The intersection of artificial intelligence (AI) and semiconductor technology is producing some of the most dynamic shifts in the tech industry today.

On that note, three Motley Fool contributors who cover tech and AI stocks put their heads together to highlight their top stock picks among AI-focused semiconductor stocks today. They came up with memory technology frontrunner Micron Technology (MU -0.33%), industry legend Intel (INTC -1.19%), and AI chip design leader Broadcom (AVGO -0.59%).

Read on to review how these chip titans benefit from the ongoing AI surge.

Intel aims to catch up to Nvidia and TSMC all at once

Billy Duberstein (Intel): Intel has badly lagged other AI companies in recent years and in 2024, with the stock down over 25% on the year thus far as the Nasdaq has rallied almost 8%.

That's somewhat warranted -- after all, Intel is currently in a multiyear attempt to catch up to Taiwan Semiconductor Manufacturing in terms of its ability to produce leading-edge chips. Success would not only mean Intel's designed chips would achieve market leadership once again, but that Intel would also become a successful and profitable foundry for third parties.

Achieving process technology leadership and winning new clients is a long, tough slog, but Intel has shown good progress over the past year or so, hitting milestone after milestone on its forecast timeline to achieve five chip nodes in four years.

And just last week, Intel unveiled its new Gaudi 3 chips, the accelerators Intel hopes will compete with Nvidia GPUs. In a surprise, Intel's management claimed the new Gaudi chips outdo the current Nvidia H100 on several key metrics. While Nvidia is already moving on to its new Blackwell chips later this year, Gaudi has the potential to find some small share of the AI market, which is turning out to be a very big market.

Aside from sheer hardware chops, Intel is also leading large consortiums aiming to form open source solutions to both AI programming software, taking aim at Nvidia's CUDA software moat, as well as AI ethernet networking, challenging Nvidia's Infiniband solution.

Intel also just broke out its products and foundry segment financials separately, giving investors a sense of what Intel may be generating in profits without the burden of all its foundry capital. Outside of Intel's foundry, its chip products made about $12.4 billion in operating profit last year. That's not too bad, considering Intel has fallen behind on process technology in recent years, and that this doesn't contain much of any Gaudi revenue. It also means Intel's valuation is not demanding, with a market cap of just $160 billion.

Of course, Intel's foundry lost a whopping $7 billion last year alone. But that's not entirely surprising, given the huge capital outlays and depreciation that must be taken ahead of Intel bringing in revenue.

Will Intel pull it off? The company is investing heavily behind the effort, procuring the first high-NA EUV machine from ASML Holdings. Not only that, but Intel has also locked down the majority of ASML's high-NA EUV machines for this year -- an aggressive investment to try and pull ahead of TSMC in 2025.

For those looking for AI winners that haven't yet taken off, Intel could be a contrarian play. However, it's also riskier, and much more dependent on solid execution. Still, success over the next couple of years could lead to meaningful upside for the stock -- probably more than the others that have already run.

Micron is a great AI bet, no matter who wins the accelerator wars

Anders Bylund (Micron Technology): Every chip designer worth its sodium silicate is coming up with powerful AI accelerator chips these days. If you can't beat the raw performance of Nvidia's H100 and Blackwell chips, maybe you can offer a more power-efficient solution, or a cheaper chip, and so on. There's enough demand for AI systems and solutions to create all sorts of exploitable niche markets.

Micron Technology takes a different approach to the AI opportunity.

As a leading provider of memory hardware, Micron benefits from the AI-driven boom in high-performance computing. Every high-end number-crunching machine needs a ton of fast DRAM memory, and most also rely on large arrays of solid-state storage devices (SSDs) built around NAND memory chips.

The memory sector as a whole is on the rise, but Micron goes a couple of steps further.

  • Right now, chief rivals Samsung and SK Hynix don't have an answer to the power efficiency of company's latest high-bandwidth memory (HBM) products. This should make Micron the best choice for situations where a very large amount of high-speed DRAM memory is needed -- such as the supercomputers that train AI engines with the help of accelerators from Nvidia and its rivals.
  • In fact, Nvidia's Blackwell AI accelerators are shipping with up to 192 gigabytes of Micron HBM per module. In other words, Micron rides Nvidia's coattails in an incredibly literal way.

Now, Micron investors have seen these growth catalysts coming in and the stock trades just a hair below all-time highs. Micron's valuation ratios also run high. If you're looking for a low-priced discount deal, this isn't the stock for you.

But it's a different story if you believe that the memory market is changing on a deeply fundamental level, placing Micron in a golden age of technology leadership amid a game-changing AI boom. In this light, Micron's high valuation might just be "the new normal" with the potential of further gain over the next couple of years.

So you should review your tolerance for market risk and act accordingly. For growth-oriented investors with a robust risk appetite, Micron offers a direct investment in the AI frenzy without having to pick a winner in the accelerator chip wars.

Big tech AI chips are really all about this top semiconductor stock

Nicholas Rossolillo (Broadcom): It's been a busy week in AI and semiconductors, with Alphabet announcing a new Arm Holdings-based processor. Google also discussed how its latest Tensor Processing Units (TPUs) are training Google AI. Meta Platforms announced its latest custom AI chip for processing some of its data center computing. And even Intel tried to get in on the action with the debut of its latest Gaudi 3 AI chip.

At this point, all of the tech giants that make up the "Magnificent Seven" have, at the very least, hinted that they're working on something in the name of competing with Nvidia.

In all reality, these companies aren't so much trying to compete with Nvidia as much as they are looking for a way to save some money.

At any rate, what few investors realize is that most of these big tech chip designs aren't exactly of their own making. They actually have to tap the semiconductor industry itself in order to design these AI systems, which they are able to do so since they have such ample financial resources at their disposal.

The largest custom AI processor design house at this point is none other than chip titan Broadcom, which is responsible for helping Google with its custom TPUs, as well as doing the heavy lifting design work for Meta, and an as-yet unnamed third tech giant on another AI project.

As a result, some 35% of Broadcom's 2024 semiconductor sales are expected to be from AI.

These days, the market is focused on how well Broadcom can integrate and unlock profitability from its mega-acquisition of cloud software essential VMware. But the AI chips could be a real long-term game changer. At 28 times the current fiscal year expected earnings per share, Broadcom stock could actually still be a decent value for the long haul.