The Critical Role of Conventional Energy Infrastructure in Powering the AI Industrial Revolution

For the last year or so, AI has in some way, shape or form become a (the) topic of conversation everywhere.? In late 2022, I became interested in understanding how AI could help in application to productivity gains, so I started playing around with a well-known AI platform.? While it was a novel and cool platform, it was also a frustrating tool when trying to utilize it for anything I expected from a supposed “Intelligent” platform.? While part of me was excited about the possibilities of what it “may become”, the other part of me started to file it under the next new “passing fad” – think pet-dot-com, Palm Pilot, Netscape navigator, Ask Jeeves, MySpace, RIM…

May 24, 2023 – We’re off to the races!

On May 24th, 2023 Nvidia reports a blowout quarter after the closing bell.? The demand-pull for the GPU chips it produces that power AI is so far above what The Street had forecast (exceeding 50% Wall Street estimates) that it becomes apparent that the AI race, along with the race to secure AI infrastructure in Nvidia’s chips was already well underway.

Dot-AI Bubble?

Chatbots, Co-pilots, GPT, Generative AI…Everywhere I look!? Since that “ah-ha!” moment in the spring of 2023, it has become almost impossible to find some corner where an AI integration, plan, scheme or possible scam isn’t lurking.? Much like the dot-com era, companies and promoters recognize both the potential value of an actual AI integration and scale-up in their businesses, as well as the impacts on the market cap of their publicly traded companies.? The AI-fueled frenzy has added trillions of dollars in market cap to the North American indexes in under 10 months.? Nvidia alone grew from ~$650 bb cap before the markets closed on May 24, 2023 to over $2 trillion as of Mar 8, 2024.? Companies across the large-cap tech sector are racing to hastily integrate AI across their base businesses, while the broader sectors race to integrate AI in order to achieve optimization and cost efficiency across their enterprise.? Heck, as I post this, LinkedIn is giving me an option to “Rewrite with AI”.? While certain aspects of the AI Industrial Revolution have elements that look like, smell like, taste like and feel like the dot-com bubble, the net impact of that enthusiastic period was a continued expansion of the internet into literally everything about the way we live, play and work.? From the ashes of that burst bubble, Amazon changed the way we shop, Apple changed the way we communicate, Facebook changed how we socialize and consume content and Google became another word for search.? We didn’t regress and use less internet; it expanded everywhere and we use more.

Conventional Energy Infrastructure to power the power of the power of AI

As AI technologies advance, their power requirements are escalating rapidly, underscoring the indispensable role of conventional energy sources. ?Conventional energy infrastructure and the ballooning demands of the AI industrial revolution in North America underpin my belief that a focus on natural gas utilities, pipeline operators, nuclear power plants, and oil & gas producers are necessary to meet these demands.? Further, an expansion of our conventional energy infrastructure to meet future needs is undeniable.

Current and Projected Power Demands

Recent forecasts indicate that data center power consumption in the US is expected to reach 35GW by 2030, nearly doubling from 17GW in 2022. This surge is largely attributed to the increased demand for AI and machine learning-ready infrastructure. Analysts estimate the future power consumption related to AI deployments could escalate from approximately 1GW in 2023 to 7GW by 2026, highlighting a significant growth in energy requirements driven by AI technologies.

Conventional Energy Infrastructure Expansion Requirements

To put this into context, I did a bit of research to break down what meeting the forecasted data center demand could require in terms of infrastructure expansion:

Nuclear Power Plants: ?Approximately 18 new nuclear power plants, each with a 1GW capacity, would be essential.

Natural Gas-Fired Electrical Plants: ?A typical 1GW capacity for a mid-size natural gas plant means about 35 new facilities would be needed to accommodate the increase in demand. Natural gas plays a pivotal role, contributing the largest share of utility-scale electricity generation capacity in the US at 43%.

Transmission Pipelines: ?Ensuring adequate natural gas supply to power plants will necessitate extensive expansion of pipeline infrastructure.

The Irreplaceable Nature of Existing Energy Infrastructure

Despite advancements in technology, conventional energy infrastructure remains irreplaceable for powering AI technologies due to its reliability, scalability, and efficiency. This provides a stable energy supply essential for the growth and continuous operations of AI industries.

Projected Expansion and Investment Opportunities

Given the expected doubling of data center power consumption by 2030 and the sharp increase in AI-related energy demands, significant investment is required in conventional energy infrastructure to bridge the gap between current capacities and future needs. Strategic investments in these sectors are crucial for supporting technological advancements and meeting the growing energy demands of the AI revolution.

But what does it all mean Basil?

The AI Industrial Revolution is grounded in the vast computational power of Nvidia GPUs housed within large-scale, energy-intensive data centers. These centers are fundamentally reliant on reliable, steady and scalable power from conventional energy infrastructure.? Beyond the competition amongst industry and corporations, Sovereign Countries around the world are also in a competitive race for AI dominance, treating AI advancements as both a commercial and strategic imperative. Governments are actively investing in AI research, fostering industry partnerships, and establishing national AI strategies – including the buildout of massive AI data centers. This intense focus on AI from a sovereign perspective underscores its perceived critical importance across various domains, from economic development to national security, with many referring to it as an "AI arms race."

I Gotta Have More Cowbell!

In my view, I believe that in order to fuel the expansion of the AI Industrial revolution, there is an impending need for significant capital investment in the expansion of conventional energy sectors—including natural gas, pipelines, nuclear energy, oil & gas, and power utilities.? Robust demand for energy to power AI technologies presents a significant growth opportunity for the conventional energy sectors, which could yield strong returns due to the sustainable and ever-increasing nature of AI demand-pull. My perspective sees conventional energy as an underpinning element for the continued advancement and integration of AI across industries, making it a strategic play for forward-looking investors.

#energyinfrastructure #pipelines #nuclearpower #naturalgas #oil&gas

?