When Excavator Makers Become Tech Winners, Something Is Wrong

The Nasdaq, up 11%, and the Dow Jones, up 10%, have also delivered first-half performances that would look respectable even in calmer times, let alone in such turbulent ones.

What explains this rally?

The Semiconductor Engine

Much of the answer lies in SOXX, the exchange-traded fund tracking semiconductor stocks. Since the start of the year, it has gained more than 80%, although investors have taken some profits over the past two days.

Its largest holdings include Micron, AMD, Nvidia, Broadcom, Intel and Applied Materials. Micron alone has gained more than 200% this year, while AMD is up more than 131%. The first half of 2026 was therefore the strongest in SOXX’s history. Its message to investors is simple: right now, almost nothing outside this sector is worth owning.

But even this market has its problems. SOXX has weakened sharply over the past three days and investor sentiment remains highly volatile. The latest shift came after reports that Meta plans to resell unused computing capacity from its data centers.

At first, the news lifted Meta’s stock sharply. Within 24 hours, however, the same story began to look very different. If Meta has enough spare capacity to sell it, that also suggests something more troubling: it does not have enough internal demand for it.

In other words, Meta appears to have built more computing power than it can use itself. The news suddenly raised a question the market did not want to hear.

What if today’s investment boom is not based only on current demand, but also on demand that has yet to materialize? What if part of the investment in AI is beginning to resemble the classic overinvestment cycle the semiconductor industry knows so well? At first, everyone believes capacity will always be scarce. So they build, order and pay at almost any price. In the end, it turns out that there is suddenly more capacity than actual customers.

A Warning from Wall Street

Then came Michael Burry, the investor made famous by The Big Short. On his Substack account, he published a list of positions he is currently shorting – stocks he expects to fall significantly. The list was long: Micron, the SOXX ETF, Nvidia, Applied Materials and Caterpillar among them.

Burry was essentially repeating what he has been saying for much of the past year. In his view, the developments around AI increasingly resemble a bubble, and not just any bubble.

He sees the main weakness in circular investment, with some tech companies sending money to others, which then use it to buy chips, cloud services and computing power. On the surface, the whole system looks stronger than it really is. The real problem emerges once it becomes clear that part of this demand is not being driven by real AI applications, but merely by the need to keep the narrative alive.

As usual, Burry is certainly onto something, and his reasoning is highly rational. The problem is timing. In recent years, he has got it wrong several times in a row. Still, Burry can afford a premature call. Across his career as a whole, he remains comfortably in profit. One mistimed bet on a market decline will not bring him down.

The Unexpected Winner of the AI Boom

The most interesting name on his list is Caterpillar. The company is familiar from countless American construction-site scenes: yellow excavators, bulldozers, generators and heavy machinery.

At first glance, it is about as far removed from artificial intelligence as a concrete mixer is from a language model. Caterpillar used to be one of those sturdy industrial stocks that formed part of a portfolio’s stable foundation.

It is a major company with strong revenue and a proven business model, making bankruptcy – and therefore a fall in its share price to zero – highly unlikely. The only drawback used to be its very stability.

That has changed with the arrival of artificial intelligence. Caterpillar’s stock has risen by more than 61% since the start of the year. For a company of this kind, that would once have been almost unthinkable. Demand for new building projects has certainly not doubled.

On the contrary, high interest rates in the US are holding back construction. Caterpillar, however, has found a new line of business: selling backup power generators for huge data centers built specifically for AI.

When even a company such as Caterpillar starts trading as a winner of the technological revolution, it is clear that the story has gathered enormous momentum. Not because Caterpillar is a bad investment, but because the market has started looking for winners in places where, only two years ago, almost no one would have thought to look.

At that point, investors are no longer buying only the prospect of future profit growth. They are also buying the abstract idea that a single successful division can change the fundamental nature of an entire corporation.

But even in the age of AI, Caterpillar will not stop being Caterpillar. It will still mainly be a company that makes money because someone, somewhere, needs to build, mine or move something.

So all summer, markets will be watching closely to see who is right: the eternal pessimist Burry or the optimistic Donald Trump, who has clearly built his policies and political campaign on the premise that his administration is economically successful because Wall Street is trading at historic highs.

For Trump, this is a very bold and dangerous game. The higher and faster stock market indexes rise, the greater the risk of a brutal crash.