Ten Stocks
for the
AI Buildout

If you want to understand where the equipment money is actually flowing inside semiconductor manufacturing, the answer right now is advanced packaging. That's where chips get stacked on top of each other in three dimensions to achieve the density that AI workloads demand. And the company that inspects and measures those packages, making sure they're built correctly, is Onto Innovation.

Think of it like a quality control station on a very expensive assembly line. Every chip that gets packaged for AI needs to be checked. Onto makes the tools that do the checking.

The company recently crossed $1 billion in annual revenue for the first time. More importantly, one memory customer nearly doubled its backlog with a $240 million volume purchase agreement for Dragonfly 2D inspection and 3D bump metrology systems through 2027.

Global AI investment fueling a robust upcycle in semiconductor capital equipment spending. Mike Plisinski, CEO

I see two waves of demand here. The first is the HBM3 to HBM4 scaling cycle, which is already underway. The second is co-packaged optics, where optical components get integrated directly onto the chip package, pulling Onto's metrology equipment into an entirely new application. Onto already generates 61% of its revenue from AI-related applications, and that share is climbing.

Shares were down about 15% on geopolitical and recession concerns when I first recommended Onto. That entry has only looked better with every earnings report since.

ACM Research is a diversifier in this portfolio, and it benefits from something that actually hurts most other semiconductor equipment companies: the Chinese buildout.

Here's the dynamic. Western memory makers, Micron, SK Hynix, and Samsung, are staying disciplined on output. China's CXMT wants to flood the market with its own memory. To do that, CXMT needs equipment. And ACM Research, which has built deep relationships inside the Chinese semiconductor industry, is one of the primary suppliers of that equipment. ACM also has growing traction in Korea with SK Hynix.

Revenue has grown from $157 million in 2020 to $901 million in its most recent fiscal year. The company's long-term target is $4 billion. Forward revenue guidance implies 21% to 30% growth year over year.

We have the customers, the products, the capacity and the capital to execute on our global business plan. David Wang, CEO

The stock sold off hard on short-term concerns, and that's exactly the entry point I was looking for. Analyst consensus price targets still imply meaningful upside from where shares trade.

I'll be straight with you: this one is small and speculative. But the story is compelling enough that I wanted it in this portfolio.

Aehr makes burn-in and test equipment for semiconductors. For the past few years, its primary market was electric vehicle power chips, and that market has been brutal. Revenue has been falling. The company did $66 million in revenue in 2024, $59 million in 2025, and is expected to do around $47 million in 2026. On the surface, that looks like a broken story.

But here's what I'm watching. Aehr is actively pivoting into silicon photonics for data center optical interconnects, memory testing, and custom ASICs for AI accelerators. A recent quarter showed bookings of $37.2 million with a book-to-bill ratio exceeding 3.5×. That's an extraordinary bookings number relative to current revenue.

More than $37 million in quarterly bookings and a book-to-bill ratio exceeding 3.5×. Demand continues to accelerate. Gayn Erickson, CEO

The company also secured a production win for its Sonoma PLBI systems for next-generation AI accelerators and added a new silicon photonics customer for data center optical interconnects. Management expects a significant revenue rebound in 2027. The pivot is the story here.

Aixtron is the most obscure name on this list. It primarily trades on Germany's stock exchange, with a thinly-traded OTC listing in the U.S. under AIXXF. Volume averages roughly 8,000 shares per day, so use limit orders if you're buying. Don't market-buy this one. I wanted it in this portfolio because very few investors own it yet, and I think that changes over the next 24 months.

The company in one sentence: Aixtron builds vapor deposition equipment for non-silicon semiconductor materials, specifically gallium nitride (GaN), silicon carbide (SiC), and indium phosphide (InP). Those three compounds are the exact materials that matter for two different AI waves happening at the same time.

First, indium phosphide is the starting point for optical transceivers. Silicon is a poor material for generating light, so every optical module in a data center eventually traces back to an InP wafer. As optics scale toward co-packaged architectures and 1.6T and 3.2T networking, demand for InP capacity explodes. AXT Inc., which makes the wafers at the front of that supply chain, has seen its stock appreciate roughly 2,000% in a year. That tells you how acute the bottleneck already is.

Second, gallium nitride and silicon carbide are the workhorse materials for 800V power management, which is the architecture Nvidia's Vera Rubin platform requires. Every data center rack that moves from 54V to 800V becomes a demand pull for Aixtron equipment.

The catch, and the reason the stock is still depressed, is that Aixtron's legacy business served the electric vehicle market, which has been brutal for years. Earnings have been cut roughly in half since 2022. That's the entire bear case. I believe the next 24 months flip that entirely as the optics bottleneck and the 800V power transition both land on Aixtron's order book at the same time.

If you're buying, be patient with those limit orders. Don't chase.