A new white paper from the Exponential Roadmap Initiative (ERI), led by Johan Falk and Owen Gaffney alongside a group of contributing experts, delivers a clear message: as artificial intelligence scales at unprecedented speed, it must do so in a way that supports a stable planet.
Titled “What an Earth-Aligned Economy Now Requires from AI Hyperscalers,” the paper argues that the rapid expansion of AI infrastructure coincides with a shrinking window to avoid dangerous climate tipping points. Its conclusion is direct, AI deployment must actively support societal goals, especially efforts to remain within planetary boundaries.
What “Earth aligned AI” means
The paper introduces the concept of an Earth aligned economy, where AI is designed and applied to benefit both people and the planet. This includes reducing pressure on the climate, biodiversity, and natural resources, while supporting fair and sustainable economic growth.
In practice, Earth alignment requires AI to be intentionally guided toward planetary stability, from cutting emissions and protecting ecosystems to improving access to tools that help measure and manage sustainability transitions.
Three expectations for hyperscalers
The paper outlines three core expectations for companies building and operating large-scale AI infrastructure:
Firstly, AI must be powered by clean energy and help accelerate the next generation energy system faster than AI load growth, while avoiding investment that drives fossil fuel lock-in.
Secondly, AI must become significantly more resource efficient and transparent across its full value chain, with clear disclosure of emissions and resource use to enable accountability, inform decisions, and ensure a credible transition from temporary increases to absolute decline.
Thirdly, AI applications must prioritise Exponential Solutions capable of delivering gigatonne scale climate impact, rather than those that accelerate overshoot.
Toward an Earth-aligned AI pathway
The paper shifts from principles to action, stressing that Earth-aligned AI will require coordinated effort across the ecosystem. Progress is already underway, but must scale faster through collaboration.
It proposes three tracks for implementation: clean power and energy system acceleration; AI efficiency, transparency, and procurement standards; and scaling Exponential Solutions. A dedicated “Hyperscaler Track” could help align standards, metrics, and partnerships.
The message is clear: not less ambition, but better aligned ambition. Hyperscalers that deliver clean energy, radical efficiency, and net-positive climate impact can help shape an Earth-aligned economy, and earn their licence to scale.
