AI expected to increase data center energy demand by 160%

With our unlimited demand for artificial intelligence, researchers at Goldman Sachs calculated how our energy consumption is going to be impacted.

Analysts at Goldman Sachs recently published a paper analyzing the impacts of data centers on our energy usage, and their research estimates that data center power demand will grow 160% by 2030.

It isn’t that our demand for data has been slow in the past. Instead, data center workloads nearly tripled between 2015 and 2019. Although during that period, data center demand for power remained flattish, at about 200 terawatt-hours per year; largely because of efficiencies in how they used power.

Since 2020, efficiency gains have flattened, and power consumed by data centers has risen, A LOT. Some AI innovations could improve computing speed faster than they ramp up their electricity use, but using AI will still cause an increase total consumption of power.

Stated by the International Energy Agency, a single ChatGPT query requires 2.9 watt-hours of electricity, compared with 0.3 watt-hours for a simple Google search.

Source: StatistaReutersSimilarweb

Monthly usage averages 1.7b visits per month, and growth has relatively flattened since it's inception in late 2022. Ongoing advancements will increase convienence for users, expanding it's network effect.

Today, data centers currently consume 1-2% of overall energy, but this percentage will likely rise to 3-4% by the end of this decade. This will have a significant impact on our global footprint– the carbon dioxide emissions of data centers may more than double between 2022 and 2030.

Hopefully companies will gather their energy from renewable resources, like nuclear, which has been a hot topic recently with companies like Oklo, a nuclear energy startup that recently IPO'd and partnered with OpenAI.

What will happen to our energy sector?

Over the last decade, US power demand growth has been flat, even though our population and economy have grown. Efficiencies have helped; examples like the LED light, which has driven lower power use. But that is going to change. Between 2022 and 2030, Goldman Sachs estimates the demand for power will rise roughly 2.4% — and around 0.9 percentage points of that figure will come from data centers.

The link has been copied!