Summary:
During a recent episode of The AI Investor Podcast (full episode below), 24/7 Wall St. Analyst Eric Bleeker discussed several semiconductor companies that could benefit from the next wave of AI-driven data center expansion.
Among the stocks highlighted by Bleeker is Navitas Semiconductor, which appears to be in revenue decline but may see a major turnaround as new Nvidia data center architectures drive demand for advanced power management. Wall Street forecasts show Navitas revenue potentially rebounding sharply by 2027 and 2028 as the company pivots away from low-margin businesses and toward data centers.
Another stock Bleeker recommends exploring is On Semiconductor.
“This one’s a $22 billion company,” Bleeker explains. “They’re relatively flat year to date. But EPS estimates from Wall Street have a strong rebound: $2.91 this year, $4.03 in 2027, $5.36 in 2028. That’s a very nice curve, and it’s because there is supposed to be a rebound in automotive, which is the primary end market now for a lot of these applications, and industrials as well. So they’re going to benefit from that.”
Watch the video above or read the full transcript below for all five of Bleeker’s recommendations.
Transcript:
Eric Bleeker: I’ve got five stocks. I’m gonna go through each relatively quickly. I want to talk about what opportunities look like. So we’ve got Navitas Semiconductor (NASDAQ: NVTS). They’ve got a $2.3 billion market cap. They’re up 19% year to date, 345% across the past year. Keep in mind, we’re lapping the tariff situation, so that might influence it, but they’re down across the past five years.
Here’s the background I’m talking about. Their revenue was $83 million in 2024. It dropped to $46 million in 2025. It’s supposed to go—
Austin: It’s gonna go down to $38 million, right? This looks like a business in terminal decline. A hundred percent. Three years where your revenue is down more than 50%.
Eric Bleeker: But people are catching on to what exactly this power situation, what this new architecture from Nvidia (NASDAQ: NVDA) | NVDA Price Prediction with Rubin is going to mean. So what’s gonna happen is Wall Street’s now estimating they’ll jump up to $65 million in 2027 and $122 million in 2028. Austin, what’s going on right now is they’re shedding their low-margin business and moving toward data centers. They’re doing a pivot to get away from their old business and toward their new business. So this is one example of what this ramp is going to look like because we’re gonna see Rubin really come into effect in the second half of 2027 and into 2028. The businesses right now that have zero revenue are going to see a relatively steep scale.
So another business I want to talk about is On Semiconductor (NASDAQ: ON). This one’s a $22 billion company. They’re relatively flat year to date. But Austin, EPS estimates from Wall Street have a strong rebound: $2.91 this year, $4.03 in 2027, $5.36 in 2028. That’s a very nice curve, and it’s because there is supposed to be a rebound in automotive, which is the primary end market now for a lot of these applications, and industrials as well. So they’re going to benefit from that. This stock is also down 20% since mid-February. You’re looking across all these names in the AI space going, okay, in these key trends everything’s running. This is an opportunity to get companies with revenue growth mostly focused in 2027 and 2028 and get it at a starting point today.
So On Semiconductor, Austin, is extremely diversified. There’s very low contribution from data centers today, but what it can become is a third tent pole to their business. They’ll have industrials, automotive, and pretty soon data centers. You’re getting it at a relatively opportunistic cost.
Austin: This whole wave right now feels very similar. People could be forgiven for being skeptical of this thesis. Let me explain. You’re saying here are companies that are really beaten down. They’re heavily exposed to the EV space, but as we move into this power management space, they’re going to benefit from these mega data centers because they have better power management technology or they play into the supply chain in an important way. They’re pre-revenue, so we’re kind of buying the rumor. That might seem risky. But let’s look at something we talked about just a few months ago: turbines. Not an incredible tsunami industry like this, but all this capital makes for strange bedfellows. Turbines are an industry where you previously would have invested for aviation, but actually the power generation aspect for data centers made them a really appealing investment. If you realized they were going to be power generation on site, you could have done fantastically well. So this feels very similar to me.
Eric Bleeker: Correct. And the other things we’ve done is we invested relatively at a trough in memory and said we believe this is going to take off. We want to get in before it takes off. Optics—many of these stocks were down 50% a year ago, but we saw the curve was coming because of new architectures and developments in data centers. This is a call right now that there is a massive architecture shift that the market is still fundamentally mispricing. Zero of it has been realized, so you’re able to buy in a relative trough before this acceleration happens and becomes blindingly obvious to the market that this is a major trend.
We’ve done this a few times in other industries. This is our attempt to do it again in something new. As much as it always feels late in AI, there are always new things emerging, and this is one of the most exciting opportunities today. So let’s do a few more.
Wolf speed—we got a great email on this from Ari. I’m going to read it in full on our next episode. They’re a pure play on silicon carbide. It’s one of these materials we’ve talked about that is going to be increasingly important for this kind of new power management need with 800-volt data centers. They mine the material, slice it into wafers, and fabricate it. I think last quarter they said data center revenue was up 50% sequentially. So you can start seeing the curve that’s happening, moving away from this EV market and toward data centers.
Next is Aehr Test Systems (NASDAQ: AEHR). If you’re listening to this instead of watching the podcast, it might sound like a weird ticker, but it’s AEHR. It’s a $1.2 billion company. It’s up 75% year to date. It’s got the same curve where it had $66 million in revenue in 2024, $59 million in 2025, $47 million expected this year, and then a massive V expected in 2027. Again, it’s because it’s shifting from EV end markets to new kinds of data center demand.
Aehr Test Systems does test and burn-in equipment for these new kinds of compounds we’re talking about, and they’ve recently had wins in optics, memory products, and even the custom ASICs we’ve talked about. The amount of wins in the right kinds of industries—I can’t think of another stock that has this many.
The question is just how big their revenue can get. But I think they’re going to start moving. They’re moving already this year with people seeing them as an optics play and for ASICs.
Austin: This is a company in 2028 that could do over $100 million in revenue easily. If you look at their 2024 revenue, they’re at $66 million, then 2025 $59 million, 2026 $47 million. Again, this looks like a business in structural decline, but there’s 75% expected revenue growth next year in 2027. That would take them up to $82 million, way above their 2024 revenue figure, and even 50% growth from there puts them well above $100 million in 2028.
Eric Bleeker: Yeah, and there’s a non-zero chance to have an extremely steep revenue ramp beyond anything Wall Street expects. Then you get those huge multiples based upon extremely steep growth. This thing is up 75% year to date, but it’s up for a reason because almost no stocks have the structural tailwinds of industries that are winning the way this stock does.
Again, I don’t think anyone is really pricing in much of this power management situation we’re talking about, which I think can be really big. I think they’re moving on things like optics right now.
Lastly, STMicroelectronics (NYSE: STM). It’s a $30 billion company. Same thing. This is a company that made $4.46 in EPS at the peak of EVs in 2023. Right now, in the last year, that number has crashed to $0.53. But they’re going to have an opportunity to turn it around. Like On Semiconductor, they’re mostly industrials and autos, but they’ve got photonics as a catalyst as well and a decent-sized power semiconductors business.
Right now they’re not seen as an AI winner. If anything they’re kind of seen as a structural loser, but they’re going to get to flip this narrative around and they’re extremely large and diversified. So Austin, that’s it. I just wanted to do a quick tour of this space.
Top Gaining Stocks
Top Losing Stocks