industrial sites in a matter of days, Carbon Clean is focused on what can be deployed today, not just what might be possible tomorrow.
Power beyond the grid
“A solar panel in space generates 13 times the amount of energy than that same panel does on Earth”
- Martin Soltau, Co-CEO of Space Solar
Martin Soltau takes the discussion further into the future. Space Solar is looking to harvest solar energy in orbit and transmit it wirelessly back to Earth. At commercial scale, it could provide reliable, dispatchable clean power beyond the limits of weather, land use and existing grid geography. This reframes the challenge from generating more power to meet AI-driven demand, to redesigning the energy system to ensure the power can reach the data centres.
AI as part of the fix
The discussion also explored how AI could make the energy system smarter, helping power move more efficiently from generation to use.
Emerging uses include:
- Grid optimisation, using real‑time data to assess how much power transmission lines can safely carry, and reducing the need for costly network upgrades.
- Smarter integration of renewables, improving forecasting of when sources such as solar and wind will be available.
- Electric vehicles, by managing when and how vehicles charge, AI can help reduce peak demand, and reduce strain on the grid.
- Battery and storage innovation, by accelerating material development and improving how batteries are charged and discharged.
Mastercard