How To Think & Invest Around Energy Constraints?

The last few weeks have seen the drumbeat of news about the supply constraints for energy gather strength.

Zuckerberg discussed Meta's attempts to build their own "silicon" (i.e.: semiconductors) to replace Nvidia's GPUs for their AI efforts. However, he also discussed energy at length.

This was strange but also unsurprising. As the Amazon headline above demonstrates, more than a few technology leaders and thinkers are spending a lot of time thinking about how to keep the lights on.

For his part, the Facebook founder is clearly worried about bottlenecks for further technological advances developing in a world where the electricity demand to power data centers is projected to increase by 13% to 15% compounded annually through 2030.

The full interview is here:

All of this points to a strengthening of a theme that we have made for some time:

The biggest constraint on the AI productivity boom may not be semiconductor chips or trade wars or even complex supply chains largely centered around geographically vulnerable and strategically under threat Taiwan.

The real risk may come from other, far more mundane and old fashioned parts of the economy - reliable electricity and the electric grid.

i.e.: you don't just need to have the Nvidia processors and build the data centers, you also need to power them, 24/7, 365.

Normally we are pretty big fans of going AGAINST this argument. Underestimating the elasticity of supply in a well functioning economy is frequently the road to ruin.

By that we mean: if prices rise, that usually brings more supply into a market.

The idea behind this thesis is pretty simple:

If prices soar for an asset then the market will pretty quickly respond. Widgets or a commodity or whatever is in short supply will gradually and then suddenly become ubiquitous.

Markets are not perfect but they are exceptionally good at supplying what is in high demand and over time this will lead to the price for those goods coming down dramatically.

As we have already discussed, you are seeing this right now with the effort by Big Tech companies to replace Nvidia's core product: GPUs.

You also saw this during the pandemic. At the outset we didn't have nearly enough PPE or home office gear or tablets for schoolchildren to be home schooled etc. All of this was rapidly resolved and far faster than most of the original estimates.

Like all good market rules, however, there are some exceptions. The Nvidia is a good example. Building more GPUs is NOT like manufacturing more KN95 masks or office chairs. Rather it is a hugely expensive, highly technical and incredibly difficult sophisticated manufacturing process. It will take years.

Many billions will be spent by the likes of Microsoft and Meta trying to come up with viable alternatives so they do not have to pay out the nose to Nvidia. They will likely get there eventually but until them, Nvidia will keep reaping huge profits.

But other constraints are different.

For instance, as this newsletter has argued before it can take a decade or longer to open a new copper mine or oil field and almost as long (or longer in some states!) to connect a new solar field or wind farm to the energy grid.

Those constraints seem very far away from most of us working in the service economy and far away from wherever we get the power to turn on the lights or power our computers or charge our electric vehicles.

Today our argument is:

This is only getting harder and yet also strategically and economically more important.

That is unusual. It is also unfortunate.

Nonetheless, we are all going to begin discovering that electricity, something we take for granted the vast majority of our lives, is both incredibly important and very finite.


Electricity demand is skyrocketing and we are struggling to rapidly scale up electricity generation. That only means one thing for prices. And regular people may not know much about the permitting process for

What to do about it?

This is the tough one.

From one perspective if we aren't going rapidly scale energy then we will just cap how fast and how productive our economy can grow. That is a shame and also disappointing in terms of what to do about it.

But there are a few options, albeit imperfect ones:

  • We have made the argument about nuclear power and uranium before. If you missed it or are interesting you can read our most recent reprisal here.

  • We also continue to like energy more broadly. This is a sector of the stock market that has done incredibly well since 2020 and while it may have market swings we think it could continue to outperform the market and many of its competing sectors.

  • If you want a more specific play on energy companies you can look at midstream energy companies, many of which are involved in either storage or transporting energy like natural gas or oil.

If electricity demand rises and there are real constraints that lead to structurally higher prices over the medium term then those who can store it and move it will be well placed to benefit.

Companies that control the means of production will also do well but it is the middlemen with better optionality that might see the greatest profits. They will also perhaps avoid the political firing squad for longer as well.


Have questions? Care to find out more? Feel free to Download our App (!!) or reach out at or join our Slack community to meet more like-minded individuals and see what we are talking about today. All are welcome. You can also get our newsletter as an RSS feed.


    Why Do Prices In Inflationary Spirals Rise Like Rockets & Fall Like Feathers?


    What Could Replace Nvidia? Why Is That Important?