Hi, I’m Carl Tannenbaum, Chief Economist for Northern Trust.

A few years ago, lightning struck an electrical substation near our home. As a result, we were without power for almost three days. We were cooking over fire, using candles, and socializing without electronic distractions. It was as if we had retreated a hundred and fifty years in time.

It is easy to overlook how important access to cheap power is for societies. It is essential both for commerce and the quality of life.

But as economies embrace artificial intelligence, or AI, the price and availability of power is becoming a concern. AI has taken businesses and the markets by storm. Its promise of additional productivity could boost profits and raise standards of living. For AI adherents, the sky is the limit.

Those lofty ambitions, however, may be grounded by earthy limitations. It is going to take a lot of electrons to deliver AI solutions. Data centers are expected to account for twenty percent of the increase in global electricity usage through the year twenty thirty.

These facilities will compete with residences and factories for power, and everyone will likely pay more for energy in the years ahead. Producing more electricity will require more fuel, including increased extraction of coal and natural gas.

Construction of new plants, windmills, solar farms, and dams will require capital and could raise environmental concerns. At the same time, the equipment needed to turn fuel into electricity is in short supply. Turbines used in gas‑fired power plants are on backorder, and competition for these components could attract regulation.

Transmission presents another significant challenge. Power cables are typically made of copper, and Bloomberg anticipates a global shortage of six million tons of that metal over the next ten years. Much of the world’s electricity infrastructure is also aging and may not safely handle increased loads without significant modernization.

AI proponents correctly assert that the technology can also save electricity through improved process engineering. AI‑driven conservation routines are already assisting factories and offices, and new generations of chips and algorithms may require fewer computations, reducing overall power demand.

Still, as we get excited about the possibilities of AI, we need to avoid taking the availability of power for granted. There are many problems to solve and trade‑offs to reconcile.

After the most recent blackout, several of my neighbors invested in backup generators. That gave me the bright idea of purchasing a fleet of them and selling electricity to local server farms. It would be the best performing investment in my portfolio.

And that’s The View From Here.