Perhaps the biggest story in the market over the last couple months has been the mean reversion/catch-up/rotation trade between the market's "haves" (megacap tech) and "have nots" (small caps, cyclicals, value, etc). The idea behind the move is simple enough. When one segment of the market outperforms by a large margin over a long period of time (think Mag 7 since 2022), eventually traders' figure there is opportunity in that gap and decide to "go the other way" - at least for a while. This trade is definitely not new as we've seen many such rotational moves over the years.

The Argument

One of the biggest features in the market that seems to be driving this move - at least in part - is the massive "argument" going on these days over the state of the AI trade.

On one hand, you have the fast-money "masters of the universe" who run money at hedge funds and Wall Street bank trading desks. You know, the gang that defines long-term as their lunch appointment on Thursday. And the crowd that seems to play fast and loose with facts, interviews, and media publications. In this world, a trade "narrative" is all that matters. The narrative doesn't have to be true or even makes sense. No, as long as it can be easily and often repeated, the trade can live on.

This is the crowd that has decided that AI just isn't going to work - especially not for the likes of Microsoft (MSFT), or Meta (META), or Oracle (ORCL), or ServiceNow (NOW), or Salesforce (CRM), or heck, even Palantir (PLTR) - the company that just reported perhaps the best earnings in the history of software. The hedgie crowd that according to CNBC's sources at major funds on Wall Street, has made $24 billion (and counting as of last Wednesday) in profits by shorting software stocks. The crowd that has lopped a cool $1 trillion off the value of the software sector. And in the process, knocked the iShares Software Sector ETF (IGV) down -32% in 3 months.

But don't just take my word for it. According to an analyst at DA Davidson, "Hedge fund are all net short software right now." And then there's this from Goldman's Prime Brokerage unit, "This week's notional short selling in US Single Stocks was the largest on our record." Hmmm...

We've Seen This Movie Before

If memory serves, this is the same crowd that tried to sell us the AI bubble theme at the beginning of last year. And of course, that trade didn't work out so well and the performance of a great many hedge fund geniuses suffered because of it. Turns out that calling a top into a generational change in technology can be tricky!

Yet undeterred, here they are at it again. In the face of record earnings. Record revenues. Record growth. And of course, record expenditures on hardware investments. It's that last part where our furry friends in their bomber jackets seem to be making a stand - and profiting from their narratives.

The Issue

At issue here is the idea of what's called Capex, which is short for capital expenditures. Cutting to the chase, there is an arms race happening in AI. A race for dominance as the current thinking is that the next generation of computing will be a winner-take-all, or at the very least, a "winner-take-most" situation.

The idea is that the AI arms race is going to end up the same way the competition for internet search did a generation ago. Think back to the early days of the internet when folks used more than Google (GOOGL) or Siri to find stuff on the internet. Everybody knows Google has dominated search since then and printed a gazillion dollars in high-margin profit along the way.

The current bearish narrative is that AI will wind up looking similar - with one dominant player. Personally, I disagree vehemently with this concept as "AI" isn't a product you buy off the shelf. No, I see AI as the next generation in computing where there won't be just one winner or dominant player. There will be all kinds of winners and losers along the way. Because, the times they a changin'!

You can already see this happening with OpenAI. Three years ago, ChatGPT was all the rage and Sam Altman and friends appeared to be the only real player in the game. But then came DeepSeek. And Gemini. And Claude, Grok, etc. And don't look now fans, but the betting markets list ChatGPT as a distant third in the list of AI LLM usage by mid-year (FTR, Gemini is the now the big leader with around 70% and xAI is in second place).

I'm of the mind that there will be all kinds of winners in the next five years in all sorts of businesses. I.E. AI is so much more than just LLM (Large Language Model) chatbots.

But I digress. Let's get back to the "argument" and that arms race.

Record Spending

Below are two charts that detail the capex spending by big tech. If you are going to participate in this argument, you'd best know the details.

The first chart shows the level of capex spending from 2021 through 2025. The key is that capex really began accelerating in 2024 and 2025. For example, Amazon.com went from $10B in mid-2023 to $40B in 2025. Microsoft (MSFT), Meta (META), and Alphabet (GOOGL) all at least tripled their capex spending. Talk about going all in!

The next chart is the base of the bears' argument. The table shows the dollar amounts each company plans to spend in 2026 alone. If my calculator is correct, the total from these six companies is $735.5 Billion (and this doesn't include Elon Musk's xAI). Yowza!

To put into perspective how large this spend is, Bloomberg writes, "The largest US-based automakers, construction equipment manufacturers, railroads, defense contractors, wireless carriers, parcel delivery outfits, along with Exxon Mobil Corp, Intel Corp., Walmart Inc. and the spun-off progeny of General Electric — 21 companies — are projected to spend a combined $180 billion in 2026."

The bears - you know, the ones shorting AI tech like crazy this year - tell us this is absolutely nuts. That this level of capex spending can't possibly pay off. That the big firms spending all this money simply don't know what they are doing. And that this will end in tears for them all.

Never mind the fact that recent earnings reports generally show unprecedented demand for the current AI and data center services. Demand that at present, can't be met. Demand that according to what I've read, appears to be growing exponentially. Hence the need to expand the ability to meet the demand by, yep, you guessed it, making capital investments.

But Wait, There's More...

While I respect a good argument (I am an analyst with two good hands after all), I don't have any use for folks thinking they can predict the future. Or call the tops or bottoms of major market/macro moves. Especially when they seem to be talking out of both sides of their mouths.

You see, while our furry friends tell us that AI is going to be a ginormous bust, they told us just last week that AI will, wait for it... destroy the software industry as we know it.

So, which is it? The huge capex spending on AI can't possibly pay off - or - AI is such a powerful tool that it will destroy companies like Microsoft (MSFT), Salesforce (CRM) and ServiceNow (NOW)?

The bottom line is that the traders have been hitting the sell button early and often lately - for a trade. They contend that (you may have heard this one before) the sky is falling - especially in tech! (No really, it's actually going to fall this time!)

The masters of the trading universe contend that they have picked the exact moment that the so-called bubble will burst. So you'd best sell everything you can now - before it goes to zero. (Ok, yes, I'm guilty of being perhaps a tad sarcastic here.)

Time to Pick a Side

So, who has it right? The CEOs of the world's biggest companies or the trader bros running big bank trading desks and hedge funds?

As an "investor" and NOT someone placing a bet for the next 5 days, I'm of the mind that you need to decide if you believe the bear narrative(s). If you believe that the CEO's of Microsoft, Alphabet, Meta, Amazon, et al have become too emotional in the race for dominance, then by all means, get out of those AI-oriented positions and move on to Coca-Cola (KO), Procter & Gamble (PG), and Wal-Mart (WWMT). Or maybe play it safe and plunk your portfolio into SPY and AGG until the dust settles.

On the other hand, if you are a growth investor looking to own the fastest growing companies, you may want to either (a) continue holding your AI-oriented positions or (b) do some dip-buying in the names that have the best fundamental outlooks.

Thought for the Day:

Instead of just muddling through, why not make a concerted effort to enjoy your day?

Wishing you green screens and all the best for a great day,

David D. Moenning
Founder, Chief Investment Officer
Heritage Capital Research, a Registered Investment Advisor

Disclosures

At the time of publication, Mr. Moenning held long positions in the following securities mentioned: MSFT, META, TSLA, AAPL, ORCL, WMT - Note that positions may change at any time.

NOT INDIVIDUAL INVESTMENT ADVICE. IMPORTANT FURTHER DISCLOSURES