Is AI's mega-rally sustainable?

Global equity markets today are at their most concentrated for over 50 years. The rush for AI exposure has pushed the mega-cap technology names ever higher, which due to their ever-increasing size, means there has never been a time when the largest stocks account for so much of the world’s equity indexes.

Historically, spikes in market concentration have always unwound. We believe we might be seeing green shoots of that today.

Magnificent utilities

When reviewing year to date returns, it should come as no surprise that Nvidia and another tech stock SBC are amongst the best performing securities in the S&P 500 Index. What perhaps is surprising, is that three electric utility companies are also amongst the best stocks to have owned so far this year.

Aside from the remarkable news that a utility has appreciated more than 100% within 5 months, it’s also worth highlighting the move in copper which is up over 25% year to date. To some extent this explains why the S&P500 has slightly outperformed the tech-heavy Nasdaq year to date.

AI needs (a lot) more power

The AI theme is broadening. Investors have realised how energy intensive AI actually is, with the chart below from Dominion Energy highlighting just how much more energy is required for an AI equipped datacentre, versus a legacy version.

When combined with an underinvested energy generation and transmission grid, it’s no wonder forecasters are calling for a boom in capital expenditures in the utilities sector. This represents investment in power plants & generation, transmission lines, infrastructure… steel, concrete, power management technology and infrastructure… many ‘old economy’ industrials might just do okay as well out of AI.

Market broadening

The last time the market was anywhere near this concentrated was during the tech bubble, from which the subsequent collapse saw a powerful reversal of market concentration.

“This time is different” is a refrain we often hear when making comparisons to March 2000. The companies driving the current boom are universally regarded as awesome, with balance sheet strength, huge profits and monopolistic market positions. We don’t disagree that the mega-cap tech names are for the most part, magnificent.

However, there are laws regarding large numbers and the challenges of maintaining super-normal growth rates. It just becomes much, much harder to maintain high growth on already super-sized revenues. Forecasts from RBC suggest that earnings growth differential is set to slow in coming years.

When put together, the broadening of the AI demand thesis, plus the increasingly high hurdles that mega-cap growth companies (of any persuasion) have in terms of maintaining high growth rates, we don’t believe this time is different. At least from a market concentration unwind perspective.

When high market concentration unwinds, equal weighted index exposures always outperform cap-weighted exposures. The mega-cap tech names don’t need to crash like Cisco and AOL did 20-odd years ago, they simply need to underperform relative to the rest of the market.

In our view, the only thing better than an equal weighted exposure in the current environment is an equal-weighted exposure paired with a cap-weighted short (we are short index options via the buywrite overlay). While this combination has been a powerful headwind for our Global fund returns the past 12-18 months, as market concentration unwinds we believe it may prove a powerful tailwind for future years.