Nuclear Energy Redux: Future of AI

About a year ago, I arrived on a simple thesis:

AI growth can only be powered by an investment in Nuclear energy.

There was one underlying assumption I was making:

We mostly agree that climate change is real and we cannot keep burning fossil fuels.

My research suggested that fast forward to 2050, few things would happen in high likelihood:

1. AI and robots will perform most day to day tasks that humans will not do due to safety or labor price reasons.
2. It would require significantly more power to drive such a highly AI-driven economic fabric in the next 15 years than our power infrastructure has scaled to in the last 45 years.
3. While solar, wind, tidal and other forms of renewable energy sources will be stopgap solutions, this power hunger can only be satisfied in a scalable manner by the proliferation of nuclear energy.
4. Supporting the proliferation of nuclear energy would require the CapEx and OpEx of nuclear reactors to come down by order of magnitude while minimizing safety concerns and blast radii.

I arrived at my little truth: The future of AI and nuclear energy are intertwined.

Assuming we don't want to keep burning fossil fuels and push our planet to the brink, it is inevitable that nuclear energy and its associated investments would skyrocket like a hockey stick if AI were continue to grow the way it is growing now.

As a result of this thesis, I went long on micro nuclear reactor startups that were pushing the frontier in this space. Most notably, I kept taking long positions in NuScale Power buying on every dip for 6 months.

This stock has gone up 224% in the last year. And today, it went up 40% as news broke of 亚马逊 and 谷歌 leaning in to small nuclear reactors to serve the future power needs of their massive datacenters. While this return beats the market performance significantly, I think we are just getting started on #nuclearenergy redux.