The market spent the better part of two years rewarding almost anything connected to artificial intelligence, and investors stopped asking whether businesses were actually improving, they started asking how much AI exposure a company had and buying accordingly. That works until it doesn’t. This is why the recent tech selloff feels less like a collapse in fundamentals and more like a violent correction in expectations, and when that happens, I look for businesses still executing while everyone focuses on short-term disappointment.
CrowdStrike: Investors Are Selling The Future, Not The Business
CrowdStrike (NASDAQ: CRWD) dropped sharply after reporting revenue growth of 26%, EPS growth of 51%, record operating cash flow of $384 million, record free cash flow of $279 million, and raised full-year guidance — a combination that reads like exactly what investors claim to want, which makes the reaction worth understanding rather than simply accepting at face value.
The frustration has less to do with endpoint security and more to do with what comes next. Management keeps talking about a second act built around autonomous systems and AI-driven security operations, and investors who had already priced in early confirmation of that transition sold when the quarter delivered strong current-period results without material evidence that autonomous security is generating meaningful revenue yet.
After rallying from roughly $350 in March to nearly $780 before earnings, the stock had become one of the market’s biggest winners. Expectations were stretched. When the company delivered another strong quarter rather than a transformative one, investors took profits.
Yet the longer term picture remains constructive. Shares remain well above the 50-day moving average near $521 and the 200-day near $479, which is not where a stock trades when the money that matters has decided to leave. The market is questioning how valuable CrowdStrike’s next growth engine becomes. It is not questioning the quality of what’s already built — and that distinction is where the opportunity lives.
Cadence Design Systems: The Toll Booth Nobody Talks About
Most investors debating the AI race spend their time arguing over which chip company wins. Cadence Design Systems (NASDAQ: CDNS) gets paid before that race even begins.
Every advanced semiconductor requires sophisticated design software before it can be manufactured, tested, or deployed at scale. And as AI models grow larger and more computationally demanding, the chips required to run them grow more complex in direct proportion, making Cadence’s electronic design automation tools more critical to players like Nvidia (NASDAQ: NVDA), Advanced Micro Devices (NASDAQ: AMD), Broadcom Inc (NASDAQ: AVGO), Marvell Technology (NASDAQ: MRVL) and others in the ecosystem.
The stock surged from roughly $270 in April to more than $410 before pulling back toward $380. A correction after a move like that should surprise nobody. What matters is where the stock corrected to. Shares remain comfortably above the 50-day moving average near $336 and the 200-day near $325, and volume during the decline never suggested the kind of institutional distribution that precedes a genuine breakdown. The chart looks like a stock digesting a powerful breakout, not beginning a downtrend.
AI companies compete with each other. Chipmakers fight over market share. Cloud providers battle for dominance. Every single one of them still needs increasingly sophisticated chip designs to pursue any of those ambitions. The market sees a software company. I see a toll booth collecting from every lane of the AI highway simultaneously — and toll booths don’t lose relevance when traffic increases.
Arista Networks: The Infrastructure Play Everyone Keeps Overlooking
Investors love talking about GPUs. They spend far less time discussing what happens after the GPUs arrive. What happens is that data has to move, constantly, at enormous scale, between servers that need to communicate faster than the chips themselves can process. That is where Arista Networks (NYSE: ANET) sits, providing the networking infrastructure that allows massive AI clusters and data centers to function at the performance levels the workloads actually demand. As AI models grow larger and clusters scale wider, networking becomes more important rather than less, because moving data efficiently between compute nodes increasingly determines overall system performance.
The recent selloff dragged Arista lower alongside names facing entirely different structural risks, as investors conflated AI infrastructure spending concerns with businesses that sit in fundamentally different positions within that spending chain. Shares pulled back from approximately $180 toward the $150 area, but remain above both the 50-day moving average near $153 and the 200-day near $141, with the longer-term uptrend structurally intact and the recent weakness resembling consolidation rather than conviction selling.
The market is treating Arista as another AI stock exposed to the same risks as the names driving the narrative. The business is considerably more durable than that framing suggests, because as long as data centers keep expanding, networking is not discretionary spending that gets deferred when budgets tighten. It is the infrastructure that makes everything else function.
The Setup Across All Three
CrowdStrike’s business remains strong while the market debates its second act. Cadence keeps collecting from every company building the next generation of chips regardless of who wins the competition between them. Arista sits inside one of the most critical bottlenecks in modern computing infrastructure with a position that strengthens as the workloads it supports grow more demanding.
The selloff created uncertainty. But for investors willing to look past the short-term narrative and focus on what the businesses are actually doing, it may also have created the entry that the run-up never offered.
