Are we in an AI bubble, or is the boom just getting started?
This week on Ticker Take, we put that question to Gil Luria, Head of Technology Research at D.A. Davidson. Luria doesn’t think we’re in one, at least not yet. The AI labs’ models are already good enough to drive real economic activity, he says, and that keeps expanding as usage grows. Big tech’s AI-exposed stocks, from Nvidia to Microsoft to Meta, are trading at low multiples relative to their own history, which is hard to square with bubble pricing.
But he pointed to a few risks. Some individual assets, like the very high price tags already attached to private AI labs ahead of a possible IPO, may be running ahead of reality. And the AI buildout is shifting from cash-funded to debt-funded, since the biggest spenders have started to exhaust their own cash flow. Leverage, he says, is “one of the common things to most bubbles.” That doesn’t mean we’re in one now, but the risks are rising.
With that in mind, here are three stocks Luria likes if we’re not in a bubble, and three if we are.
No Bubble: Micron (MU)
Micron’s business has changed as much as Nvidia’s has, according to Luria. He just feels the company gets less credit for its transition. Its memory chips used to be a commodity, mostly for data storage. Now, they are essential to AI inference: more memory means a bigger context window and a faster model. Luria argues that shift isn’t priced in. Micron trades around six times earnings; if growth continues, he sees room for the multiple to reach 20 times. And because the company has locked in roughly half its business three to five years out, he sees limited downside even in a slowdown.
No Bubble: Oracle (ORCL)
Oracle’s edge, in Luria’s view, is leverage. If AI demand keeps outpacing available compute supply, Oracle stands to benefit more than its peers because of how much debt it’s taken on to build out capacity. He notes Oracle’s compute backlog is larger than Amazon’s or Google’s, yet the market is pricing that backlog as close to worthless. The worry that it won’t close is one of the things that’s been holding the stock back.
No Bubble: Palantir (PLTR)
Luria calls Palantir simply “the best company in the world” right now. Its core product combines data from disparate sources without compromising privacy, and its army of deployed engineers helps customers get results from AI faster than competitors can. That shows up in the numbers: recent revenue growth of more than 90 percent, still accelerating, a pace Luria says no other software company can match.
Bubble: Apple (AAPL)
Apple’s case doesn’t depend on AI getting any better than it is today, Luria says. People will keep buying iPhones and Macs regardless, and whatever AI they do use will likely run through an Apple device. He also points to a leadership shift: incoming CEO John Ternus comes from the hardware side, and Luria expects a wave of new hardware, including a foldable iPhone and possibly a glass iPhone, that has nothing to do with AI at all.
Bubble: Meta (META)
Meta’s ad business is already strong, growing more than 20 percent and taking share from Google. The AI spending is the wildcard: Luria notes Meta is pouring hundreds of billions into AI without yet fielding a top-tier frontier model. If the AI trade cools and that spending pulls back, he expects cash flow, and the stock’s multiple, to jump.
Bubble: Palantir (PLTR), again
Palantir shows up on both lists. Even if AI stops improving today, Luria says the company doesn’t need better models or more compute to keep helping customers. It just needs to keep doing what it already does well.
The Ticker Take
Instead of picking a side in the bubble debate, Luria is looking for companies that hold up no matter which side wins. It’s a playbook worth considering, especially if that debate keeps playing out in the headlines.
Jon Erlichman is a BNN Bloomberg contributor and the host of Ticker Take on YouTube.
