Falling AI token prices threaten to accelerate power grid strain

What's Happening

• McKinsey and Boston Consulting Group report that falling AI token prices will drive higher-volume enterprise adoption, Business Insider notes.
• This surge in usage threatens to overwhelm electric grids even as developers make individual models and chips more efficient.
• McKinsey projects global data-center electricity demand will grow 24% each year through 2030 before slowing to a 5% growth rate in the following decade.

Why It Matters

Efficiency gains in AI hardware and software create a rebound effect. Cheaper computing costs encourage companies to deploy AI across more departments, multiplying total energy consumption. Grid operators face a compounding crisis as data centers become the primary driver of electricity demand growth.

The Context

• Data centers represent the fastest-growing load segment in OECD power markets.
• A Boston Consulting Group survey of 1,300 executives found that mature AI adopters manage token spending to guarantee specific returns.
• Half of these advanced companies encourage employees to use paid AI tools, while 22% enforce strict usage limits to control costs.

What They're Saying

• McKinsey researchers note that making individual AI tasks cheaper and less power-intensive guarantees the technology will spread to more applications, driving up total power demand.

What to Watch

• McKinsey analysts caution that data-center growth projections after 2030 remain uncertain while enterprises evaluate whether AI deployments deliver sufficient business value.

The Bottom Line

Making AI cheaper and more efficient will accelerate data center energy consumption as enterprises scale their deployments.