---
title: "McKinsey and BCG Say Cheaper AI Could Keep Pushing Power Demand Higher"
description: "McKinsey expects data-center power demand to grow 24% a year through 2030 as cheaper tokens drive wider AI use."
author: "rews desk"
published: 2026-10-03T09:59:01Z
modified: 2026-10-03T12:46:12Z
url: https://rews.cc/a/mckinsey-and-bcg-say-cheaper-ai-could-keep-pushing-power-dem-a14cce
language: en
tags: ["ai", "energy", "infrastructure", "data-centers", "economy", "business"]
publisher: "Rews (https://rews.cc)"
---

# McKinsey and BCG Say Cheaper AI Could Keep Pushing Power Demand Higher

*McKinsey expects data-center power demand to grow 24% a year through 2030 as cheaper tokens drive wider AI use.*

By rews desk · October 3, 2026 · https://rews.cc/a/mckinsey-and-bcg-say-cheaper-ai-could-keep-pushing-power-dem-a14cce

## In brief

- McKinsey and BCG said in reports this week that cheaper AI will likely push electricity demand higher, not lower
- McKinsey expects global data-center power demand to grow 24% a year through 2030, then 5% a year to 2040
- BCG surveyed 1,300 executives; 75% of the most AI-mature firms actively manage token spending toward a return
- Half of those firms push paid AI tools on staff, while 22% set usage limits
- The Senate on Tuesday blocked a bill to make data centers pay for the grid costs they drive

McKinsey & Company and the Boston Consulting Group said in reports published this week that the falling price of using AI will most likely push electricity demand higher, not lower, as cheaper models give companies more reasons to use them at greater scale.

The two firms were writing about the same mechanism from opposite ends. Companies pay AI providers by the token, the small unit of text that models read and generate, and token prices keep dropping. McKinsey said cheaper tokens would make AI more widely used in the short term and raise electricity demand, even as companies work to make their models, chips and data centers more efficient.

Data centers are already drawing more power. McKinsey called their electricity demand “the fastest-growing load segment in OECD power markets” and said that in several markets data centers are the biggest reason demand is expected to grow through 2030. The firm expects global data-center electricity demand to grow 24 percent a year through 2030, then to slow to 5 percent a year between 2030 and 2040.

BCG’s report looked at how the businesses paying for those tokens behave. In its survey of 1,300 C-suite and other senior executives across more than 20 sectors, the companies furthest along with AI — a group BCG calls “future-built” — were not simply letting token spending rise by default. Three-quarters had made an explicit decision on token spending and managed it toward a specific return.

Half of that group actively encouraged employees to use paid AI tools to maximize adoption, compared with 25 percent of the companies lagging with the technology. Another 22 percent of the future-built companies set limits or controls on usage to manage costs.

There is a tradeoff in McKinsey’s account. As companies make each AI task cheaper and less power-intensive, AI becomes useful in more places, and total power demand rises. What happens after 2030 is less clear, the firm said, because companies are still working out whether AI delivers enough value. Taken together, the reports suggest cheaper AI will not automatically mean less energy use.

The projections land in the middle of a fight over who pays for the new load. On Tuesday, the Senate [blocked a bill](https://rews.cc/a/senate-blocks-bill-to-make-data-centers-pay-for-grid-costs-a-9d16f0) that would have made data centers cover the grid costs they drive.