AI boom's hidden cost: A new inflation worry for the Fed?

Consider this: what if the AI excitement, which is supposedly changing everything, eventually ends up reflected in your monthly electricity charges? That would certainly be noteworthy, wouldn't it?

Let's explore how that scenario could actually materialize.

When most people think about AI, they picture advanced chatbots, robotics breakthroughs, and soaring tech shares. Yet, there's a far less appealing aspect to this. Someone has to construct massive facilities to house all those computers and maintain them continuously. These are data centres, and they demand electricity for operation and cooling systems to prevent overheating.

Imagine dozens of these structures appearing in a single town, which is exactly what's happening globally. The local power grid was likely not designed for such a surge in demand. Consequently, utility providers might need to generate more power, upgrade substations, and lay new cables—all of which come with significant costs. Depending on how these expenses are distributed, households and businesses could end up footing part of the bill.

Electricity isn't the only concern. These projects require vast amounts of raw materials like steel and copper, plus specialized equipment and skilled labor. If tech companies compete for the same resources and workers at once, prices could escalate further.

You might assume this is a distant, hypothetical scenario, but it's not. Dallas Fed researchers earlier this year estimated that data centres' impact on electricity prices alone could add approximately 0.04 to 0.13 percentage points to annual US inflation by 2030 under plausible conditions.

That estimate isn't an inflation crisis on its own. However, with the Fed already struggling to bring inflation down to 2%, it represents yet another potential complication.

This is all without considering that businesses might begin passing higher power and construction costs to their customers. That would give the Fed another reason to maintain a tighter monetary policy, supporting the "higher for longer" stance. This could also sustain elevated Treasury yields, adding pressure on expensive tech stocks—an ironic twist, isn't it?

Of course, there's a counterargument: AI could boost business efficiency and lower costs over time. That's plausible. But while the promise is reduced prices and greater efficiency, the journey there might be costly.