AI’s Impact and the Path to Net-Zero
The rapid growth and evolution of generative AI has been a defining feature of the global economy in recent years. Goldman Sachs estimates that $527 billion will be spent on AI infrastructure in 2026 alone [1] and the International Energy Agency estimates that electricity demand from AI datacenters will more than double from current levels by 2030 [2]. The blistering pace of AI development and investment has created severe strains on electricity generation – with at least 50% of new AI-related power likely to come from fossil fuels given current renewables capacity. In some places, datacenter demands have already outstripped the electric grid supply, turning instead to off-grid diesel or natural gas generators for more power. Each new datacenter also requires land and consumes huge quantities of water used for evaporative cooling. Billions of gallons of freshwater are now used each year to prevent AI supercomputers from overheating, directly affecting local watersheds and downstream availability [3]. As a result, annual GHG emissions and freshwater consumption have dramatically increased for companies leading the AI charge, many of which also previously adopted ambitious net-zero targets by 2030.
Coincidentally, AI systems themselves may also deliver a promising benefit for companies pursuing net-zero goals. Recent reports highlight the immense potential to improve GHG accounting for Scope 3 emissions across supply chains and industries by enabling more ubiquitous and seamless reporting [4]. (Note: SBTi Corporate Net-Zero Standard V2.0 is expected to be published in June 2026 and the current draft recognizes companies using avoidance and removals offsets to compensate for ongoing emissions as they work towards net-zero).
In response, several large tech companies have elevated offset purchasing – namely from engineered carbon removal projects – presenting both challenges and opportunities for a supply-constrained sector of the VCM. Companies competing to secure removal credits have quickly deployed capital into a variety of nascent project types, where many still depend on future innovation to verifiably sequester carbon at scale. Forward offtakes for early-stage carbon removal projects carry obvious delivery risks but also delay the timeframe when climate mitigation actions are achieved. In contrast, nature-based projects that avoid or reduce emissions are immediately scalable and allow companies to invest in climate action now – while ensuring equally high levels of integrity. Carbon credits from forestry and grassland projects are particularly well-suited to mitigate the land and water use impacts associated with AI by supporting vital ecosystem functions.
In light of AI’s increasing appetite for power and water, the carbon market should recognize that engineered removals and nature-based solutions serve uniquely different – but equally important – roles in a robust offsetting strategy.
To inquire about The Climate Trust’s portfolio of high-quality carbon offsets, contact us at info@climatetrust.org.
References: