ETF Smart Grid Infrastructure

The world needs more power to feed the "AI beast".

The age of electricity

Strong growth in electricity demand is heralding a new “Age of Electricity”, with demand set to soar through 2027.”

Global electricity consumption is expected to increase at the fastest pace in years over the 2025-2027 forecast period of this report, fuelled by growing industrial production, rising use of air conditioning, accelerating electrification, and the expansion of data centres worldwide.

Global electricity demand rose by 4.3% in 2024 and is forecast to continue to grow at close to 4% out to 2027.

Over the next three years, global electricity consumption is forecast to rise by an unprecedented 3 500 TWh. This corresponds to adding more than the equivalent of a Japan to the world’s electricity consumption each year.

Most of the additional demand for electricity through 2027 will come from emerging economies, which are expected to make up 85% of the growth.”

Source: International Energy Agency

Catalysts

a) The risk of a "power shortfall" for US data centers.

The large number of LLM introductions in the coming weeks (from xAI, Anthropic, OpenAI, and others) underscores the exponential pace of LLM development, and the need to rapid increases in infrastructure and computational requirements.

"A long history of algorithmic improvements suggest we shouldn't understate the incremental demand that comes from lower costs, more advanced capabilities, and continued scaling."

According to Morgan Stanley, they estimate a ~36 gigawatt shortfall in terms of US power access , through 2028. A typical nuclear reactor produces 1 gigawatt (GW) of electricity.

b) US “Reshoring” of manufacturing capacity for the production of humanoids.

As Humanoid production comes closer, power will become a bottleneck. The Industrial economy accounts for 26% of US Electricity demand (vs Data Center ~4%) and estimates suggest the combination of rising US Industrial Production & Humanoid adoption can drive 2% annual growth in US Industrial electricity demand, reversing prior 25yrs of declines.

c) Electricity demand is growing on an absolute basis but also on a relative basis.

The electricity share in final energy demand is growing by 1% each year i.e. every year 1% of hydrocarbons energy consumption is switched to electricity.

Source: BloombergNEF

d) Annual expenditure on distribution networks (Grid infrastructure) more than triples to about $535 billion by 2050, from $147 today.

Sources: Bloomberg New Energy Finance,

Largest positions and geographical exposure

Conclusion

Among the many bottlenecks for the AI revolution, energy might be the most important and the most difficult to address. IF estimates of data center energy consumption turn out to be true (or even in the vicinity of truth), our current energy infrastructure will not be able to support those demands.

Unfortunately, there’s no easy way to increase energy capacity quickly. Data centers have two options: on-grid energy and off-grid energy. On-grid energy passes through the electric grid and is doled out by utilities. Off-grid energy (or “behind the meter”) bypasses the electric grid, such as on-site solar, wind, and batteries. Or even better, building a GW data center next to a 2.5 GW nuclear power plant!

The problem with on-grid energy is the time it takes to expand grid capacity and transmission wait times are generally about 5 years.

So what are Data Centers using for Energy?

The short answer is they will buy anything they can get their hands on!

From an investor’s point of view, that’s where you want to be.

Reply

or to participate.