ASML Guidance Signals Durable AI Capex

ASML’s second-quarter results showed exactly that. The Dutch company reported €9.3 billion in net sales and €2.9 billion in net income, then raised its 2026 outlook to €43 billion to €45 billion in sales with gross margin of 54% to 56%. Just three months earlier, it had already been looking for €36 billion to €40 billion in sales and 51% to 53% margins. That kind of revision is what you want to see if you own a company with a deep moat and a nearly irreplaceable role in the supply chain.

Why does this matter economically? Because ASML’s machines are not ordinary factory equipment. They are the ultra-precise lithography systems that etch the tiniest features onto semiconductors, enabling the high-performance chips used in AI training and inference. Each one can cost around $400 million, which sounds extreme until you remember that these systems are the bottleneck for the most valuable chips on earth. If chipmakers and foundries are willing to keep buying them, it tells you the capex cycle is still alive.

The broader AI buildout is reinforcing that message. Industry commentary now suggests the bottleneck in AI has shifted from chip design toward advanced packaging, a sign that the buildout is becoming more complex, not less. Nvidia is still the headline name in AI chips, but AMD and Intel are also pushing deeper into long-term server chip contracts, while the infrastructure layer — from cooling to data-center capacity — is expanding alongside it. That is good for the entire semiconductor ecosystem, but especially for the companies that provide the picks and shovels.

ASML’s stock action shows investors are already voting with their feet, even if the path is choppy. The shares have climbed sharply over the past year, but the latest trading data show some short-term volatility, with the stock slipping from recent highs and the 14-day RSI sliding toward oversold territory. The 50-day moving average remains well below the longer-term trend, which suggests the bigger uptrend is intact even if momentum has cooled. For long-term investors, that kind of pullback often matters less than the underlying earnings power.

And that is the real story here: ASML is not just participating in the AI boom — it is monetizing the bottleneck. Companies like Nvidia may capture the headlines, but ASML captures the manufacturing necessity. As long as AI infrastructure keeps scaling, advanced chips will need more lithography capacity, more packaging, and more capital spending. That supports a durable earnings model, not a one-quarter trade.

There are risks, of course. Semiconductor spending is cyclical, export controls can complicate sales, and any slowdown in AI capex would eventually hit the whole chain. But the latest numbers suggest the opposite is happening: demand is broadening, not fading, and the companies closest to the physical making of chips still have pricing power.

For investors thinking in years, not weeks, ASML remains one of the most compelling ways to own the AI buildout. If you want exposure to the long runway in semiconductors without betting on any single chip designer, this is still a name worth watching — and holding for the long term.