AI Technology and Energy Demands

AI technology continues to drive change and new uses across many industries. While generative AI is still in its early stages, people are starting to see its impact reach beyond just the workforce and now seeing its effects on the energy grid.

Growing Usage of AI

According to a recent survey by McKinsey & Company, 78% of global companies are now using AI for their businesses in at least one function – up from 72% in early 2024. During that same period, companies have started integrating generative AI across multiple functions within their organizations.

It’s clear the use of AI is only going to continue to grow. While there’s plenty of attention on the benefits and possibilities of investing in AI technology, there’s less focus on the changes needed to support that growing demand. Let’s explore another side of AI – one that presents both challenges and opportunities: the energy demand and infrastructure changes needed to support and sustain the rising AI usage.

Energy & Water Consumption

Let’s take a step back. How much energy are we talking about?

As AI usage increases, there is conflicting information about how much energy an individual AI query (such as asking ChatGPT or other Large Language Model (LLM) a question) actually uses. The simple answer is that it depends on which model is being used (newer models tend to be more efficient) and how complex the inputs are.

Epoch AI, an AI research firm, dug into this question in their analysis of “how much energy does ChatGPT use?” Their findings were interesting: a short query of less than 100 words was estimated to use less energy than 5 minutes of running a 10W lightbulb or 5 minutes of using a laptop – like the one used to write this blog post. However, a longer query (of ~7,500 words) used more than twice the energy cost of the laptop in watt-hours.

In addition to electricity, there’s also water usage involved – primarily for cooling data centers where AI systems are housed. This is another growing area of concern as demand increases.

Changes that Present Opportunities

Modernizing and restructuring the power grid presents room for opportunities. As AI usage expands, growth will be needed in several supporting areas: chip manufacturing, semiconductors, and building additional data centers to support the increased demand. Smaller components of the grid will be affected too – think things like meters, sensors, monitoring systems, etc. There are potential opportunities, not just to upgrade existing infrastructure, but to also explore new energy sources – potentially with a cleaner impact on the environment. Some, like Sundeep Bajikar, a global technology investor, think the grid needs to catch up already. Either way, these changes could create entirely new supply chains and market opportunities, especially for those in the industrial and manufacturing sectors.

Change is happening and accelerating as the AI usage grows. Companies that anticipate needs and can strategically adjust will be better positioned for the opportunities the growing energy demand provides.

“Change is the only constant in life.” – Heraclitus

Please reach out to a member of our Manufacturing & Distribution team for more information on the topic outlined above. For more information regarding our Manufacturing & Distribution experience, visit our Manufacturing & Distribution industry page.

About the Author

Related Industries