• MIT Technology Review's 6 October 2026 note says rising electricity demand is driven 'in part by AI data centers', but the note gives no figure for the AI share.
• The IEA estimates data centres used about 415 TWh in 2024, roughly 1.5% of global electricity consumption.
• The IEA base case projects about 945 TWh by 2030, just under 3% of global consumption.
• In that base case, data-centre growth is less than 10% of global electricity demand growth from 2024 to 2030.
• The IEA says the United States and China account for nearly 80% of global data-centre growth to 2030, so exposure is regional.

AI's electricity appetite is real but smaller globally than the headlines imply, and it is very uneven by region. MIT Technology Review's 6 October note on its 2026 climate-tech list says rising electricity demand is driven in part by AI data centers. The IEA's primary figures put that in proportion. For teams shipping AI, the practical constraint is the local grid, not the global total.

What the MIT note says, and what it does not

The note calls rising electricity demand, driven in part by AI data centers, one of the biggest stories of the year. It is a framing sentence in a methodology note about how a company list was chosen, so it gives no load figure and no estimate of AI's share. That is normal for such a note and is not a defect. It does mean the sentence cannot be cited as evidence of scale.

The IEA's base case sets the scale

The IEA estimates data centres consumed about 415 TWh in 2024, roughly 1.5% of global electricity consumption. Its base case projects about 945 TWh by 2030, just under 3% of the total. It also says data-centre growth accounts for less than 10% of global demand growth between 2024 and 2030. These are projections under stated assumptions, and they cover all data centres, not only AI workloads.

The pressure is concentrated in a few markets

The IEA says China and the United States account for nearly 80% of global data-centre consumption growth to 2030. It projects US consumption rising by about 240 TWh, up 130% from the 2024 level. A global share under 10% can therefore sit alongside serious local strain. Siting, interconnection queues and regional supply decide whether a given deployment can get power, so planning should start from the specific region.

What the list's companies address

The MIT list is a curated set of climate-tech companies, and the five it describes in detail work on batteries, storage and wind: WeLion, Envision Energy, Energy Dome, Moment Energy and Form Energy. Form Energy, for example, is described as making iron-based batteries for multi-day storage. MIT says it chose companies on potential impact, scalability and commercial progress. Inclusion is an editorial judgment, not evidence that a technology relieves AI-driven demand.

Verdict: plan on regional capacity, not global averages

Treat the claim that AI is straining electricity supply as true in places and overstated as a global statement. Anyone budgeting for AI deployments should use regional grid-operator or regulator data alongside the IEA's projections, which carry their own scenario assumptions. The MIT note is useful as a pointer to a live issue. It is not a source for how large AI's load is or how it will be met.