AI can't outprompt a shortage of power, water, and land

The rapid expansion of AI infrastructure is encountering constraints in energy, water, and land resources, according to analysts at Forrester. The firm predicts governments will soon compel datacenter operators to cover the expenses of consumed resources. A 2027 sustainability outlook from Forrester underscores the need for a reevaluation of infrastructure funding, management, and regulation.

Recent PwC estimates suggest a global datacenter capital expenditure of $22 trillion to $50 trillion by 2050, which is comparable to a year's worth of U.S. economic output. However, the ability of grids and communities to accommodate these plans remains questionable. Forrester warns that unlimited AI ambitions will face a reckoning with limited resources.

Governments and utilities are expected to demand that datacenter operators finance grid improvements, secure reserved capacity, and ensure protections against speculative demand. Community impact assessments are also projected to become a mandatory requirement before new projects commence. Forrester's senior analyst, Abhijit Sunil, explains that utilities cannot anticipate every datacenter's timely arrival or promised load consumption.

If generation and transmission are built for speculative demand, households and businesses may ultimately pay for the unused infrastructure. The White House has already urged hyperscalers to procure their own power, fund upgrades, and pay for reserved capacity, even when it remains unused. By 2027, at least two countries will implement ratepayer-protection principles through tariffs or similar obligations.

Cloud and colocation expenses will become more location-specific, contractually binding, and susceptible to power-market risks. For Sunil, datacenter operators should anticipate long-term minimum payment commitments, financial guarantees, energy tariffs, and demands to decrease power consumption during grid stress. Consequently, cloud economics will increasingly depend on location, electricity contracts, grid maturity, and providers' capacities to generate or reduce power.

Enterprises rapidly expanding AI capacity will likely inherit these costs. Despite the United States' potential shift away from certain environmental oversight, Forrester emphasizes that community scrutiny will play a crucial role in determining which facilities are built.