Sustainability update August 2026

Study finds AI’s potential climate benefits outweighed by fossil fuel productivity gains

 A modelling data study published by Nature Partner Journals’ npj Climate Action found that AI-driven productivity gains in fossil fuel industries generated significantly more carbon emissions than those avoided by AI optimisation in renewable energy markets. Comparing the technical potential for AI to boost clean power generation against its application in fossil fuel production, the study found that large-scale emissions asymmetries between renewables and fossil fuels occurred across 64 empirically derived scenarios. Net annual carbon pollution increased over a range between 0.47 and 1.8 gigatonnes, equivalent to around 1% to 5% of the energy sector’s yearly emissions – net emissions reductions only occurred where fossil fuel sector AI productivity gains were zero.

While previous research explored the indirect climate benefits of AI optimisation in renewable energy production, npj Climate Action’s study was the first to assess AI’s climate impacts across the whole energy sector. If renewable and fossil fuel energy producers adopted AI optimisation at similar rates, the study calculated that productivity gains for renewables would need to be four times greater than those for fossil fuels for emissions to break even. In all its carbon pricing modelling scenarios, the study showed that emissions asymmetries could be narrowed by increasing pricing thresholds but not eliminated. Datacentre energy demand wasn’t factored into the study’s calculations, but it nevertheless found that the emissions generated by AI datacentres in optimising fossil fuel production were significantly above current estimates. Overall, the study concluded that reduced emissions from AI optimisation were possible but only after a systemic overhaul of the energy sector’s governance framework: the emissions asymmetries generated by AI productivity gains reflected “an energy system still structurally anchored in fossil fuels”.

Renewable applications for AI optimisation are still predominantly pilot schemes, but fossil fuel producers have deployed them at scale. The International Energy Agency (IEA) estimates that AI could increase the volumes of recoverable oil and gas reserves by around 5% and reduce the costs of deepwater exploration and mining. Reflecting this potential, Mike Sommers, head of the American Petroleum Institute, described AI’s role in subsurface exploration potential as “the next fracking boom”. Oil giant Saudi Aramco manages the world’s largest proven crude oil reserves and invested heavily in AI optimisation over 2024 and 2025 after identifying 442 potential use cases across its operations. Independent research and energy intelligence company Rystad Energy estimated in May that digitalisation and AI optimisation could create around $500 billion of cumulative value for fossil fuel exploration and production companies between 2026 and 2030 with additional annual value capture forecast to be $80 billion higher in 2030 than it was in 2025.

“AI could help create $500 billion of additional value for fossil fuel exploration and production companies by 2030.”

Greenbank avoids exposure to companies enabling fossil fuel expansion through AI, focusing instead on businesses enhancing the efficiency and sustainability of AI infrastructure, and those whose products contribute to wider decarbonisation and resource efficiency. The research by npj Climate Action reinforces the need to assess the specific application of AI rather than assuming it’s inherently beneficial or harmful from an environmental standpoint. Ultimately, the environmental value of AI depends on whether it’s being used to accelerate decarbonisation or boost carbon-intensive economic activities. Investors should continue to scrutinise how companies deploy AI and whether its use aligns with long-term sustainability objectives.

Disagreements and delays continue to undercut the ambition of UN plastics treaty talk

After participating nations failed to reach consensus on a binding global plastics treaty during UN-led talks in Geneva last year, core delegates met informally in Nairobi to resume negotiations and identify possible areas of convergence ahead of the next formal session. More a diplomatic reset than an opportunity for a breakthrough in negotiations, the Nairobi talks succeeded in keeping the process alive but failed to resolve the fundamental question of whether a globally binding treaty on plastic pollution should principally regulate the production of new plastics or the management and use of manufacturing materials.

While talks chair Julio Cordano praised the delegations in Nairobi for their “continued engagement and commitment”, critics argued that a revised reference text published after the meeting signalled even less ambition than the draft roundly rejected in Geneva. It was also noted that civil society representatives were largely shut out of the informal talks. Text brackets indicating continuing disagreements or divergent views covered almost everything related to production, chemical use, and the impacts of plastics on public health. High-ambition countries and environmental groups want a globally binding treaty that tackles plastic production at source rather than prioritising waste management. Plastics and petrochemical producers have formed a small but powerful lobby fighting against globally mandated production caps. This impasse highlights the challenges of the UN’s reliance on consensus decision-making. 

Details of human health impacts were heavily bracketed in the post-Nairobi “aid to negotiations” reference text, leading to concerns among health experts that they could be omitted from future negotiations. A health and plastics report published by The Lancet in 2025 concluded that plastics posed “a grave, growing, and under-recognised danger to human and planetary health”, responsible for more than $1.5 trillion in annual health-related economic losses. Advocates of tighter health controls point to inconsistent regulation on the use of toxic chemicals in plastic production, high volumes of plastic contaminants in global waterways and agricultural systems, and the increased prevalence of plastic waste exports to poorer nations with fewer regulatory checks and balances: the UK alone exported 675,000 tonnes of plastic waste in 2025, its largest annual shipment in eight years.

Another round of informal talks will take place in Bangkok in late September before delegates reconvene for the next formal session in March 2027. Greenbank continues to support calls for ambitious regulation on this issue via the Business Coalition for a Global Plastics Treaty.

“Godzilla” El Niño could push 50 million more people into acute food insecurity by 2027

El Niño is a periodic climate phenomenon where the surface waters of the central and eastern tropical Pacific Ocean become unusually warm. Rising heat and moisture rapidly alters global atmospheric circulation leading to significant changes in regional weather patterns. Recent predictive modelling for this year’s iteration – nicknamed “Godzilla” by climatologists – indicated that an abnormally intense climate shift exacerbated by elevated ocean surface heat could raise overall global temperatures by up to 3.6C. Experts believe that the resulting climate disruption could be the most impactive since the 1997 El Niño cycle which triggered widespread droughts and wildfires throughout southern Asia, devastating rainfall across the Americas, and increased coral bleaching worldwide.

Even accounting for adjustments in the forecasting models, the UN’s World Food Programme (WFP) warned that the rapidly evolving weather system could increase the number of people in acute food insecurity by over a fifth – current WFP estimates place that at 225 million people in 45 countries. In WFP assessments, acute food insecurity refers to hunger that falls below the level of famine. Exposure is highest in southern and eastern Africa, where important farming and pastoral areas face significant agricultural drought risks. The 2023-2024 El Niño cycle brought severe drought to the Southern African Development Community, straining livestock, eradicating around 80% of staple crops in the worst affected areas, and pushing over 8 million people into acute food insecurity. The disruption to monsoon rainfalls in Asian countries can significantly impact globally important rainfed crops such as rice, maize and pulses. India is currently on course for its weakest monsoon for decades – with rainfall 15% below the long-term average, crop yields and rural incomes are expected to fall while food prices are forecast to rise sharply. In Central America and the Caribbean, the risk of agricultural drought is highest across the region’s “Dry Corridor”, which incorporates Colombia, Venezuela, Cuba, the Dominican Republic, and Haiti. The 2015-2016 El Niño cycle devastated Haitian harvests and pushed 3.5 million people into acute food insecurity across the Dry Corridor.

"225 million people across 45 countries are estimated to be facing acute food insecurity.”

El Niño cycles expose the same vulnerabilities and often trigger failed harvests, livestock losses, rising household debt, and migration in search of new sources of food and water. This time out, the risks to areas with limited coping capacity are compounded by a demonstrably warmer planet and the combined effects of increased geopolitical tensions and global economic stress. These feedback loops between climate change and food system instability are explored further in a recent report by the Cambridge Institute for Sustainability Leadership, to which Rathbones contributed.