Why Broadcom and TSMC look poised to be AI chip winners.
The hottest artificial intelligence (AI) stock no doubt has been Nvidia, whose graphic processing units (GPUs) have become the backbone of the AI infrastructure buildout. However, with spending in the space still frantic as companies rush to compete, Nvidia won’t be the only winner.
While Nvidia has been a great stock, I want to highlight two more companies that have a lot of potential upside from the AI boon.
Broadcom
While Nvidia is the leader in mass market GPUs, Broadcom (AVGO -3.03%) is looking to be the leader in customized AI chips. These are application-specific integrated circuits, or ASICs, that are designed specifically for a customer’s precise needs. These chips are created to run specific AI workloads, and as such are more efficient than general GPUs.
Broadcom’s first customer in the space was Alphabet, for which it helped design and provide key technologies for its tensor-processing units (TPUs). These TPUs are now used for both AI training and inference. Alphabet claims that what separates its TPUs from GPUs is that they carry several specialized features, including a matrix multiply unit (MXU) and proprietary interconnect topology that make them ideal for accelerating AI training and inference.
The company has announced that it has added more large AI chip customers this year, which are believed to be Meta Platforms and China’s ByteDance, owner of TikTok. ByteDance needs AI chips that are compliant with U.S. export restrictions as Nvidia’s most advanced GPUs are not permitted to be shipped to China. Meta, meanwhile, values its strategic relationship with Broadcom so much so that the company added its CEO to its board of directors earlier this year.
There are also multiple reports that OpenAI is set to become Broadcom’s fourth big AI chip customer as it has been in discussions with the company to design its own chips. There have also been rumors that the company has secured a fifth AI ASIC customer with a planned ramp up in 2026.
Custom AI chips are a large and growing market, which has helped Broadcom continue to ramp up its AI revenue forecasts, which started the year at $7.5 billion but are now sitting at $12 billion. With the addition of new customers, this is a huge opportunity for the company over the next several years.
Image source: Getty Images.
Broadcom is also benefiting from providing networking components used in the buildout of AI data-center infrastructure. The company believes its ethernet switches will be at the heart of handling AI workloads and transferring data between GPUs, and that as GPU clusters get larger, the importance of switching becomes more important.
AVGO PE Ratio (Forward 1y) data by YCharts
Broadcom has a huge AI opportunity in front of it, while the stock is attractively priced trading at a forward price-to-earnings (P/E) ratio of under 28 based on next year’s analyst estimates and a price/earnings-to-growth (PEG) ratio of 0.75. Generally, a stock with a PEG under 1 is considered undervalued, and growth stocks often command PEG ratios much higher than 1.
Given its valuation and growth prospects, Broadcom’s stock looks like it could be the next big AI winner.
Taiwan Semiconductor
Taiwan Semiconductor Manufacturing (TSM -4.74%), or TSMC for short, is the largest semiconductor contract manufacturer in the world, and as such it plays a vital role in the chipmaking process. It manufactures chips for most of the top semiconductor companies, including Nvidia and Broadcom, while Apple, which designs its own chips for use in its devices, is its largest customer.
As such, the the company is benefiting from the proliferation of chips needed for both the AI infrastructure buildout and any hardware upgrade cycle that may accompany it for smartphones and other devices. By and large, it doesn’t matter who is winning the AI chip race, as most of the big players are its customers.
Given the insatiable need for chips, not surprisingly TSMC’s services are in high demand, and the company is working to increase capacity to meet this increasing demand. Meanwhile, thus far there is no indication that demand is waning as major tech companies continue to ramp up their AI infrastructure investments and more and more compute power is needed as large language models (LLMs) become more advanced.
With demand so high for AI chips, capacity at TSMC’s foundries (chip manufacturing facilities) is tight, which is also giving the company a lot of pricing power. According to Morgan Stanley, the company is set to raise prices by 10% for AI semiconductors and chip-on-wafer-on-substrate products, by 6% for high-performance computing, and by 3% for smartphones.
TSM PE Ratio (Forward 1y) data by YCharts
The combination of increasing capacity to meet demand and strong pricing power sets up TSMC to be one of the biggest AI beneficiaries over the next several years. Meanwhile, the stock is relatively inexpensive trading at a forward P/E ratio of about 21 based on next year’s analyst estimates and a PEG ratio of just over 1. Given its robust outlook, the stock looks like a solid buy at current levels.
Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool’s board of directors. Randi Zuckerberg, a former director of market development and spokeswoman for Facebook and sister to Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Fool’s board of directors. Geoffrey Seiler has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet, Apple, Meta Platforms, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy.