When a group of stocks regularly outperforms the broader market, Wall Street analysts like to bundle them together so it's easy for investors to separate them from the rest of the pack. In 2017, CNBC financial analyst Jim Cramer coined the FAANG acronym to describe five of the largest technology companies at the time:
- Facebook, which now trades as Meta Platforms
- Apple
- Amazon
- Netflix
- Google, which now trades as Alphabet
In 2023, Bank of America analyst Michael Hartnett identified a new group of stocks and named them the "Magnificent Seven." Those stocks delivered an average return of 112% last year, which crushed the 24% gain in the S&P 500 index. The Magnificent Seven includes:
- Meta Platforms
- Apple
- Amazon
- Alphabet
- Microsoft
- Nvidia (NVDA 0.03%)
- Tesla
Image source: Nvidia.
2024 might be dominated by the "AI Five"
Despite most of the Magnificent Seven heading higher so far in 2024, Tesla stock trades down by 33% and Apple is down by 8% so far. It prompted one analyst -- Glen Kacher from Light Street Capital -- to rethink the stock market's leadership altogether.
Considering artificial intelligence (AI) is such a dominant theme among the market's top-performing stocks, Kacher has identified five companies operating at the forefront of the fast-growing industry. He calls them the "AI Five:"
- Nvidia
- Taiwan Semiconductor Manufacturing (TSM 0.14%)
- Advanced Micro Devices
- Broadcom
- Microsoft
Here's why two stocks in that group are a buy right now.
1. Nvidia
Nvidia opened 2023 with a market capitalization of around $360 billion. It added $1.8 trillion to that figure in the 16 months since, making it the third-largest company in the world as measured by market cap behind Microsoft and Apple. It's all thanks to Nvidia's graphics processing units (GPUs) for the data center, which are the most sought-after in the industry for handling AI workloads.
The H100 GPU sent Nvidia's data center revenue soaring 217% year over year in fiscal 2024 (ended Jan. 28) to $47.5 billion. Now, the company is ramping up shipments of its new H200, which provides up to twice the inferencing speed while consuming half the amount of energy of its predecessor. Inferencing is the process of feeding live data to an AI model so it can make predictions, leading to faster and more accurate outputs for the end user.
Nvidia's GPUs are at the top of most data center operators' wish lists, including Amazon, Microsoft, and Google. Social media giant Meta Platforms even agreed to buy 350,000 H100 chips in the coming year (with an estimated price tag of $9 billion) to train its Llama large language models (LLMs). The only thing that stopped Nvidia's revenue from growing even faster last year was a lack of supply; it simply couldn't keep up with the unprecedented demand.
Investors' attention has already turned to Nvidia's new Blackwell architecture, which was unveiled in March. A GPU called the B200 will start shipping next year, and it will come in a DGX configuration that links eight of them together to provide 3 times faster training performance and 15 times faster inferencing speed than the equivalent DGX H100 setup. In other words, it will allow developers to bring their AI applications to market substantially faster.
Wall Street expects Nvidia to deliver $24.44 in earnings per share during fiscal 2025 (which began in February), marking an 88% increase from fiscal 2024. Based on Nvidia's current stock price of $880.08, it trades at a forward price-to-earnings (P/E) ratio of 36.
That's a 20% premium to the current 30.1 P/E ratio of the Nasdaq-100, but Nvidia's blistering growth means its stock could look cheap by the end of the fiscal year. However, investors who wait until then might find they've missed the boat -- as has been the case for the past year.
2. Taiwan Semiconductor Manufacturing (TSM)
TSMC is one of the most important companies on the planet right now. It manufactures more than half of the world's chips, including those for data centers, smartphones, cars, and general consumer electronics. Most importantly, TSM is the go-to manufacturer for leading semiconductor designers like Nvidia and Advanced Micro Devices.
TSMC's revenue sank by 8.7% year over year in 2023 to $69.3 billion. Many end markets were grappling with supply gluts that began in 2022, when rising inflation and interest rates prompted consumers to trim their spending on big-ticket electronics. But things began to turn in the fourth quarter, because while revenue was down 1.5% year over year, it was up 13.6% sequentially.
That included growth of 17% in the high-performance computing (HPC) segment and 27% growth in the smartphone segment. The HPC result was driven by strong demand for AI-related data center chips, and the smartphone result was partly attributable to Apple's latest 3-nanometer A17 Pro chip.
The HPC segment accounted for 43% of TSMC's total revenue in 2023, and that figure consistently increases each year as tech giants race to fill their data centers with AI chips. In fact, the company expects HPC to be the fastest growing piece of its entire business this year, which means it will represent an even larger share of total revenue.
But smartphones will also be an important source of long-term demand as AI processing migrates from the data center to the device. Apple's A17 Pro chip was the first to utilize TSMC's 3-nanometer fabrication process, and it's designed to speed up AI workloads in applications like the Siri voice assistant on the iPhone. Plus, Apple is currently negotiating with companies like Alphabet and OpenAI to determine which AI chatbot will be installed on future models of the iPhone by default.
Running those models on-device will require complex chips capable of delivering significantly more processing power, and that's good news for TSMC.
TSMC estimates its revenue will grow by more than 20% in 2024, sharply reversing the decline in 2023. Wall Street thinks it will grow by another 20% in 2025 to top $100 billion for the first time. Here's the best part: Based on TSMC's $5.18 in 2023 earnings per share, its stock trades at a P/E ratio of just 27.2, which is substantially cheaper than both Nvidia and the Nasdaq-100 index.
