AI winners hit a higher bar as investors ask tougher questions

Key takeaways

• Strong earnings are not always enough when valuations already expect near-perfect execution.

• Artificial intelligence is boosting demand, but also lifting spending, competition and investor scrutiny.

• The key lesson is to separate business quality from share-price expectations.

Good earnings used to be the easy part. A company beat estimates, lifted guidance, smiled politely, and the share price usually did the rest. Not this week.

Broadcom, Palo Alto Networks and CrowdStrike all reported strong results around 2 and 3 June 2026. Each sits close to one of the market’s biggest themes: artificial intelligence (AI), cybersecurity and the infrastructure needed to run the digital economy. Yet their shares fell after the numbers.

That is the useful lesson for investors. The market is not saying these businesses are broken. It is saying expectations are no longer cheap. When a theme becomes popular, investors stop asking “is this growing?” and start asking “is it growing fast enough, profitably enough, and clearly enough to justify the price?” Wall Street can be a demanding dinner guest.

The bar has moved higher

Broadcom makes chips and infrastructure software. In simple terms, it supplies some of the parts that help data centres move, process and manage huge amounts of information. That matters because AI is not magic in the cloud. It is electricity, chips, networking equipment and software, stacked together at industrial scale.

Broadcom’s fiscal second-quarter results showed the strength of that demand. Revenue rose 48% to 22.2 billion USD, while AI semiconductor revenue jumped 143% to 10.8 billion USD. The company also guided for AI semiconductor revenue of 16.0 billion USD in the next quarter.

Those numbers are not small. They are “finance department needs a bigger spreadsheet” numbers. But the shares still fell sharply in extended trading after revenue slightly missed Wall Street forecasts and the company kept its longer-term AI sales target broadly unchanged. That was enough to disappoint a market that had already priced in a lot of good news.

This is the first common thread. In AI infrastructure, investors are rewarding acceleration, but they are punishing any hint that expectations have run ahead of reality. Broadcom is still growing fast. The question is whether the share price had already assumed even faster growth.

Cybersecurity is strong, but not simple

Palo Alto Networks and CrowdStrike tell the same story from a different corner of technology. Both companies help organisations protect their systems, data and users from cyberattacks. As AI spreads, this job becomes more important. More digital tools create more doors. Cybercriminals, sadly, do not take long summer holidays.

Palo Alto reported fiscal third-quarter revenue of 3.0 billion USD, up 31% from a year earlier. Its next-generation security annual recurring revenue, a measure of repeat subscription income from newer security products, rose 60% to 8.1 billion USD. Remaining performance obligation, which is contracted revenue not yet recognised, rose 36% to 18.4 billion USD.

Again, strong numbers. Yet the stock fell. Part of the issue is that acquisitions boosted growth, including CyberArk and Chronosphere. Acquisitions can make a company stronger, but they also make the picture harder to read. Investors want to know how much growth comes from the original business and how much comes from buying another one. When that line becomes less clear, the market often reaches for a discount.

CrowdStrike also beat expectations. The company reported first-quarter revenue of 1.39 billion USD and ending annual recurring revenue of 5.51 billion USD. It also raised its full-year revenue outlook. But its shares fell in extended trading after operating expenses rose 15%, partly as the company invests more in AI and product development.

That is the second common thread. AI is not only a revenue opportunity. It is also a spending cycle. Companies need engineers, data, infrastructure, security tools and new products. The market likes investment when it creates growth, but it still wants proof that the growth will turn into durable profits.

The market reaction is not as strange as it looks

For retail investors, the most important point is that share prices react to expectations, not just results. A good quarter can disappoint if investors expected a great one. A great quarter can disappoint if the valuation already assumed perfection. This is why earnings season can look irrational from the outside and painfully logical from the inside.

Broadcom shows how AI hardware expectations have become extremely high. Palo Alto shows how cybersecurity investors now care about the quality and visibility of growth, not just the size of it. CrowdStrike shows that even strong recurring revenue can be overshadowed by rising costs if investors worry about margins.

The broader industry message is clear. AI remains a powerful demand driver. Data centres need chips and networking. Companies need better security. Customers want platforms that do more with fewer tools. But the easy phase of the trade may be maturing. The next phase is about execution, pricing power and cash flow, not just attaching “AI” to a slide deck and hoping the stock market applauds.

Risks to watch

The first risk is valuation. If a company trades at a high price because investors expect rapid growth, even a small disappointment can cause a large move. The early warning sign is a stock falling on good news.

The second risk is spending discipline. AI can improve products, but it can also raise costs before it raises profits. Investors can watch whether operating margins, free cash flow and hiring trends support the growth story.

The third risk is visibility. For acquisitive companies such as Palo Alto, investors need to understand what is organic growth and what is bought growth. Both can be valuable, but they are not the same thing.

Investor playbook

• Watch whether guidance rises faster than expectations, not only whether it rises.
• Compare revenue growth with margin trends to see whether scale is improving.
• Separate AI demand stories from AI profit stories. They are cousins, not twins.
• Treat sharp post-earnings moves as information about expectations, not automatic verdicts on quality.

The price of perfection

The lesson from Broadcom, Palo Alto and CrowdStrike is not that strong technology companies suddenly became weak. It is that the market has become more selective. AI and cybersecurity remain important long-term themes, but popular themes can carry heavy expectations.

For investors, the useful question is not whether a company has exposure to AI. Many do. The better question is whether that exposure creates profitable growth at a price that already leaves room for imperfection. Good businesses can still be expensive stocks. In a hot market, even excellent earnings may need to bring their own fire extinguisher.

This material is marketing content and should not be regarded as investment advice. Trading financial instruments carries risks and historic performance is not a guarantee of future results.