TSMC Emerges as a Key Player in the AI Chip Manufacturing Landscape

Nvidia's Huang says buying TSMC's stock is 'very smart.' The pros share their take

Taiwanese chipmaker Taiwan Semiconductor Manufacturing Co made headlines last week following Nvidia CEO Jensen Huang's remarks that buying its stock would be "very smart. " "Well, first of all, I think TSMC is one of the greatest companies in the history of humanity, and anybody who wants to buy TSMC stock is a very smart person," he said on Friday during his Taiwan visit. Shares in TSMC rose on Monday and Tuesday and are up 9.77% since the start of the year. Last month, TSMC's market capitalization crossed $1 trillion, thanks to a strong sales forecast driven by growing artificial intelligence demand. Its market value now stands at 30.42 trillion New Taiwan dollars, which is just under $1 trillion, according to LSEG data. 2330-TW 1Y mountain TSMC shares TSMC manufactures advanced chips, including for major chip designers such as Nvidia and AMD . The Taiwanese firm has now taped out a new graphics processing unit and silicon photonics processor for Nvidia's next-generation Rubin-architecture supercomputers, meaning chip design has been finalized and manufacturing can begin. The firm has cut the use of Chinese chipmaking equipment in its most advanced chip plants to avoid potential U.S. curbs that could impede production, Nikkei reported, citing comments from sources familiar with the matter. FactSet data shows that most analysts are bullish on TSMC. Of the 44 analysts covering the stock, 42 give it a buy or overweight rating, while one has a hold and another has a sell rating. Analysts' average price target is NT$1,354.27, giving it 15.7% potential upside. The spotlight on TSMC has now raised questions about whether those not already invested should buy the stock. A neutral view Tariq Dennison, co-founder and investment advisor at GFM Asset Management, is neutral on TSMC. It's "notable when an industry leader says something so positive about another company in the same industry," he told CNBC Pro . However, Dennison said, he tends to "get cautious, or even bearish," when too many people are bullish on a stock or sector. Dennison attributed Huang's view to TSMC's "very clear lead" in the manufacturing of the most advanced semiconductors, including 2 to 3 nanometer technology, and the likelihood that it will "keep that lead for many years to come. He also inferred that Huang expects TSMC to "keep growing and monetizing that lead for many years to come." That's a "much more optimistic assumption than I would make," the wealth manager said. He said he believes TSMC is of high quality and its growth prospects are "well priced in a 23 times forward earnings, versus the 4.25% yield on the 10-year U.S. Treasuries." However, he cautioned that its valuation underprices the risks of geopolitical uncertainties and unforeseen technological disruptions. The bulls However, Arthur Lai, head of technology research Asia at Macquarie Capital, is bullish on TSMC. Huang's "comments revalidate our raised CoWoS [chip-on-wafer-on-substrate] forecasts and reinforce that Nvidia's execution is a multi-year demand driver for AI packaging and interconnects," he said. CoWoS is an advanced packaging technology developed by TSMC. "We see continued upside for the advanced packaging and networking supply chain," Lai added in an Aug. 26 note. Lai had raised his target price on the stock by 2% to NT$1,310 following the company's second-quarter results in July. He simultaneously increased his 2025 and 2026 net profit forecast for TSMC by 11% and 5%, respectively, given better-than-expected AI growth and gross profit margin impact from foreign exchange charges. That was thanks to TSMC's solid near-term AI demand, even as the company anticipates softness in its revenue in the last quarter of the year in light of "tariff-related uncertainties and a still cautious consumer segment," Lai wrote in a July note. Phelix Lee, equity analyst at Morningstar, is also bullish on TSMC. He has a five-star rating on TSMC and considers it "attractively valued" given its quasi-monopoly in manufacturing AI and other premium semiconductor chips, he told CNBC Pro. Morningstar gives stocks a rating of between one and five stars, with a top rating indicating that its shares are undervalued. In July, Lee hiked his fair value estimates for TSMC to NT$1,800 from NT$1,700, after the company raised its full-year revenue growth guidance by about 30% in U.S. dollar terms. He also raised his 2025-2029 revenue estimates by 5% and earnings per share estimates by 9% in anticipation of AI's contributions. "TSMC is undervalued as the market is overestimating tariff effects and underestimating the longevity of AI investments," he wrote in a July note. -- CNBC's Dylan Butts contributed to this report.

Billionaire Stanley Druckenmiller Sold His Fund's Stakes in Nvidia and Palantir, and Has Piled Into This Essential Artificial Intelligence (AI) Stock Instead | The Motley Fool

Duquesne's billionaire boss completely jettisoned Wall Street's two hottest AI stocks in favor of a trillion-dollar artificial intelligence business that's notably cheaper. The month of August has been chock-full of important data releases. The inflation and jobs report from the federal government, Federal Reserve Chair Jerome Powell's highly anticipated speech at Jackson Hole, and earnings releases from many of Wall Street's most influential businesses have deluged investors with information. It's also made it easy for something of importance to slip through the cracks. For instance, Aug. 14 marked the deadline for institutional investors with at least $100 million in assets under management (AUM) to file Form 13F with U.S. regulators. A 13F offers a concise snapshot of which stocks Wall Street's shrewdest money managers bought and sold in the latest quarter (in this case, the quarter ended June 30). Although billionaire Warren Buffett has earned quite the following on Wall Street, he's not the only billionaire fund manager with a keen eye for value. Duquesne Family Office's billionaire boss Stanley Druckenmiller is also known for his outsize returns and ability to snag amazing deals. What's particularly interesting about Druckenmiller has been his trading activity within the artificial intelligence (AI) space. Over the previous year, based on 13F filings covering trading activity from July 1, 2024 to June 30, 2025, Duquesne's investment guru sent Wall Street's two top-tier AI stocks, Nvidia (NVDA 1.10%) and Palantir Technologies (PLTR -0.98%), to the chopping block. But at the same time, he purchased shares of an essential AI titan for four consecutive quarters and made it a top-five holding. As of the midpoint of 2024, Duquesne's billionaire chief was overseeing $2.9 billion in AUM spread across 64 holdings. This included 214,060 shares of Nvidia, following its 10-for-1 forward split in June 2024, as well as 769,965 shares of Palantir. By March 31, 2025, both of these holdings were completely exited. The logical explanation for selling two of the hottest AI stocks on the planet is that Druckenmiller was locking in gains. The average stock for Duquesne Family Office has been held for less than seven months, which suggests its billionaire investor isn't afraid to cash in his fund's chips when the opportunity presents itself. Shares of Nvidia have catapulted higher by roughly 1,120% since the end of 2022, as of the closing bell on Aug. 22. Its AI-graphics processing units (GPUs) have become the preferred choice by businesses to deploy in their AI-accelerated data centers. Strong ongoing demand for AI-GPUs, coupled with Nvidia's well-defined compute advantages, have allowed it to charge a premium for its hardware. Meanwhile, Palantir's stock has gained well over 2,300% during the same time frame. Its AI and machine learning-fueled Gotham platform, which assists federal governments with military mission planning and oversight, has done a lot of the heavy lifting. With no clear one-for-one replacement for Gotham, and no true competition to its enterprise subscription Foundry platform at scale, Palantir has commanded quite the valuation premium. What might be of concern to Nvidia and Palantir shareholders is if Druckenmiller sent these stocks packing for reasons that go beyond profit-taking. As an example, Druckenmiller opined in May 2024 that "AI might be a little overhyped now, but under-hyped long term." Duquesne's billionaire boss astutely recognizes that every next-big-thing trend over the last 30 years has endured an eventual bubble-bursting event early in its expansion phase. In other words, investors have a propensity to overshoot when it comes to early utility and adoption rate expectations. This eventually leads to lofty expectations that aren't met and the bursting of the bubble. No company has benefited more from the AI revolution than Nvidia. If a bubble were to form and burst, there's little question its shares would feel the impact. Though Palantir's multiyear government contracts and subscription revenue would protect it from an immediate revenue drop-off, poor investor sentiment during a bubble-bursting event would still, likely, weigh on its shares. Valuation is another catalyst that may have enticed Druckenmiller to sell. Nvidia and Palantir are trading at respective price-to-sales (P/S) ratios of 30 and 117. Historically, businesses on the front line of a game-changing innovation have peaked at P/S ratios of 30 to 40. In short, Nvidia and Palantir are priced for perfection. Although Druckenmiller remains skeptical of some of the companies leading the evolution of AI over the short run, there is one essential artificial intelligence stock that he simply can't stop buying for his fund. Duquesne's 13Fs show buying activity spanning four consecutive quarters for trillion-dollar chip fabricator Taiwan Semiconductor Manufacturing (TSM 1.09%), which is best-known by its abbreviation, "TSMC." Here's a breakdown of Druckenmiller's buying activity of TSMC over the previous year: In just one year, TSMC has become a top-five holding for Duquesne's now greater-than-$4 billion investment portfolio. The obvious allure of Taiwan Semiconductor is the role it's playing in AI chip manufacturing. Its chip-on-wafer-on-substrate (CoWoS) technology is essential for the packaging of high-bandwidth memory needed for AI-accelerated data centers. TSMC's monthly CoWoS capacity is expected to effectively double to between 65,000 units to 75,000 units by the end of 2025 from 35,000 per month in 2024. Further expansion is expected to 100,000 monthly units come 2026. This added capacity will help Nvidia, Advanced Micro Devices, and other chipmakers ramp up sales of their AI GPUs. To build on the above, TSMC is dealing with quite the backlog for its services, with some orders stretching well into 2026. Backlogged orders provide a healthy level of operating cash flow predictability. While advanced chips accounted for 60% of TSMC's net sales during the June-ended quarter, there's still plenty of opportunity outside of the AI arena. More than a quarter of its net sales trace back to chips and solutions used in next-generation smartphones. The proliferation of 5G wireless networks has led to a steady smartphone upgrade cycle. Another 10% (combined) of Taiwan Semiconductor's net sales derive from Internet of Things devices and chips used in automobiles. Vehicles and household appliances becoming more reliant on technology is excellent news for TSMC. Duquesne's billionaire chief might also be attracted by TSMC's valuation. Though its forward-year price-to-earnings (P/E) ratio of 21 represents a modest premium compared to recent years, TSMC's annual earnings growth rate of 20% or greater makes its forward P/E quite palatable.

Billionaire Dan Loeb Sold 29% of Third Point's Stake in TSMC in Favor of Another Trillion-Dollar Artificial Intelligence (AI) Stock

Enjoy 0% intro APR on purchases and balance transfers for 15 months -- plus up to 5% cash back in popular categories. Earnings season -- the six-week period where many of the S&P 500's components report their operating results -- is often viewed as the most important data dump every quarter. But don't overlook the invaluable information provided by Form 13F filings with the Securities and Exchange Commission. A 13F is a quarterly filing that shows investors which stocks, exchange-traded funds (ETFs), and select options Wall Street's genius fund managers have been buying and selling. Though 13Fs aren't without their flaws (e.g., they may contain stale information for active hedge funds), they can clue investors into the stocks and trends that are intriguing successful money managers. While Warren Buffett tends to be the most-followed of all billionaire asset managers, he's not the only one with a lengthy track record of spotting good deals hiding in plain sight. Third Point's billionaire investor Dan Loeb is another money manager who rightly garners attention. What's made Loeb such an interesting fund manager to follow is his penchant for investing in large- and small-cap growth stocks, as well as his lure to artificial intelligence (AI) stocks. Based on 13Fs, Loeb has been a decisive seller of world-leading chip fabricator Taiwan Semiconductor Manufacturing (TSM +0.00%), which is more commonly known as "TSMC," over the last year. But he's also done quite a bit of buying in the AI arena, with another trillion-dollar AI stock on the menu. Billionaire Dan Loeb has been sending shares of TSMC to the chopping block Though examining share purchases and sales over the prior three-month period can be informative, looking back six or 12 months can help investors spot patterns. Third Point's last four 13Fs show that billionaire Dan Loeb has been a clear seller of TSMC stock. Between the end of June 2024 and the midpoint of 2025, Loeb has sent 595,000 shares of Taiwan Semiconductor Manufacturing to the chopping block, which reduced his fund's stake in the company by 29%. The 1.43 million shares still held makes TSMC Third Point's fifth-largest holding by market value. One of the likeliest reasons for selling TSMC stock is simple profit-taking. From the start of May 2024 through the midpoint of 2025, Taiwan Semiconductor stock rallied about 65%. Loeb upped his fund's position in TSMC notably during the second quarter of 2024. With the average hold time of Third Point's top-10 positions clocking in at roughly 11 months, it's not surprising to see Loeb actively locking in gains. However, there may be more at play than just a desire to take profits. For example, TSMC is valued at nearly 22 times forward-year earnings. Though a forward price-to-earnings (P/E) ratio of 22 is actually quite reasonable when compared to other market-leading AI stocks, it's a premium compared to the forward P/Es of 13, 16, and 20 that TSMC ended the year with in 2022, 2023, and 2024, respectively. The perceived bargain that Loeb nabbed isn't as apparent in the forward P/E ratio any longer. It's also possible that Loeb has concerns about TSMC's outlays to expand its manufacturing presence in the U.S. Among the reasons the company is expanding production in the U.S. is to avoid tariffs from the Donald Trump administration. The cost to build and expand U.S. chip-manufacturing facilities, coupled with potentially higher production costs and lead times, has the potential to weigh on TSMC's operating results. Nevertheless, Taiwan Semiconductor Manufacturing appears to be ideally positioned to benefit from the AI revolution. Even if there are demand hiccups, its lengthy backlog and revenue diversity - it's a leading provider of chips for next-generation smartphones, automobiles, and internet of Things devices -- set the company up for long-term success. Dan Loeb made this trillion-dollar AI stock a top-three holding over the last two quarters On the other end of the spectrum, Third Point's Dan Loeb entered 2025 with no shares of Wall Street's largest public company and the leader of the artificial intelligence revolution, Nvidia (NVDA 0.01%). But during the first and second quarters, he oversaw the purchase of 1.45 million and 1.35 million shares of Nvidia, respectively. Loeb's fund now owns 2.8 million shares of this trillion-dollar colossus of a company, and it's grown into his fund's third-largest holding. For more than two years, the lure for investors with Nvidia has been its AI-graphics processing unit (GPU) dominance. The company's Hopper (H100) and Blackwell GPUs dominate in AI-accelerated data centers, with Nvidia CEO Jensen Huang noting that "demand is extraordinary" for Blackwell Ultra (the successor to the Blackwell chip). No other competitors have come close to matching the compute capabilities of Nvidia's AI hardware. Huang is attempting to bring a new advanced chip to market on an annual basis. The ramp up of Blackwell Ultra is ongoing, with the debut of Vera Rubin and Vera Rubin Ultra expected in the second-halves of 2026 and 2027, respectively. The latter two will run on the all-new Vera processor and should help solidify Nvidia's position atop the compute pedestal. Loeb likely also appreciates the role Nvidia's CUDA platform has played in delivering eye-popping sales and profit growth. CUDA is the toolkit developers lean on to get the most out of their Nvidia GPUs, including the building and training of large language models. It's effectively become the anchor that keeps clients loyal to Nvidia's ecosystem of products. The other catalyst that may have enticed Loeb to nearly double Third Point's stake in Nvidia during the second quarter was the early April swoon for Wall Street. President Trump's tariff announcements on April 2 led to a short-lived mini-crash in the major stock indexes. This allowed opportunistic investors, including billionaire money managers, to go shopping. But continued upside in Nvidia is far from a guarantee. History tells us that every next-big-thing trend for more than 30 years has endured a bubble-bursting event early in its existence. Investors have a terrible habit of overestimating the early stage utility and adoption rate of new technologies, which eventually leads to sky-high expectations not being met. This same fate seems likely for the artificial intelligence revolution. Competitive pressures should eventually weigh on Nvidia, as well. In particular, having many of its top customers by net sales developing their own AI-GPUs could cost it valuable real estate in data centers, or at the very least may delay future upgrade cycles.

Prediction: This Unstoppable Stock Could Be the Next $2 Trillion Giant | The Motley Fool

Taiwan Semiconductor is posting rapid growth thanks to the ongoing AI arms race. As companies get larger, it will become more common for them to cross the $1 trillion and $2 trillion thresholds. Still, there are only a handful of companies within striking distance, as only 19 companies worldwide have a market cap of $500 billion or greater. One of the most promising companies in this group is Taiwan Semiconductor Manufacturing (TSM -3.11%), which has a market cap of $1.2 trillion. I think TSMC is slated to become a $2 trillion company within a few short years, thanks to its position in an important industry. Although it has delivered impressive shareholder returns over the past few years, I think it's slated for even more as it pursues a $2 trillion valuation. TSMC is a chip fabricator that offers its capabilities to some of the most important tech companies in the world. Companies like Apple (AAPL -0.19%) and Nvidia (NVDA -3.38%) utilize TSMC's production capabilities, as they design the chips in-house and then send them to TSMC to be produced. As companies like Apple and Nvidia push to have more secure supply chains, TSMC's $165 billion investment in its Arizona production facility makes it an even more attractive partner. Demand for U.S.-produced chips has been overwhelming, and production capacities at the Arizona facility that's currently operational have reportedly sold out through 2027. With Intel's (INTC -2.29%) chip foundry business struggling and TSMC's booming, it's clear which company has become the preferred partner for many of the leading tech companies. Another factor in TSMC's success is its drive to continuously push the edge of what's possible. Although Taiwan Semiconductor has the best possible technology in the world with 3 nanometer (nm) chips, it's pursuing a 2nm chip that is slated for launch later this year. This chip node is seeing massive pre-production demand, as it will consume 25% to 30% less power when configured at the same speed as a 3nm chip. This is a big deal, as it helps energy-intensive artificial intelligence (AI) workloads become more efficient as TSMC's 2nm chips roll out. Beyond that, it's working on the A16 and A14 chips that also improve energy consumption. TSMC has cemented itself at the center of all cutting-edge technology, which will allow it to succeed as AI demand rises and falls, and a different technology boom takes its place (such as self-driving cars). Despite all of its success and its lucrative market position, Taiwan Semiconductor's stock doesn't get a ton of respect. TSMC's revenue grew at a 44% pace in U.S. dollars during Q2. That's a speed that few companies outside of Nvidia can match, yet TSMC trades at a fairly low price tag. At 23.9 times forward earnings, Taiwan Semiconductor trades at basically the same price as the broader market, as measured by the S&P 500 (^GSPC -0.64%). The S&P 500 trades for 23.7 times forward earnings, so despite TSMC's rapid growth rate, it doesn't have a premium valuation. Management is quite bullish on its long-term outlook. For the next five years starting in 2025, it believes its AI-related chip revenue will rise at a 45% compound annual growth rate (CAGR) and overall revenue will increase at nearly 20%. That's monstrous growth, and shows that 2025 will not be the end of heightened chip demand. Investors already know that many of the AI hyperscalers are going to be increasing their data center capital expenditures during 2026. These buildouts take several years, so it's clear that overall demand for AI computing capacity is still increasing. This is a multiyear growth driver for TSMC, and will be one of the primary reasons why the company eventually reaches a $2 trillion market cap. I think Taiwan Semiconductor is one of the best buys on the market today, and investors should consider scooping up shares before the market decides to give it an even higher premium due to its growth and position in the AI arms race.

This Underrated AI Stock Has Zero Hype and Massive Free Cash Flow | The Motley Fool

However, all these chipmakers hand their designs to TSMC for large-scale manufacturing, turning them into real products. That's why it's not only one of the best, but one of the safest ways to invest in the AI infrastructure buildout. It wins no matter which chip designer takes the lead, and it's generating a ton of cash doing it. Even Nvidia's CEO Jensen Huang went out of his way to praise the company. He called TSMC "one of the greatest companies in the history of humanity," adding that "anybody who wants to buy TSMC stock is a very smart person." That is not the kind of praise Huang throws around lightly. TSMC is the top foundry in the world, producing most of the world's advanced chips. Rival Intel (INTC 2.20%) has been trying to build its own foundry business, but it is losing money and hasn't been able to gain any ground. In fact, the U.S. government recently made a large investment in the struggling company, reportedly to help bolster it. Samsung, meanwhile, has struggled with production yields. It also recently lost one of its advanced chip designs, as Alphabet switched to TSMC for its Tensor G5 chip used in its Pixel smartphones. Neither Intel nor Samsung has shown that they can match the scale or reliability of TSMC. That's why TSMC has locked in almost every large AI chipmaker as a customer. Chip designers are constantly looking to shrink node sizes, and TSMC is the only foundry that has shown it can consistently produce advanced nodes with strong yields. Nodes are a reference to the size of the transistors used on a chip, measured in nanometers. With smaller nodes, more transistors can be packed onto the chip, which improves performance and power efficiency. Smaller nodes are becoming an increasingly larger part of TSMC's mix. Chips built on 7-nanometer or smaller nodes are already nearly three-quarters of TSMC's revenue, while its 3nm chips alone are almost one-quarter. Meanwhile, it is already preparing to move into 2nm. One of the most overlooked parts of TSMC's story is its cash generation. In 2024, it produced more than $26.5 billion in free cash flow. That was after spending heavily on building new fabs. So far this year, it's already generated over $15 billion in free cash flow despite continued heavy capex spending. It's also paying a growing dividend off that mountain of cash. Most people think of foundries as low gross margin businesses; however, TSMC is changing that narrative. Its leadership in advanced nodes has given it strong pricing power over the years. Nobody else can deliver chips at the same density and yield, so customers are willing to pay up. That's why its margins have stayed strong and have been increasing. Investors don't talk about TSMC with the same excitement they talk about Nvidia or AMD. That could be because it's not a brand consumers recognize, or perhaps because the foundry business isn't just quite as exciting. It's also not a U.S.-based company, with its headquarters in Taiwan. However, TSMC has been one of the biggest beneficiaries of the AI buildout, and it should continue to be a big winner moving forward. Last quarter, its revenue climbed 44% to $30 billion, while its profits soared. Meanwhile, management expects AI chip demand to grow more than 40% annually through 2028. The company is working closely with its largest customers to increase capacity, so it should have good visibility into this growth. Overall, TSMC is one of the most important companies in the AI supply chain. Without it, the current AI infrastructure buildout wouldn't be possible. It's growing rapidly, expanding margins, and generating a boatload of cash. Despite that, the stock is one of the most attractively valued AI plays in the market, trading at a forward price-to-earnings (P/E) ratio of 24.5 times based on analysts' 2025 estimates and a price/earnings-to-growth ratio (PEG) of less than 0.65. Stocks with PEG ratios below 1 are generally considered undervalued. Investors would be smart to heed Jensen Huang's advice and be a buyer of TSMC.

Prediction: This Monster Artificial Intelligence (AI) Chip Stock Will Soar in September (Hint: It's Not Nvidia) | The Motley Fool

Over the past several weeks, investors have been bombarded with a wave of updates as companies reported earnings results for the second calendar quarter. For technology investors, artificial intelligence (AI) remains the dominant theme fueling the sector higher. As I write this (mid-day on Aug. 27), all of the "Magnificent Seven" have posted earnings -- with the lone exception being Nvidia (NVDA -3.38%), which reports later today. Still, the breadcrumbs left by big tech point to an undeniable trend: Spending on AI infrastructure is accelerating. Let's explore why I'm optimistic about Broadcom's upcoming earnings report, and assess whether the stock is a compelling buy at current levels. One of Broadcom's key AI growth drivers comes from its application-specific integrated circuits (ASICs) business. These custom silicon solutions allow customers to design chips that are optimized for their unique workloads. By integrating purpose-built performance with compute power efficiency, Broadcom's ASICs help hyperscalers lower their total cost of infrastructure relative to relying solely on off-the-shelf accelerators from the likes of Nvidia. This becomes highly desirable as training and inferencing workloads scale and become increasingly complex as more sophisticated AI use cases unfold. Broadcom's networking division is also positioned to benefit materially from the ongoing AI infrastructure cycle. As big tech continues to pour hundreds of billions of dollars annually into GPU deployment, Broadcom's supporting infrastructure becomes an indispensable unsung hero. The company's portfolio of high-performance switches, interconnects, and optical components delivers low-latency, high-bandwidth connectivity to keep next-generation accelerators running at full speed. In essence, the company's networking gear represents a foundational layer of AI data center construction -- ensuring scalability and efficiency as workloads expand. With a forward price-to-earnings (P/E) multiple of 45, Broadcom certainly isn't trading at a discount. In fact, its multiple sits near peak levels seen during the AI revolution. Even so, the company's board of directors authorized a $10 billion stock buyback program back in April. Share buybacks at elevated valuations can point to a strong signal: Management remains confident in Broadcom's long-term growth trajectory, underscored by ongoing hyperscaler investment. On a more subtle note, sometimes companies repurchase their own shares when management thinks the stock is undervalued. These dynamics could suggest that Broadcom is positioned for sustained, robust earnings growth, which could fuel further valuation expansion -- even in the face of a premium multiple. For the last few years, the AI trade has largely surrounded Nvidia and the cloud hyperscalers. Yet as infrastructure spending accelerates, the scope of the AI opportunity is broadening to other mission-critical enablers such as Broadcom. Custom chips, high-performance networking equipment, and integrated systems are now just as essential as securing GPUs -- and Broadcom sits squarely at this intersection. In my eyes, Broadcom is approaching its own "Nvidia moment" -- a potential inflection where the narrative begins to recognize Broadcom as a supporting pillar of AI infrastructure and not simply an ancillary beneficiary of these tailwinds. Against this backdrop, I predict that Broadcom's September earnings report will reinforce its strategic importance in the AI landscape -- fueling investor enthusiasm and a further rerating of the stock. For these reasons, I see Broadcom as a compelling opportunity to buy and hold over a long-term time horizon.

Prediction: This Artificial Intelligence (AI) Semiconductor Stock Will Join Nvidia, Microsoft, Apple, Alphabet, and Amazon in the $2 Trillion Club by 2028. (Hint: Not Broadcom) | The Motley Fool

This company is growing quickly, and its stock is a bargain at the current price. Big tech companies are set to spend $375 billion on artificial intelligence (AI) infrastructure this year, according to estimates from analysts at UBS. That number will climb to $500 billion next year. The biggest expense item in building out AI data centers is semiconductors. Nvidia (NVDA -3.38%) has been by far the biggest beneficiary of that spend so far. Its GPUs offer best-in-class capabilities for general AI training and inference. Other AI accelerator chipmakers have also seen strong sales growth, including Broadcom (AVGO -3.70%), which makes custom AI chips as well as networking chips, which ensure data moves efficiently from one server to another, keeping downtime to a minimum. Broadcom's stock price has increased more than fivefold since the start of 2023, and the company now sports a market cap of $1.4 trillion. Another year of spectacular growth could easily place it in the $2 trillion club. But another semiconductor stock looks like a more likely candidate to reach that vaunted level, joining Nvidia and the four other members of the club by 2028. Broadcom is a massive company with operations spanning hardware and software, but its AI chips business is currently steering the ship. To that end, AI revenue climbed 46% year over year last quarter to reach $4.4 billion. Management expects the current quarter to produce $5.1 billion in AI semiconductor revenue, accelerating growth to roughly 60%. AI-related revenue now accounts for roughly 30% of Broadcom's sales, and that's set to keep climbing over the next few years. Broadcom's acquisition of VMware last year is another growth driver. The software company is now fully integrated into Broadcom's larger operations, and it's seen strong success in upselling customers to the VMware Cloud Foundation, enabling enterprises to run their own private clouds. Over 87% of its customers have transitioned to the new subscription, resulting in double-digit growth in annual recurring revenue. But Broadcom shares are extremely expensive. The stock garners a forward P/E ratio of 45. While its AI chip sales are growing quickly and it's seeing strong margin improvement from VMware, it's important not to lose sight of how broad a company Broadcom is. Despite the stellar growth in those two businesses, the company is still only growing its top line at about 20% year over year. Investors should expect only incremental margin improvements going forward as it scales the AI accelerator business. That means the business is set up for strong earnings growth, but not enough to justify its 45 times earnings multiple. Another semiconductor stock trades at a much more reasonable multiple, and is growing just as fast. Both Broadcom and Nvidia rely on another company to ensure they can create the most advanced semiconductors in the world for AI training and inference. That company is Taiwan Semiconductor Manufacturing (TSM -3.05%), which actually prints and packages both companies' designs. Almost every company designing leading-edge chips relies on TSMC for its technological capabilities. As a result, its market share of semiconductor manufacturing has climbed to more than two-thirds. TSMC benefits from a virtuous cycle, ensuring it maintains and grows its massive market share. Its technology lead helps it win big contracts from companies like Nvidia and Broadcom. That gives it the capital to invest in expanding capacity and research and development for its next-generation process. As a result, it maintains its technology lead while offering enough capacity to meet the growing demand for manufacturing. TSMC's leading-edge process node, dubbed N2, will reportedly charge a 66% premium per silicon wafer over the previous generation (N3). That's a much bigger step-up in price than it's historically managed, but the demand for the process is strong as companies are willing to spend whatever it takes to access the next bump in power and energy efficiency. While TSMC typically experiences a significant drop off in gross margin as it ramps up a new expensive node with lower initial yields, its current pricing should help it maintain its margins for years to come as it eventually transitions to an even more advanced process next year. Management expects AI-related revenue to average mid-40% growth per year from 2024 through 2029. While AI chips are still a relatively small part of TSMC's business, that should produce overall revenue growth of about 20% for the business. Its ability to maintain a strong gross margin as it ramps up the next two manufacturing processes should allow it to produce operating earnings growth exceeding that 20% mark. TSMC's stock trades at a much more reasonable earnings multiple of 24 times expectations. Considering the business could generate earnings growth in the low 20% range, that's a great price for the stock. If it can maintain that earnings multiple through 2028 while growing earnings at about 20% per year, the stock will be worth well over $2 trillion at that point.