The headlines belong to AI and semiconductors right now. Chips are soaring, data center buildout stocks are making new highs, and the momentum crowd is firmly in control. But underneath the surface of a market that looks healthy, something odd is happening. The new-low list has been outrunning the new-high list even as the S&P 500 pushes above 7,500. That's not a healthy market. That's a narrow one.
Jeff Clark of TradeSmith has seen this setup before. His read: when gains concentrate in a thin slice of the market, the rotation trade is coming. And when it does, the money that rushes out of the hot names has to land somewhere. He thinks it lands in stocks that have already been left behind—and he has three specific names in mind.
The Setup: When Enthusiasm Gets Discounted to Infinity
The bull case for AI stocks isn't fiction. Real money is flowing into data centers, chips, and infrastructure. The question Clark is asking is a different one: for how long? Once a data center is built, you don't build another one next door. Memory chips are a cyclical commodity—yet the market has priced them as if the cycle has been suspended permanently. Clark's view is that the market is extrapolating today's spending to infinity, and that a correction is overdue. That doesn't mean the AI trade is over. It means the easy money in the hot names may already be made, and the opportunity is now sitting in the stocks no one is talking about.
Figma: A Software Survivor Priced Like a Casualty
Figma (NYSE: FIG) went public at $33 a share, shot to more than $140, and has since retraced nearly all of those gains—spending time near $20 before a recent earnings pop pushed it back above $22. The surface-level read is that software is under pressure from AI, and Figma is getting caught in that tide. Clark's read is almost the opposite.
Figma isn't being destroyed by AI. It's integrating it. The platform, used by designers and product teams to build digital products and prototypes, has leaned into AI tooling rather than ignoring it, and the results are showing up in the numbers. The company's user base is growing more than 50% year-over-year, and its most recent earnings report came in at 10 cents per share against an expected loss of 17 cents. Net dollar retention has climbed to 139%, meaning existing customers are spending more. Revenue growth is accelerating, not slowing.
For Clark, the thesis is simple: the stock was never worth $140, but it was also never worth being abandoned. Near $20, it's pricing in too much fear and not enough of what the business is actually doing. His target entry is around that level, and he sees it as a name worth holding for the long run.
Kratos Defense: A Drone Pure-Play That Got Ahead of Itself
The defense budget expansion story is real, and Kratos Defense & Security Solutions (NASDAQ: KTOS) sits right at the center of it. The company's unmanned aerial systems—jet-powered drones, hypersonic vehicles, and related defense technology—have the Department of Defense as their primary customer, and that customer is spending aggressively. Kratos reported 22.6% revenue growth in its most recent quarter, with a record backlog and raised full-year guidance.
But the stock ran from roughly $35 a year ago to $120 at its peak, and then gave most of it back. It's trading near $53 today, which Clark acknowledges is not cheap on traditional metrics. This is not a value stock in the Graham-and-Dodd sense. What it is, he argues, is a growth stock with earnings expanding north of 45% annually, trading at a steep discount to where market enthusiasm put it just a few months ago.
Clark's preferred entry is closer to $45 to $50. The defense sector as a whole has pulled back from early-2026 highs as investors wait for the spending surge to show up more aggressively in earnings. Clark sees that patience as the setup. Drone technology spending isn't going away, and the pullback creates a better entry than anything available when KTOS was making headlines at the top.
SoundHound AI: Round-Trip Ticket, Better Destination
SoundHound AI (NASDAQ: SOUN) has put investors through a full round trip. A year ago, the stock was trading near $8, ran all the way to the low $20s on AI enthusiasm, and has since come back down to roughly $8. Anyone who bought near the top knows exactly how painful that ride has been.
But Clark's focus isn't on where the stock has been; it's on whether this entry price makes sense relative to what the company is building.
SoundHound's technology is the conversational AI voice layer embedded in cars, restaurant kiosks, and consumer devices—the software that responds when a driver asks for the nearest gas station or a customer places a voice order. The company is not yet profitable. What it is, Clark says, is doing the right things operationally: growing revenue, expanding into new verticals, and positioning itself as the leading pure-play on voice AI at a price point that reflects none of that potential. At $8, the stock is trading where it was before the original wave of AI enthusiasm, and the business is meaningfully larger now than it was then.
The risk is real, as profitability is still quarters away at minimum, and the stock has shown it can be volatile in both directions. But for investors who believe voice AI will become embedded infrastructure, Clark's argument is that the round trip back to $8 is exactly the kind of entry point that "buy low, sell high" was invented for.
The Bigger Picture
The three names share a common thread: each ran hard on genuine enthusiasm, pulled back further than the fundamentals justify, and now sits in the uncomfortable zone where patience is required. That discomfort is the point. The stocks generating today's headlines are priced for perfection. These aren't—and for investors willing to wait for the rotation Clark sees coming, that gap may be exactly where the opportunity lives.
The article "Investors Abandoned These 3 AI Stocks Too Early, Says Jeff Clark" first appeared on MarketBeat.
