Nvidia vs. Taiwan Semiconductor: The Battle for AI Chip Dominance

Prediction: 3 Artificial Intelligence (AI) Stocks That Will Be Worth More Than Nvidia in 3 Years | The Motley Fool

These three tech giants have sustainable competitive advantages and could leave Nvidia in the dust. Nvidia (NVDA -2.25%) has been one of the biggest beneficiaries of the booming demand for anything related to artificial intelligence (AI). The company's GPUs are valuable equipment used to train large language models, which provide the backbone for generative AI. Big tech companies and cloud providers are snatching up as many Nvidia chips as they can get their hands on. The results have been nothing short of phenomenal for Nvidia. Over the last two years, Nvidia has seen its market cap climb from about $424 billion to $3.1 trillion. It's currently sitting right above Microsoft as the second-largest company in the world behind only Apple. While Nvidia gets a lot of attention, it's important to remember it's not the only AI stock you can invest in. There are dozens of other great companies participating in the growth of AI. And the long-term prospects for some of them are even better than Nvidia's. That's why I predict these three companies could end up passing Nvidia's value over the next three years. Meta Platforms (META -1.30%) is one of Nvidia's largest customers. CEO Mark Zuckerberg recently committed to accumulating 350,000 of the chipmaker's H100 GPUs by the end of this year. The company's capital expenditures, which management expects to come in between $37 billion and $40 billion this year, are rivaled only by the big public cloud companies like Microsoft and Alphabet (GOOG 0.30%) (GOOGL 0.33%). And management expects Meta capex to keep climbing in 2025. Those massive investments won't pay off directly for some time, but Meta is positioned better than just about any other company to integrate AI capabilities into its products. It sees opportunities to improve the advertising business, grow its business messaging service (with custom AI chatbots), build the most popular consumer-facing AI assistant, and increase engagement across Facebook, Instagram, and its messaging apps. That should lead to strong revenue and earnings growth over the long run, even with a step up in depreciation expenses from the increased investments in data centers. Importantly, the growth in spending will slow down over time, as Meta determines exactly how much data center capacity it needs for training and using its generative AI developments. The stock currently trades at a forward PE around 26, which is reasonable for a company with its growth prospects. But Meta could outperform earnings estimates if its generative AI features increase ad prices and engagement on Facebook and Instagram and open the door for new revenue opportunities through business messaging and AI assistant interactions. With a market cap of about $1.35 trillion, it could find itself quickly growing to surpass Nvidia over the next three years if one of its AI efforts starts showing very good results. Taiwan Semiconductor Manufacturing (TSM -1.29%), also known as TSMC, is the largest chip manufacturer in the world. When a company like Nvidia designs a new chip, it contracts with TSMC to produce that chip. TSMC's biggest advantage is its scale. It takes in over 60% of all spending on chip manufacturing. That gives it more money to reinvest in building its capabilities to produce faster, more powerful, and more power-efficient chips. With its cutting-edge design capabilities, it's able to maintain a dominant market share, since chip designers can't get the same results from anyone else. TSMC's relatively agnostic to who designs the chips. It can make an Nvidia chip just as easily as it can make a custom-designed chip from Meta, Microsoft, or Alphabet. And it does exactly that. That leaves it in a much less precarious position than Nvidia when it comes to the future of AI data center chips. Many of Nvidia's largest customers have their own chip designs specifically for training and using large language models. These chips aren't as flexible in their usage as Nvidia's, but they're more power-efficient and less expensive to acquire. That makes them increasingly valuable as companies like Meta, Alphabet, Microsoft, and others dial in and scale their AI development. In the meantime, TSMC is able to benefit from the growing spend and increased competition for its limited resources. Its shares trade around 26 times forward earnings, but they arguably deserve a higher multiple as high demand and competition should benefit its bottom line and its earnings are much better protected on the downside. By comparison, Nvidia trades for a forward earnings multiple of 48. The true value for both companies is likely a multiple somewhere in the middle. If TSMC sees multiple expansions and Nvidia sees contraction, TSMC could surpass Nvidia's market cap over time. Alphabet is already the fourth-largest company in the world, but its market cap sits about $1 trillion behind Nvidia's. Still, there are good reasons to think the Google owner will increase in value faster than the chipmaker going forward. At the core of Alphabet is Google Search. And while it's facing regulatory pressure, it's unlikely to lose the crown as the most popular search engine in the world. That ensures Google remains a key part of marketers' advertising budgets. Importantly, artificial intelligence assistants like OpenAI's ChatGPT or Meta AI aren't as big of threats as investors once feared. Google has used its own AI, based on its Gemini LLM, to provide direct answers to many search queries, linking to relevant sources. Management says its AI Overviews result in greater search usage and higher user satisfaction. It's also using AI to support new ways to search, including taking videos with your phone or circling content on an app or web page. Both should lead to increased usage. But AI has the potential to significantly grow Google's cloud platform, which surpassed $10 billion in quarterly revenue for the first time last quarter. Google is winning many high-profile customers, including Apple, for its AI development platform on Google Cloud. Meanwhile, its Gemini for Workspace software is helping increase its average revenue per user and attract new customers. Alphabet is certainly spending heavily in order to win those customers. It expects to spend about $50 billion on capital expenditures this year, mainly investing in servers and data centers. But the payoff could be huge, as it invests ahead of the curve of demand for AI development. Google Cloud can become a much bigger part of the business, providing a nice complement to the Search business. With shares trading at just 22 times forward earnings, there's room for multiple expansions. That's especially true if you consider the company's bottom line should sustain very high growth due to significant share buybacks supported by Alphabet's massive annual free cash flow. It's actually surprising investors don't value Alphabet higher than Nvidia already.

Forget Nvidia -- Taiwan Semiconductor Is the Best Way to Invest in Artificial Intelligence (AI)

Nvidia (NASDAQ: NVDA) has captured the heart of the artificial intelligence (AI) investment trend by providing investors with jaw-dropping growth rates and incredible dominance within the graphics processing unit (GPU) industry. However, challenges are starting to arise within its business (even though it's still dominant), with competitors building up their AI offerings and customers starting to design their own AI chips to train software models. All of this leads me to believe there's another company that's a better AI investment, Taiwan Semiconductor Manufacturing (TSM -1.29%). Regardless of which AI company you're talking about, it's likely a TSMC customer because TSMC is the world's largest contract chip manufacturer. This means it acts as a fabrication shop for all things chip-related. With massive demand across the board, it's set to benefit whether Nvidia continues its dominance or others knock it from its perch. TSMC's new product will be an important innovation TSMC has sat atop the semiconductor throne for a while. It has competitors, but none offer the scale and technology that it does. It can produce the world's smallest and most powerful chip, which has a distance between traces of 3 nanometers (nm). It also has a 2nm chip in development that will significantly improve on the energy efficiency of the 3nm chip, which is fantastic news for companies building out their AI computing power. The GPUs in these data centers are energy-hungry and cost a lot to run. If customers can outfit them with products that have similar performance but improve energy consumption by 25% to 30%, they will be a massive hit. Pre-production demand has already surpassed that of previous-generation chips (3nm and 5nm), so this is a huge catalyst that will be realized once TSMC's 2nm chips enter production in 2025. This demand adds to management's projection that AI demand will grow at a 50% compounded annual growth rate (CAGR) through 2028 and will account for more than 20% of its revenue by then. Over the long term, management sees its revenue growing by a CAGR of 15% to 20%, which is a fantastic rate by all measures. But that pales in comparison to what Nvidia is seeing, so why would investors want to pick TSMC over Nvidia? It all has to do with steadiness. TSMC is a more consistent company Both Nvidia and TSMC are cyclical businesses, but Nvidia's volatility during a cycle is far greater than that of TSMC. Take a look at the chart below, which graphs the percentage of each company's revenue and earnings per share (EPS) that fell from all-time highs. TSM Revenue (Quarterly) data by YCharts As you can see, Nvidia's dips are far deeper than TSMC's, which shows its decline at the bottom of a cycle is far greater. This likely has to do with TSMC's product breadth, as its chips go into far more products than just GPUs. Nvidia is a one-trick pony, even if that trick is amazing. While some investors are OK with Nvidia's booms and busts, I'm not. Recognizing your appetite for risk is a huge part of investing, and if you're a bit concerned about how Nvidia will perform due to its massive run-up already, then TSMC is a great substitute to capitalize on the same tailwinds that are propelling Nvidia higher.