We often hear about the significant power demands of artificial intelligence (AI) and new technologies, and there’s no denying that these tools consume substantial energy. However, we rarely discuss the other side of the equation: the energy these technologies save by delivering substantial productivity efficiencies. This hesitation to embrace new technology is common – resorting to “Luddism” often serves as a safety net for preserving the status quo.

A simple example is the time saved in drafting an email like this. Prior to current AI tools, this process took us about 7 hours to consider, acquire and verify any data, review assumptions, write, review, format, and send. Now, with the help of AI, the same work can be completed in approximately 4 hours. While these tools are indeed power-hungry, they allow us to save three hours of energy per task — a meaningful improvement in efficiency.

At the Cleantech Forum conference in Paris this week, our partner Daniel Doll-Steinberg presented an initial model comparing the energy required for this example and projecting future efficiencies. At a high level, we estimate that the energy needed to complete a task using a combination of human effort and a few minutes of AI time (4 hours total) is about 60% less than the energy required to complete the same task in 7 hours manually.

Looking toward to 2030, while humans and laptops would still require the same time input, advances in power generation, storage, and delivery (likely also driven by AI) could reduce energy needs by 66%. Our projections suggest that, in the future: Human plus AI collaboration could reduce time spent on tasks to 6% of current levels, with energy usage, from source, dropping to just 5%; and AI alone could potentially complete tasks using only 0.5% of the time and 3% of the energy required today.

This shift demonstrates not only the potential for time efficiency but also for significant energy savings, making the adoption of AI both practical and inevitable.

We expect similar impacts across many human-centred functions. For example, we are actively investing in companies like Ryftpay and Wamo, which are poised to transform the services they offer. These companies will not only reach clients they might not have been able to serve in the past, but their technology-driven efficiencies will also help optimise resource usage across various industries.

As an investment fund committed to this space, we aim to actively participate in the debate on the future of AI-driven efficiency. Our initial model, which we acknowledge is a starting point, will be made available to academics and other stakeholders for further analysis and refinement. Embracing these changes is essential; resisting them would be, as history has shown, a mistake.