Preface
The second section from our annual review to be broken out is our commentary the intersection of AI-linked Capex and fiscal spending. This is certainly a topic that we will continue to discuss in the future and we are monitoring it closely on a daily basis, it really is right up our alley.
Fiscal/AI Thematic Additions: April 1st 2024
Status: Intact & Expanding Thesis
Takeaways:
POWER GENERATION NARRATIVE HAS MULTIPLE TAILWINDS BUT HAS BEEN EXAGGERATED BEFORE
PATH MATTERS IN TECHNOLOGICAL ADVANCEMENT, CAN'T SKIP STEPS
UTILITY STOCKS OVERVALUED, POOR RISK/REWARD COMPARED TO OTHER OPTIONS
FOCUS ON AI EFFICIENCY GAINS AND POWER DEMAND BENEFICIARIES SIMULTANEOUSLY TO MINIMIZE RISK
INVEST IN ENERGY NODES AND CONNECTIVITY OVER STRAIGHT TRANSMISSION
At the end of Q1 we had some interplay between our Fiscal Spending basket and our AI basket as we turned our attention to what the government might need to prioritize in order to facilitate the proliferation of AI data centers and projected demand.
This dovetailed nicely with our point made in September 2023 that the US grid was reaching the point where it would no longer be able to support technological advancement. Its dated nature made it likely that regardless of who was in office or which party held the senate or house they would have to prioritize investments in electricity. Whether that was renewables and electrification or revamping the grid to reinforce our economic growth.
As the quote from a couple months ago makes clear, I was skeptical that this was a truly sustainable narrative despite already being bullish on the names for other reasons. Buying integrated utilities on an AI thesis just seemed…insane. But also, if it worked, very asymmetric. Seeing Vistra (VST US EQUITY) become highly correlated with NVDA was one of the most amazing things this year.
So, overall, I am very happy that I did not let my skepticism inform whether or not I added these names to our basket, because...
But to say I am no longer skeptical is a bit too far. This episode in history from the late 1990s rings in my ears when I hear the discussion:
As does the resultingly unimpressive curve for power demand since that was written...
But hey, I am not one to shy away from a good narrative. And It is no secret that electricity use and GDP growth are closely correlated, which makes the idea that AI is going to drive significant gains in power generation pretty astoundingly bullish if correlations hold..
So I wanted to explore this overlap, because the fact is that I am bullish on these companies and grid investment via government spending, AI or no AI.