Episode 121 – AI power plays

There are now more data centres being built in the US than office buildings. In this episode of Perspectives, Escala Chief Investment Officer, Tracey McNaughton, looks at the political implications of data centre demand for energy.

(00:56) That’s the first move in nine months but it was widely anticipated. It was called a “risk management” move. What’s your take?

(03:03) Stephen Miran, a Trump-aligned recent appointee, has stirred the debate by advocating for aggressive rate cuts. He wanted a half point cut, not the quarter point that was consensus.

(03:48) What does a less independent US central bank mean?

(05:58) We have seen this challenge to central bank independence before haven’t we. President Nixon pressured Fed Chair Arthur Burns to keep interest rates low ahead of the 1972 election.

(06:52) Trump is increasingly interfering in the operation of the market. It is strange to see the US government listed now as the third largest equity holder of Intel. So what’s the playbook? How do investors position for this?

(08:13)  Let’s start with demand, because that’s where Jensen Huang’s comment lands hardest. When he says AI needs 100 times more power, what exactly is being measured here?

(10:38) So this is real megawatts being pulled off the grid, and it’s happening fast - faster than new supply can be created.

(11:17) You’ve just described what looks like a classic market failure. In most sectors, when demand explodes, prices rise, and that price signal encourages more supply. But in power markets, the feedback loop seems broken. Why isn’t the invisible hand working here? What’s stopping utilities, or private investors, from rushing in to build more capacity and cash in on the higher demand from AI?

(12:19) And that tightness shows up as rising prices for wholesale electricity and eventually household bills. So AI demand, if unmanaged, risks spilling into the broader economy as a cost-of-living issue.

 

(12:40) How are regulators starting to respond? Do you see the beginnings of more heavy-handed policy - where governments start deciding who gets power and at what price?

(13:29) We have heard recently Pennsylvania is thinking about going its own way in what sounds like a break from the national approach. What exactly are they proposing, and what does it tell us about how fragmented U.S. energy policy might become?

(14:20) So effectively, Pennsylvania is saying: “Come here, set up your data centres, and we’ll make sure you have the power you need, even if it means going against the grain of national policy.”

(14:55) When I listen to all this, what I hear is that electricity is no longer just a background utility. It’s becoming strategic - like oil was in the 20th century. Are we moving into an era where countries, and even states, treat power as a weapon of economic advantage?

(15:40) So for investors, for policymakers, even for ordinary households - what does this mean?

(16:22) Tracey, this has been eye-opening. AI isn’t just a technology story - it’s an energy story, and a policy story, and maybe even a political story if states start going their own way.