Uncover the key players shaping the future of technology with breakthrough AI developments.
The intersection of artificial intelligence (AI) and semiconductor technology is creating some of the most dynamic change in the technology industry today.
On that note, three Motley Fool contributors covering tech and AI stocks have pooled their brains to highlight today's top AI-focused semiconductor stocks.They came up with a pioneer in memory technology micron technology (MU -3.91%)an industry legend intel (INTC -5.16%)a leader in AI chip design broadcom (AVGO -2.78%).
See how these chip giants are benefiting from the ongoing AI explosion.
Intel aims to catch up with Nvidia and TSMC at once
billy duberstein (Intel): Intel has lagged far behind other AI companies in recent years and into 2024, with its stock price down more than 25% year-over-year so far. Nasdaq It rose nearly 8%.
That makes sense to some extent. After all, Intel is currently in a multi-year effort to catch up. taiwan semiconductor manufacturing In terms of their ability to manufacture cutting edge chips. Success not only means that Intel-designed chips will once again gain market leadership, but it also means that Intel becomes a successful and profitable foundry for third parties.
Achieving leadership in process technology and attracting new customers will be a long and tough challenge, but Intel has shown good progress over the past year or so, reaching milestones with a projected schedule of five chip nodes in four years. have been achieved one after another.
And just last week, Intel announced its new Gaudi 3 chip, an accelerator that Intel hopes to compete with. Nvidia GPU. Surprisingly, Intel's management claimed that the new Gaudi chip outperformed his current Nvidia H100 in several key metrics. Nvidia has already started work on new Blackwell chips later this year, but Gaudi could capture a small share of what is becoming a very large market for AI.
In addition to significant hardware reductions, Intel is also building a large consortium that aims to form open source solutions for both AI programming software to take aim at Nvidia's CUDA software moat and AI Ethernet networking to compete with Nvidia's Infiniband solution. I am leading.
Intel also just released separate financials for its product and foundry divisions, allowing investors to see what kind of profits Intel is generating without committing to all of its foundry capital. Outside of Intel's foundries, the company's chip products generated operating profits of about $12.4 billion last year. This isn't such a bad thing, considering Intel has lagged behind in process technology in recent years, and that includes little of Gaudi's revenue. It also means that Intel's valuation, with a market capitalization of just $160 billion, isn't harsh.
Intel foundry of course lost A whopping $7 billion last year alone. However, this is not at all surprising considering that Intel has to make huge capital expenditures and depreciation before it can bring in any revenue.
Will Intel be successful? The company is investing heavily in this effort, procuring its first high NA EUV machines from: ASML Holdings. Not only that, but Intel has locked down most of ASML's high NA EUV machines this year. This is an aggressive investment that will try to get ahead of his TSMC in 2025.
For those looking for an AI winner that hasn't yet caught on, Intel could be a contrarian play. However, it is also high risk and highly dependent on reliable execution. Still, success over the next few years could lead to big gains in the stock price, perhaps more so than other stocks that are already up.
No matter who wins the accelerator wars, Micron is a big bet on AI
Anders Byland (Micron Technology): These days, every chip designer worth sodium silicate is coming up with powerful AI accelerator chips. If you can't beat the raw performance of Nvidia's H100 and Blackwell chips, they may be able to offer more power-efficient solutions, cheaper chips, and more. There is enough demand for AI systems and solutions to create all sorts of exploitable niche markets.
Micron Technology is taking a different approach to the AI opportunity.
Micron, a leading provider of memory hardware, is benefiting from the AI-driven boom in high-performance computing. All high-end math machines require large amounts of high-speed DRAM memory, most of which also rely on large arrays of solid-state storage devices (SSDs) built around his NAND memory chips. .
The memory space as a whole is on the rise, but Micron is going a few steps further.
- Currently, its main rival Samsung SK Hynix There is no answer regarding the power efficiency of the company's latest high-bandwidth memory (HBM) products. This should make Micron the best choice in situations where very large amounts of fast DRAM memory are needed, such as supercomputers training AI engines using accelerators from NVIDIA and its rivals.
- In fact, Nvidia's Blackwell AI accelerators ship with up to 192 gigabytes of Micron HBM per module. In other words, Micron is riding the coattails of Nvidia in an incredibly literal way.
Now, Micron investors are seeing these growth catalysts arrive, with the stock trading just below its all-time high. Micron's valuation ratio is also high. If you're looking for a discount deal at a low price, this isn't the stock for you.
But it's a different story if you think the memory market is changing at a fundamental level and Micron is entering a golden age of technology leadership amidst a revolutionary AI boom. From this perspective, Micron's high valuation may just be the new normal, with the potential for it to rise further in the coming years.
Therefore, you should review your tolerance for market risk and act accordingly. For growth-minded investors with strong risk appetite, Micron offers direct investment in the AI frenzy without picking winners in the accelerator chip wars.
Big Tech's AI Chips Are All About This Top Semiconductor Stock
Nicholas Rossolillo (Broadcom): This week has been a busy week for AI and semiconductors. alphabet announce something new arm holdingsbased processor. Google also explained how its latest Tensor Processing Units (TPUs) are training Google AI. meta platform We announced our latest custom AI chip to handle some of the computing in our data centers. 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 are at least hinting that they're working on: something In the name of competing with Nvidia.
In reality, these companies aren't so much trying to compete with Nvidia as they are looking for ways to save money.
In any case, what most investors don't realize is that most of these big tech chip designs aren't of their own making at all. Designing these AI systems actually requires leveraging the semiconductor industry itself, but it's possible because there's plenty of money at our disposal.
The largest custom AI processor design company today is none other than chip giant Broadcom. Broadcom is behind Google's custom TPUs and is responsible for the heavy-lifting design work for Meta, and is a third, as-yet-unnamed tech giant. In another AI project.
As a result, roughly 35% of Broadcom's semiconductor sales in 2024 are expected to come from AI.
Lately, the market has been watching how Broadcom can unlock profitability from consolidating its big acquisition of VMware, which is critical to cloud software. But AI chips could be a game-changer in the long run. At 28 times this fiscal year's estimated earnings per share, Broadcom stock may actually still be worth a decent amount over the long term.
Suzanne Frey, an Alphabet executive, is a member of the Motley Fool's board of directors. Randi Zuckerberg is a former head of market development and spokesperson at Facebook, sister of Meta Platforms CEO Mark Zuckerberg, and a member of the Motley Fool's board of directors. Anders Bylund has held positions at Alphabet, Intel, Micron Technology, and his Nvidia. Billy Duberstein has worked at ASML, Alphabet, Broadcom, Meta Platforms, Micron Technology, and Taiwan Semiconductor Manufacturing. Nicholas Rossolillo has held positions at ASML, Alphabet, Broadcom, Meta Platforms, Micron Technology, and his Nvidia. The Motley Fool has positions in and recommends ASML, Alphabet, Meta Platforms, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool recommends Broadcom and Intel and recommends the following options: A long January 2023 $57.50 call on Intel, a long January 2025 $45 call on Intel, and a short May 2024 $47 call on Intel. The Motley Fool has a disclosure policy.