Data-center share of global energy consumption is set to increase significantly. In 2022, data centers made up about 2% of global electricity demand, and this is projected to reach as much as 7% by 2030. There are already six states in the US where data centers consume more than 10% of electricity supply, with Virginia leading at 25%.
This energy demand is translating into opportunity in two ways. It’s facilitating greater adoption of clean energy, and hyperscalers are making their data centers more energy efficient—driving changes across their supply chains that may translate into opportunities for investors. The risk, however, is that many new AI data centers are being powered by natural gas—a non-clean energy source—which could put hyperscalers’ long-term zero-carbon commitments at risk.
For bond investors, AI’s energy and capital intensity are a double-edged sword. Issuers slow to adapt may face credit pressures, reduced access to capital and higher funding costs. But many companies that are building or financing AI-intensive data centers, together with utilities that are modernizing their grids and sourcing renewable energy, are issuing green bonds to address AI’s growing energy footprint.
From an actively managed perspective, the key is to identify issuers with credible transition strategies, AI-related competitive advantages, disciplined capital allocation and diversified funding access. These are the companies most likely to enhance credit quality over time and deliver attractive opportunities.
Growth in power demand is also putting pressure on infrastructure more broadly. Companies likely to benefit from upgrades include suppliers of high- and medium-voltage cables, providers of energy-efficient climate control systems, manufacturers of gas turbines (particularly with AI-driven control systems) and makers of fuel cells that can generate electricity on-site at data centers.
Water security is another large-scale environmental risk. Failure to manage it may inflict material damage on business and investment outcomes, while companies that are strategically focused on the issue could gain competitive advantages.
In other environment-related areas, AI can help improve the monitoring and measurement of CO2 emissions—for example, capturing and interpreting satellite data to assess emissions from power plants or vehicles. It can also refine estimates of wildfire emissions, strengthen the monitoring of carbon sequestration efforts and improve the integrity of carbon markets.
Risk to Employment Depends on Successful Transition
Within the social dimension of our AI impact matrix, three themes stand out: employment risk, the spread of misinformation and bias.
One of the most significant social risks is mass job displacement. According to the World Economic Forum’s Future of Jobs Report 2025, businesses expect AI and information technology to have the most transformative effect on their operations by 2030. Yet the overall impact on employment may be benign (Display).