AI: A sustainability risk and opportunity for long-term investors

Investors can help determine whether artificial intelligence becomes a net positive for societies and the environment. How they allocate capital and engage with companies and policymakers can play a vital role in the eventual impact of AI.

The debate encompassing AI’s sustainability credentials has optimists pointing to productivity gains, scientific breakthroughs, and improved resource allocation. In contrast, sceptics highlight the environmental risks from increased energy use, the erosion of social cohesion, labour impacts, and the concentration of power.

For sustainable investors, this debate makes clear that AI is a systemic force, a transversal phenomenon that can create new sustainability risks, unlock new opportunities, and reprice assets across the economy.

For investors with long horizons, the question is not whether AI will matter, but how its sustainability implications will interact with portfolio risk, return, and real-world impact.

Sustainability Risks Generated by AI

Carbon and energy system impact

The most immediate sustainability risk associated with AI is its rapidly growing carbon footprint. Operating large-scale models and associated datacentres already consume significant amounts of electricity and demand is growing with electricity use expected to double or triple by 2028 from 2024 levels in the US (US Department of Energy). 

While hyperscale operators increasingly procure renewable energy, the marginal impact of AI workloads is frequently met by fossil-based generation, especially in regions with constrained power grids.

For sustainable investors, this creates several layers of risk:

• Transition risk for utilities and power markets unable to decarbonise fast enough to meet AI-driven demand
• Reputational and regulatory risk for technology firms whose emissions rise faster than that they can implement mitigation strategies
• Systemic risk if AI-driven electricity demand drives emissions growth which is not offset by related carbon efficiency gains

Sectors most exposed to these risks include utilities and technology firms producing the data infrastructure for AI.

Water stress and local environmental externalities

Datacentres require large volumes of water for cooling, often in regions already facing water stress. As AI infrastructure expands into arid or drought-prone areas, competition for water between technology firms, agriculture, and communities becomes a potential flashpoint and introduces several issues including:

• Physical risk from water scarcity disrupting datacentre operations
• Drought risk where datacentre water use contributes to shortages and operators lose their ‘social license’
• Regulatory risk as municipalities impose water-use restrictions or pricing reforms

Investors in datacentre operators, cloud service providers, and enabling infrastructure must now assess water exposure with the same seriousness historically reserved for mining or agriculture. For a deeper dive see our paper on: “Water: The hidden constraint shaping the digital economy”.

Social cohesion and labour disruption

AI’s social risks are more diffuse, but potentially more destabilising. Automating cognitive and creative tasks threaten white-collar jobs in ways that differ from prior industrial transitions.

If productivity gains accrue primarily to owners of capital and highly skilled labour, AI could exacerbate inequality and erode social cohesion.

From an investment perspective, this matters because:

• Societies with weakened trust and rising inequality tend to exhibit higher political volatility and policy risk
• Companies deploying AI aggressively may face a backlash from workers, consumers or regulators
• Entire sectors such as customer service, back-office functions, and certain professional services may see margin pressure and labour dislocation simultaneously

These risks are especially salient for consumer-facing firms, financial services, and media.

Sustainability opportunities enabled by AI

While the risks are real, focusing only on AI’s negative impact misses its transformative potential to accelerate sustainability outcomes.

Efficiency gains 

AI can materially improve efficiency in energy, water, and material systems. Smart grids, predictive maintenance, demand forecasting, and process optimisation all reduce waste and lower emissions intensity.

In industrial settings, AI-driven optimisation can cut energy use and input costs simultaneously, strengthening both sustainability and competitiveness.

Sectors well positioned to benefit from AI include:

• Utilities and grid operators adopting AI to manage distributed renewables
• Industrial automation firms enabling efficiency gains in heavy industry
• Logistics and transportation companies using AI to optimise routing and fleet utilisation.

These gains are often incremental, but compounding, making them highly attractive from a long-term value creation perspective

Accelerated technology development and deployment

One of AI’s most profound sustainability opportunities lies in accelerating discovery. Machine learning is already being applied to materials science, battery chemistry, carbon capture, and drug development.

By reducing the time and cost required to identify viable technologies, AI can compress innovation cycles that historically spanned decades.

This creates investment opportunities in:

• Clean energy technology developers, particularly in storage, power electronics, and advanced materials
• Climate solutions platforms combining AI with hardware deployment
• Pharmaceutical and life sciences firms addressing public health and resilience challenges

The leading tech firms sit at the intersection of AI capability and sustainability-relevant application, though execution risk remains significant.

Scaling solutions faster

AI enables faster scaling of proven solutions. Once an optimisation model or design improvement is developed, it can be replicated globally at near-zero marginal cost. This characteristic is particularly powerful for sustainability, where diffusion speed often matters more than invention itself.

For investors, this favours business models with:

• High research and development costs, but low marginal deployment costs
• Platform characteristics that enable ecosystem adoption
• Strong alignment between AI deployment and regulatory or policy objectives

Renewable energy software providers, building efficiency platforms, and climate risk analytics firms exemplify this dynamic.

Squaring the AI and sustainability circle

For sustainable investors, AI is best understood as a force multiplier. It can amplify existing trends towards decarbonisation or resource strain, inclusion or inequality – depending on how it is deployed and governed.

The technology itself is neither inherently sustainable nor unsustainable; its net effect emerges from the interaction between markets, policy, and capital allocation.

We believe the implication for investors is clear. AI should be evaluated not only as a growth theme, but as a sustainability transition variable. Portfolio construction, engagement priorities, and risk assessment frameworks should evolve accordingly.

Those who integrate AI’s sustainability risks and opportunities coherently will be better positioned to navigate the next phase of long-term value creation.